Prism 7 User Guide
Prism 7 User Guide
Guide
GraphPad Software Inc.
www.graphpad.com
Table of Contents
Foreword 0
What's new? XY
..........................................................................................................................................................
Graphs 59
What's new? Grouped
..........................................................................................................................................................
graphs 61
What's new? Layouts
.......................................................................................................................................................... 62
7 What...................................................................................................................................
was new in Prism 5? 63
3
4 GraphPad Prism7 User Guide
9 Saving
...................................................................................................................................
vs. exporting 126
10 Template,
...................................................................................................................................
method and example files 126
11 Closing
...................................................................................................................................
a Prism file without exiting Prism 127
Changing a data
..........................................................................................................................................................
table format 189
Entering dates
..........................................................................................................................................................
or elapsed times as X 191
Column titles .......................................................................................................................................................... 194
Subcolumn titles
.......................................................................................................................................................... 196
Row titles .......................................................................................................................................................... 198
Sorting data .......................................................................................................................................................... 200
Changing data
..........................................................................................................................................................
table column widths 202
Specifying decimal
..........................................................................................................................................................
places in a data table 202
Excluding or highlighting
..........................................................................................................................................................
values 203
Deleting or removing
..........................................................................................................................................................
entire data sets 205
Missing values
.......................................................................................................................................................... 206
Keyboard shortcuts
.......................................................................................................................................................... 208
Rounding .......................................................................................................................................................... 209
Data objects .......................................................................................................................................................... 211
Transforming..........................................................................................................................................................
and normalizing? 211
5
6 GraphPad Prism7 User Guide
7
8 GraphPad Prism7 User Guide
Changing.........................................................................................................................................................
the plotting order of data sets 419
Nudging survival
.........................................................................................................................................................
data sets 420
Parts of whole..........................................................................................................................................................
graphs 422
Changing.........................................................................................................................................................
the look of parts of whole graphs 422
Changing.........................................................................................................................................................
which data set is plotted 425
4 Format
...................................................................................................................................
Axes dialog 426
Axis range, numbering
..........................................................................................................................................................
& titles 427
Axis range......................................................................................................................................................... 427
Axis numbering
.........................................................................................................................................................
or titles 428
Major and.........................................................................................................................................................
minor ticks 431
Hide an axis
.........................................................................................................................................................
and show scale bars 432
Discontinuous
.........................................................................................................................................................
axes 434
Additional.........................................................................................................................................................
ticks and grids 437
Axis titles......................................................................................................................................................... 443
Labeling axes..........................................................................................................................................................
with dates and times 446
Elapsed time
......................................................................................................................................................... 446
Dates (New
.........................................................................................................................................................
Prism 7 method) 448
Dates (Method
.........................................................................................................................................................
used with prior versions) 450
Logarithmic axes
.......................................................................................................................................................... 452
What is a.........................................................................................................................................................
logarithmic axis? 453
How to: Create
.........................................................................................................................................................
a log axis 455
Logarithmic
.........................................................................................................................................................
axes cannot contain zero or negative numbers 457
Regression
.........................................................................................................................................................
fits the data, not the graph 458
When to use
.........................................................................................................................................................
a logarithmic axis 459
Advice: Bar
.........................................................................................................................................................
graphs with log axes rarely make sense 461
Graphing .........................................................................................................................................................
logarithms - different than using a log axis 463
A review of
.........................................................................................................................................................
logarithms 466
Frame and origin
.......................................................................................................................................................... 468
Positioning
.........................................................................................................................................................
a graph on the page 469
Customizing
.........................................................................................................................................................
a graph's frame and axes appearance 471
Showing grid
.........................................................................................................................................................
lines 473
Setting the
.........................................................................................................................................................
graph origin 475
Graphs: Frame
.........................................................................................................................................................
and axes 476
5 Tips...................................................................................................................................
for moving or rotating objects 499
6 Aligning
...................................................................................................................................
objects on a graph 500
7 Adding
...................................................................................................................................
analaysis results to a graph 504
9
10 GraphPad Prism7 User Guide
Importing data
..........................................................................................................................................................
from a text file 551
Importing data
..........................................................................................................................................................
from the script file 552
Copy and paste
.......................................................................................................................................................... 553
Changing info..........................................................................................................................................................
constants 554
Exporting a graph
..........................................................................................................................................................
or layout 555
Exporting data,
..........................................................................................................................................................
results or info tables 557
Writing to Prism,
..........................................................................................................................................................
text, HTML, XML or SQL tables 557
Writing to text..........................................................................................................................................................
files 559
The WBlock and
..........................................................................................................................................................
WTBlock commands 561
Send to Microsoft
..........................................................................................................................................................
Office 565
Special purpose
..........................................................................................................................................................
commands 565
5 Running
...................................................................................................................................
a Prism script 567
Creating and ..........................................................................................................................................................
running a script from Prism 568
Run a script using
..........................................................................................................................................................
Automator, MacScript, or AppleScript 569
Launching a Prism
..........................................................................................................................................................
script from outside Prism (Windows) 571
Using Prism scripts
..........................................................................................................................................................
with Excel 574
How long can..........................................................................................................................................................
a script be? 576
Index 599
Then follow longer detailed tours that walk you step by step through data
analyses and graphing.
The data table with sample data, also contains a floating note that
explains how the data are arranged and gives suggestions for graphing
and analyzing.
Follow step-by-step examples, using sample data, for analyzing data with
Prism:
Descriptive statistics
Column statistics
Frequency distribution
Categorical outcomes
Contingency table analysis
Survival analysis
Nonlinear regression
Fitting an enzyme kinetics curve
Comparing two enzyme kinetics models
Automatic outlier elimination (exponential decay)
Global nonlinear regression (dose-response curves)
Ambiguous fit (dose-response)
Choose the Portfolio tab at the bottom of the Welcome dialog to browse
dozens of polished graphs, each of which demonstrates several Prism
features. When you open a portfolio graph, you'll see a brief explanation
of how that graph was made, and can explore the details. Many of these
portfolio graphs are linked to step-by-step instructions. Others give brief
explanations. Exploring these files will help you master graphing with
Prism.
Three-way ANOVA
Label subcolumns
Semitransparent colors
so you can see overlap
more.
Heat maps are a new way to plot grouped data. The plotting area is
divided into squares. The columns corresponds to different data sets in
your table, and the rows in the graph correspond to different rows in the
data table. Each square in the graph is color coded to denote the value
entered into that cell of the table. Features:
· Color scales can encode continuous data with one scale, continuous data
with two scales (perhaps one for positive numbers and another for
negative numbers) or discrete categories (1=blue, 2=red, ...).
Monochrome or color.
· Specify special colors for values that are "off the map" or for excluded
values. Optionally put "X" through those cells.
· If you entered replicate values, base the heat map on the mean, median
or geometric mean of the replicates. Or base it on the SD, SEM or CV
among replicates to make a heat map of variation.
When symbols are semitransparent, this lets you see when several
symbols are superimposed (they get darker).
Until now, you could only enter numbers into the X column of XY tables.
Prism 7 now lets you enter numbers (as always), elapsed time in
hh:mm:ss, or dates. Choose on the Welcome dialog:
If you chose to enter dates, you can enter dates in almost any format.
Prism will figure out what you meant and display the date on the table in
a standard format (that you cannot change).
XY graphs can now label the X axis with the dates or elapsed times you
entered.
To go along with three-way ANOVA, Prism can also create three way
grouped graphs.
Two of the factors must have only two levels (e.g. male/female; low
fat/high fat), but the the third factor (designated by different rows) can
have more levels (more than two rows).
· When plotting medians, Prism offers a new choice of error bars: Median
with 95% CI
· When plotting the geometric mean, Prism now offers to plot error bars
showing the geometric SD. It also shows the geometric SD factor on the
Column statistics results page.
New warnings
· If any error bar is smaller than the symbol, Prism now places a floating
note on the graph explaining why the error bars are not shown.
· Prism will warn you if you mix SD and SEM error bars on the same
graph, as this is usually not intentional.
New default
· Add major grid lines on the categorical axis of Column and Grouped
graphs.
· Some box-and-whisker graphs plot individual points for large and small
values. Prism 6 always plotted these points using solid circles. Prism 7
lets you choose any type of symbol.
· On grouped graphs, specify how large a gap you would like to see when
a data table row is blank, and specify gaps before the first data set and
after the last.
· On column graphs, specify how large a gap you would like to see before
the first data set and after the last.
· When you make one graph per data set, Prism uses the column title of
each data set as the graph title. Now this is a live link, so if you change
the column title, the corresponding graph title will update.
· For bar graphs, the choice to use (or not use) a fill pattern is now more
obvious in the Format Graph dialog. With prior versions you had to
choose a solid fill pattern to get no fill pattern. Now just uncheck the
option if you don't want a fill pattern.
Exporting
· When exporting color graphs in tiff or jpg format, Prism Windows now
lets you choose grayscale, where colors are converted to various shades
of gray (Prism Mac already offered this choice).
Axes
· Use a comma as decimal separator in axis numbering.
· Set a default for the line width of lines and arrows you draw.
Legends
· Choose square (new) or rectangular legends for bar graphs.
Text
· You can now use flush justification.
Add information to graphs
New file constants let you add file name, and dates to graphs or layouts.
With all prior versions of Prism, the confidence intervals for parameters fit
by nonlinear regression were always symmetrical around the best-fit
value. These are called asymptotic intervals or approximate intervals.
These are are what most nonlinear regression programs report. But the
true uncertainty in a parameter's value, in many cases, is asymmetrical.
In these cases, the symmetrical confidence intervals do not really express
the precision of a parameter's estimate very well.
The initial default is to show the asymmetrical intervals, but you can
change this default (in the Confidence tab of the nonlinear regression
dialog). Since the only purpose of the standard errors of the parameters
is to compute the approximate confidence intervals, we suggest that you
stop showing these once you get used to the new asymmetrical intervals.
Some scientists run many fits, perhaps using a script, or review many fits
run by others. Prism 7 provides a way to flag poor fits, so you you know
which data sets and results to inspect more carefully.
· Define "poor" fits on the new Flags tab on the parameters dialog of
nonlinear regression. Flag a fit because the R2 is too low (you define
how low that is), when the number of outliers is too high (you define
how high that is), when a dependency is too high, or when the residuals
fail a normality test, or the runs or residual test.
· When running Monte Carlo simulations choose to skip any results from
fits that were flagged as being poor.
· When comparing two models, you can tell Prism to skip formal
comparisons when one fit is flagged as being poor, and always choose
the other fit in that case.
· New choices for rules rules for initial values. You can now compute the
initial value as a constant value multiplied or divided by the initial slope,
the final slope, or the slope at the middle of the graph.
o Use the <A> or <B> notation to define different equations for different
data sets.
· The rules for initial values have been improved for the parameter Width
in the log(Gaussian) distribution, for all three parameters in the
equation for enzyme substrate inhibition, and for the Amplitude of a sine
wave.
Prism has been able to perform one- and two-way ANOVA for many
versions, and now can also perform three-way ANOVA.
Two of the factors must have only two levels (e.g. male/female; low
fat/high fat), but the the third factor (designated by different rows) can
have more levels (more than two rows).
The three-way ANOVA dialog is similar to the dialog for one- and two-way
ANOVA, but the choices for multiple comparison goal are a bit different:
Here is a portion of the results table, showing all seven P values. Yes,
seven! Three-way ANOVA is complicated.
This analysis also lets you plot the rank of each P value vs. the actual P
value, a common way to visualize the distribution of a set of P values.
Controlling the False Discovery Rate (FDR) is a great method for coping
with multiple comparisons. Prism 6 offered this as part of the multiple t
test analysis. Prism 7 offers it in three places.
As a followup to ANOVA
Prism has long offered multiple comparisons tests after ANOVA to control
the Type I error rate for the family of comparisons. Now Prism 7 lets you
use an alternative strategy for multiple comparisons following ANOVA
(one-, two- or three-way): controlling the False Discovery Rate (FDR).
Whenever you choose to use the FDR approach to decide which P values
are small enough to be a "discovery", Prism lets you choose one of three
methods for controlling the FDR.
Prism has always had a versatile transform analysis. Prism 7 adds a new
analysis to simplify transforming concentrations 296 entered into the X
column of an XY table.
· Change units, for example from molar to nanomolar. Simply tell Prism
what constant to multiply or divide by.
Most of Prism's analyses report P values, and now all these analyses let
you choose how to report P values.
· The NEJM (New England Journal of Medicine) style, which shows three
digits and includes the leading zero (0.123). P values less than 0.001
shown as "< .001". All P values less than 0.001 are summarized with
three asterisks, with no possibility of four asterisks.
· The now classic GraphPad method which reports four digits after the
decimal point with a leading zero (0.1234). P values less than 0.0001
shown as "< .0001". P values less than 0.001 are summarized with
three asterisks, and P values less than 0.0001 are summarized with four
asterisks.
· Choose how many digits you want to see after the decimal point, up to
15. P values less than 0.001 are given three asterisks, and P values less
than 0.0001 are given four asterisks.
Prism now reports the reciprocal of the relative risk and odds ratio, and
their confidence interval. Depending on the meaning of the rows and
columns, the reciprocal may be more meaningful to you.
Choice of algorithms
The methods used by prior versions were simple approximations, and the
improved methods Prism now offers produce narrower intervals that still
provide the desired confidence level. We suggest you use them routinely.
· The Row Statistics analysis can now report means only without SD or
SEM.
· One-way ANOVA now works even if any of the data set columns has only
a single value.
· Enter starting and ending dates for each subject in a survival analysis
instead of entering number of days (or weeks..).
· If you enter data with replicates (or enter mean, n and SD or SEM), the
Area under the curve (AUC) analysis will now report the AUC with a
standard error and 95% confidence interval.
· The Remove Baseline & Column Math analysis now gives you choices for
reporting the column titles. If the column titles in your data table are
"Control" and "Treated" and you ask the analysis to subtract one from
the other, you can have the second results column labeled as "B-A" or
"Treated-Control" or just "Treated".
and more.
In Prism 6, subcolumns were labeled Y1, Y2, Y3, etc. and you couldn't
change these labels. With Prism 7, you can enter your own subcolumn
titles.
Apply one set of subcolumn titles 194 to all data sets, or label each
subcolumn in each data set uniquely.
Use one color to mark values that you want to double check, and color for
cells you want to discuss with a colleague, another for key results you
want to find again. Or color code different parts of an experiment.
With Prism 6, you could highlight sheet names, but only in yellow. With
Prism 7, choose from nine color when highlighting a sheet name 151 .
You might want to choose one color to highlight the most important
results, another color to mark sheets you need to discuss with a colleague
or mentor, and yet another color to mark the graphs that were included in
a paper.
We know that some people use Prism almost every day, so we work hard
to provide time savers for frequent users. Improvements in Prism 7:
· When creating a layout 479 , assign a series of graphs so you don't have
to add them one at a time.
· With parts-of-whole graphs you can now choose which data set to graph
(these graphs can only plot one data set). Prism 6 only plotted data set
A.
· The script editor shows line numbers, making it easier to edit long
scripts.
· New script command to save all graphs or all layouts in a single pdf file.
· The Wand 136 graphs data on a new table just like you have already
graphed another table (even if that table was not analyzed).
to get your work done without thinking about saving files. Your work is
saved as you go, with no need to ever click the Save button (although
you can). Automatic backup is handled by the Mac Versions and Resume
feature, and is no longer handled by code we wrote.
This is a change Apple introduced a few years ago. If you use iWork
(Pages, Numbers, Keynote), you already know about Versions and
Resume. If you don't use iWork, spend a minute reading about these
features 111 to avoid frustration.
Prism 7 lets you enter data with 13-14 digits for each value (double
precision). In contrast, Prism 6 allowed to to enter only 6-7 digits (single
precision).
Allowing more digits prevents round off errors when importing or pasting
data with more than 6 or 7 digits of precision.
Many people now use computer displays with high resolution. Apple calls
these retina displays.
Prism 6 Mac already worked well on retina displays, but Prism 7 looks
even better with high resolution icons.
Parts-of-whole data 56
Column graphs 57
XY Graphs 59
Grouped graphs 61
Layouts 62
Time savers 50
Statistics
· Two-way ANOVA with repeated measures
in both factors
Nonlinear regression
· Fit to models defined by implicit and
differential equations.
· Write more complex models
· Weight your fit by 1/YK
Separate columns with lines
· Much easier interpolation
Survival analylsis
· Prism now computes the hazard ratio (HR) by two methods: The logrank
method and the Mantel-Haenszel method.
· The Wilcoxon test computes the 95% CI of the median. The Wilcoxon
matched-pairs test (nonparametric test of two paired or matched
groups) reports the confidence interval of the median of the paired
differences, and the Wilcoxon signed rank sum test (nonparametric test
to compare a median with a hypothetical median) now reports the
confidence interval of the difference between the actual sample median
and the hypothetical median.
· Ratio t test. The paired t test works by analyzing the difference between
each pair of values, testing the null hypothesis that the average
difference is zero. With some kinds of data, the difference between
before and after is not a consistent measure of effect. The differences
might be larger when the "before" values are larger, and smaller when
the "before" values are smaller. The ratio (after/before) may be a much
more consistent way to quantify the effect of the treatment. Actually, it
turns out that analyzing the logarithm of ratios works much better.
· Method of Pratt. There are two ways to compute the Wilcoxon matched-
pairs test when some of the pairs are identical, so the difference is zero.
Prism 6 offers both choices. The new choice is the method of Pratt for
tied rows.
Simulations are even more useful when you repeat the simulation many
times, and tabulate the results. Prism 6 makes this possible with a new
Monte Carlo simulation analysis that makes it easy to tabulate the results
of many simulations. First simulate a data table and run an analysis to
analyze that table. Then run the new Monte Carlo analysis. Specify how
many simulations you want to run, and which analysis parameters to
tabulate. You can also define a “hit”, as perhaps a P value less than 0.05,
or a confidence interval including a theoretical value. The results of the
Monte-Carlo analysis include a table with all the tabulated data (that you
can analyze further), and a table of the number of simulations that were
hits.
· New multiple comparisons test. Holm-Sidak has more power than the
Tukey or Dunnett methods, but cannot provide confidence intervals.
· Dunnett's test can now can be computed with 256 groups. Prism 5 was
limited to 20 groups.
· Main and simple effects for multiple comparisons after two-way ANOVA.
· Much easier interpolating. Many people use nonlinear regression for the
purpose of interpolating values from standard curves. If that is your
goal, most of the options provided in the nonlinear regression dialog are
neither required or helpful. Prism 6 offers a new analysis just for
interpolating curves. It offers only essential options, so is super easy to
use.
· Write more complex models. Prism 5 let’s you define different models
for different data sets by designating certain lines in the equation to
only apply to a certain data set. For example, a line in the equation
preceded with <B> applies only to data set B. Prism 6 extends this
syntax to allow more complex equations. Precede a line with <A:D> for
it to apply only to columns A-D, or with <A:K,2> to make it apply to
every second data set between A and K (A, C, E, G, I, K).
· Adjusted R2. Prism reports the adjusted R2 which takes into account the
number of degrees of freedom, and so can be compared between
models with different numbers of parameters.
Linear regression
· Convert from linear to nonlinear regression with one click. Linear
regression can be viewed a special case of nonlinear regression. Prism’s
nonlinear regression analysis can be used to fit a straight line, and there
are more than a dozen reasons why you might want to do so. But few
people think of using the nonlinear regression analysis to fit a straight
line. Prism 6 makes it easier to switch – simply click the “More choices”
button on the linear regression dialog.
· Copy the linear regression equation. Prism always reported the best-fit
values for the slope and intercept, but not in the form of an equation.
Prism 6 now reports the equation for the line, ready to copy and paste
onto a graph or into a manuscript.
· Wand to analyze data. After entering data onto an empty data table,
you may wish to analyze these new data exactly as you have analyzed
another data table. Prism has always provided a tool to do this, but you
probably never found it (it used to be called Applying a Method). Prism
6 introduces the Wand (Wizard to ANalyze Data 136 ) button. Click it to
analyze (and graph) a new data table just like you already analyzed
(and graphed) an existing table.
· Faster closing. Prism 6 Windows can close a large file much more
quickly than did Prism 5.
· Format several data sets at once. The Format Graph dialog in Prism
5 always let you either change one data set at a time, or click the “all”
button to change all data sets at once. Prism 6 also lets you choose a
defined subset of data sets at once. For example, you could select data
sets A, C, E, and G and change all of those data sets to use a dotted
line. Then choose B, D, F and H, and make those lines solid.
· Click the apply button to see changes without closing the Format
Graph or Format Axes dialog. The Format Graph and Format Axes
dialogs are very powerful, and give you a huge number of options to
change the appearance of your graph. To avoid the need to click OK,
then go back to the dialog to make more changes, simply click the new
Apply button and you’ll see the effect of the changes you made without
closing the dialog.
make any changes in the Format Axes or Format Graph dialogs or if you
move the graph on a page or stretch its axes.
· More galleries. Prism 5 lets you view all graphs at once on a gallery.
Ditto for data tables, info tables, analysis results and layouts. Prism 6
also lets you view an entire family of linked sheets in the gallery, or to
see the entire set of search results (see above) as a gallery.
· Graph Portfolio to help you master Prism. Click the Portfolio tab on
the Welcome dialog to browse a portfolio of polished graphs. Open any
of these graphs to explore how it was made.
· Open files you forgot to save! When you exit Prism (as with any
program), it asks you whether you want to save changes to every open
file. Have you ever clicked No by accident and lost your work? Not with
Prism 6! You can open "unsaved" files from the Welcome dialog (unless
you tell Prism not to keep them, or unless you waited more than three
days).
· Easier to find the sample data sets. Prism 5 introduced sample data
sets as a way to learn Prism. Prism 6 makes these more obvious on the
Welcome dialog, so more people will try them.
· More columns 187 . Prism 6 allows you to enter data in up to 256 data set
columns. Prism 5 allowed only 104.
· Paste link data or results from Prism. 229 You’ve probably encountered
situations where you need to copy values from a results table, and then
paste onto a data table, perhaps using options in the Paste Special
dialog to transpose or rearrange the values. Prism 5 could paste the
values, but there was no link, so it was not possible to trace where the
values were pasted from, and the pasted values did not update when
the original data were edited or replaced. Prism 6 lets you paste link so
values pasted onto the data table are linked to their source, and will
automatically update.
· Reformat a select set of points. 320 Select any block of data in a data
table, right click, and choose Format Points to change the color (or size
or shape...) used to graph that subset of points.
· Long column titles 194 . With Prism 6, the length of column titles has no
real limit. Prism 5 limited the length of column titles, and many people
bumped into that limit.
· Multiple line titles. 194 Each row or column title can now have two or
more lines of text.
· Bold, italics and underline. You may also now bold, italicize and
underline text in row and column titles (you could always use
superscript, subscripts and Greek).
· View all column titles at once 194 . A new tab in the Format Data Table
dialog of Prism 6 lets you see and edit all the column titles. This saves
you from the scrolling back and forth to edit titles, which was especially
tedious with XY or Grouped tables with many subcolumns.
· Faster TIFF. Exporting TIFF files at 1200 dpi from Prism Windows is
much faster and more reliable (doesn't lead to out of memory
messages).
· Smaller TIFF. Prism Windows can now compress TIFF files at 600 or
1200 dpi, and compress even when you choose the CMYK color model
(not allowed in Prism 5 Windows)
· Parts of whole tables. 174 Prism has always created structured data
tables, and Prism 6 adds a sixth kind of table: Parts of Whole. As the
name suggests, this table is used for the kind of data where it makes
sense to sum all the values in a data set column, and divide each value
in the table into this total.
· Pie charts and more. 422 Each data set column in the parts of whole table
can become its own pie chart, and each value represents one slice of the
pie. Prism will graph column A automatically. To graph the other
columns, use New...Graph of existing data. Prism can also plot these
kind of data as a donut plot, stacked bar plot, or a percentage dot plot.
· Depth charts. 485 When you assign a graph to a layout, you can specify
that it be rotated 90 degrees. This essentially creates an XY graph
where the independent (X) variable is shown on the vertical axis. As the
graph's name suggests, this kind of graph is used in fields where the
independent variable is depth, which is naturally shown on the vertical
axis.
· Label both sides of the axis. 437 Place additional ticks and labels across
the axis from the regular numbering. This is helpful, for example, when
X is time, but you also want to label particular times when an
experimental intervention was performed.
· Create a graphs with only a right Y axis. 327 Prism has always made it
easy to add a right Y axis as a second axis. But it has been difficult to
create a graph with a single Y axis positioned on the right side. To do
this with Prism 5 can requires positioning the "left" (only) Y axis in on
the right side of the graph, and then assigning data sets to the “left”
axis, even though it is on the right side of the graph. Confusing! Prism 6
makes this more straightforward. Turn off the left Y axis, and create a
right Y axis to which you assign data sets.
· Resize graph to match length of X axis 322 . The resize graph dialog in
Prism 5 lets you change the graph size to make the entire graph object
have a specified width. But sometimes your goal is to match the length
of axes, rather than the size of the entire graph object (including
labels). Prism 6 now lets you specify the desired length of the X axis,
and the rest of the graph changes size to meet that goal.
· Control the scaling factor on a layout 483 . When you place a graph on a
layout, its size usually changes. For example, the total length of a graph
might be half as long on the layout as it is on the graph. We now term
this value (50%) the scaling factor 483 . If you then add a legend, drawing
or image to the graph, Prism 6 will maintain the scaling factor, so the
graph itself won't change size on the layout (but the graph plus its
legend, drawings, etc will get larger). In contrast, Prism 5 maintained
the size of the graph placeholder, so the graph itself would get smaller
on the layout when you added a legend or text box or image to a graph.
· Easier to place a graph from one project into the layout of another. 480
Prism 5 allowed you to link a graph from one project into a layout of
another. This lets you keep each project compact, and avoids the need
to put all the data and results for one layout in one file. But that feature
has been hard to discover, and used an extra popup dialog. Prism 6
makes it easier to discover and use by placing the choice on the first tab
of the layout dialog.
· Thinner lines. We’ve added a new choice to make lines and curves ¼ of
a point thick. With Prism 5, the thinnest choice was ½ point. You won’t
notice the difference on screen, but you may notice the difference on
printouts or high resolution exported files.
In case you are upgrading from Prism 4, the pictures below show the
biggest improvements in Prism 5. For details, read this 16 page pdf
document.
of Prism.
Unlike spreadsheet
programs, and other
scientific graphing
programs, Prism's data
tables are structured or
formatted. Whenever you
create a new data table,
choose one of six table
types. This makes it much
easier to create appropriate
graphs and to choose
suitable analyses.
Consult Prism’s
comprehensive online help
to review the basic
principles of data analysis,
and to understand your
analysis choices.
Prism links data to results, graphs and layouts. It also lets you paste link
data from one part of Prism onto a data table. This has three
advantages.
· When you fix a mistake in data entry, the results and graphs are
automatically updated.
· When you open a Prism file, you can retrace every step of every
analysis, and review (or change) any of the choices. You’ll never see a
curve or a P value and wonder how it got there.
· After you’ve polished the analysis and graphing steps with data from
one experiment, you don’t have to repeat all those steps for a repeat
experiment. Prism provides several ways to instantly analyze and graph
a repeat experiment, without repeating any tedious steps and without
requiring any scripting or programming.
Add floating notes to data tables, results and graphs. These notes can be
color coded.
Use Info tables to record dates, lot numbers, concentrations, etc. These
values can be used as constants when transforming or analyzing data.
Highlight the pages that you want to review, or want a colleague to look
at. Choose a different color highlight for different purposes.
Select any cell or range of cells in data tables or results and apply a
background color.
You can send Prism files even to colleagues who don't own Prism. They
can use the Prism viewer for Windows, Mac or iPad to view all parts of
your file. Details.
This five-minute tutorial will show you a bit about how Prism is organized.
1. When you start Prism, you'll see the Welcome dialog. You'll see six tabs
representing the six kinds of data tables. Click on the Column tab. Then
choose to use tutorial data for an unpaired t test.
Notes:
· To use Prism effectively, you need to learn how to choose which of the
six types of data table 161 to select.
· Choosing the tutorial data for the unpaired t test only fills the table with
data. You'll choose an analysis in the next step.
2. Note that the tutorial data comes with a floating yellow note that briefly
explains the data format, and links to more detailed online help.
3. From the data table, click Analyze and then choose the t test analysis.
Alternatively, click the t test shortcut button above the Analyze button.
4. On the first tab of the t test dialog, confirm that the experimental
design in unpaired and that you want to assume a Gaussian distribution
and to perform an unpaired t test. For now, you can accept all the
defaults on the second tab. Click the Learn button to read explanations
of all the choices..
5. View the results. Click the analysis checklist button to help interpret
the results. Click the Results button in the upper left corner to open the
Analysis Parameters dialog to perhaps change to a paired or
nonparametric test.
Prism data tables are organized. It is essential to choose the right kind of
data table.
A Prism project (another term for a Prism file) is not limited to a single
graph or data table, but rather can contain up to 500 data tables, graphs,
analyses, and page layouts.
Prism can automatically plot error bars from raw data
One of Prism's strengths is its ability to handle replicates and error bars.
For XY and Grouped tables, replicates are placed in side-by-side
subcolumns. For Column tables, replicates are stacked in each column.If
you enter replicate values, Prism can plot either individual replicates or
error bars. It can plot error bars from the replicates you entered
automatically. You don't have to specify any calculations.
When you edit or replace data, analyses and graphs are automatically
updated
Prism remembers the logical links between data tables, info sheets,
results tables, graphs and layouts. When you edit or replace data, Prism
automatically recomputes linked analyses and redraws linked graphs.
Analyses can be chained
Results tables can be analyzed further. From any results table with green
grid lines, click the Analyze button. This lets you, say, first transform
your data, and then fit a curve (nonlinear regression) to the results.
By default, Prism creates one graph for each data table you create, but
this should not limit you. You can plot a table on multiple graphs. And you
can plot data from multiple data tables on one graph.
Recycle your work
Once you've analyzed and graphed one data table, it is easy to repeat
your work with new data by duplicating a family 134 or cloning a graph 135 .
Open a file
After you have used Prism a bit, you may see files listed
under the heading Unsaved Files. How can you open a file
that wasn't saved? When you exit Prism (as with any
program), it asks you whether you want to save changes to
every open file. If you click No, Prism still saves the file in a
special location for four days, and shows it in the list of
unsaved files.
LabArchives
Clone a graph
graph, Prism will give you the option of deleting the data.
Add your own data, and your done. From the Welcome
dialog, you can clone from any file (project) that is open,
from one you've used recently, or from projects you've save
as examples.
Portfolio
From the Welcome dialog, click on the Graph Portfolio tab to browse
dozens of polished graphs. Open any of them to see how they were made.
Most come with explanations on floating notes.
A Prism project can contain a lot more than a single graph or data table.
A Prism project (file) has five sections. Each can contain up to 500
pages (also called sheets):
Data tables
Info tables
Results
These sheets show the results of your analyses. You can copy from
here and paste parts of your results onto graphs. Learn more. 258
Graphs
Layouts
Click the New button on the Sheet toolbar and choose what kind of sheet
you want to create. The first choice is used the most. It adds a new
empty data table and linked graph 175 to your project.
You can also add a New Data Table (No Automatic Graph), a new Info
sheet 147 , a new graph of an existing data table 305 , or a new layout 479 . To
be complete, this menu also lets you add a new analysis, but you'll find it
easier to start at a data table or graph and click the Analyze button 259 .
Don't overuse this New button. You can also add to your project by
duplicating a family of sheets 134 or by cloning a graph 135 .
If you are new to Prism, start by reading this one page list 82 of
essential concepts.
Pay attention to the distinctions between the six kinds of data tables
Once you've analyzed data and made some graphs with Prism, you'll
probably want to do similar analyses and make similar graphs with
new data. Don't start again from scratch. Prism offers four ways to
recycle your work, without programming. You can duplicate a
family 134 , clone a graph 135 , use the Wand to repeat analyses 136 , or
save and recycle templates 138 .
Navigation tips
View all sheets in the gallery
Click on the Data Table, Info, Results, Graphs or Layouts folder name
in the Navigator to open the Gallery. Or click the Gallery button on
the bottom toolbar.
The gallery 144 shows all sheets in a section. Double click on the one
you want to go to. Or select several sheets to work on at once 144 .
Open the Family 100 folder to see all sheets linked to the current one.
Or click the Link button in the bottom toolbar to jump to one of those
linked sheets.
Open the Data with Results folder 262 to trace an analysis chain (when
the result of one analysis is the input to another).
Click the highlight 151 button, and the sheet name will appear
highlighted in the navigator and gallery.
Annotate any sheet with a floating note 150 , then find all annotated
sheets in a special folder at the bottom of the Navigator 100 .
If your project is large, click the search button 154 on the bottom
toolbar to find all sheets that include specified text.
Graphing Tips
Change the color (or size..) of just one symbol or bar
Right click 320 on any symbol or bar, and you can change its
appearance, apart from other symbols or bars in the same data
set. Or select a range of values in the data table, and change the
appearance of all symbols or bars plotting those 320 data 320 .
Select one (or several graphs), click the Magic 317 (Make Graphs
Consistent) button, choose an example graph, and all your graphs
will be changed at once.
Prism can plot error bars directly from raw data. Prism lets you enter
SD or SEM if you have calculated them elsewhere, but there is no
need to do so. Enter your raw data, and let Prism do the rest. Prism
offers analyses that perform row and column descriptive statistics, but
these are only to help you understand the data. You do not need to
run these analyses to plot error bars.
To add data to a graph, drag a table from the Navigator and drop onto the graph
After you drop the data or results table onto the graph, a dialog will
pop up where you can choose which data sets to include and which
axis to plot those data on (right or left).
Click Learn from any Analyze dialog to learn more about the analysis.
Our help goes beyond program help, and explains the principles
behind the analysis.
Once you've finished any analysis, click the Analysis Checklist button
in the toolbar to make sure you are interpreting the results correctly.
Select any block of tabular results, copy to the clipboard, and paste
onto any graph or layout. This will be a live link. If you change the
data, or the analysis choices, the embedded results table 265 will
update.
Exporting tips
Use the Send-to-Word and Send-to-PowerPoint buttons
Click one button 521 and you are done. No need to copy and paste,
or export and import.
Send your colleagues complete Prism files rather than exported images
Anyone can view the complete Prism file, including analysis choices,
using the Prism Viewer for Windows, Mac or iPad.
Organizing tips
Rename, reorder and delete sheets
You aren't stuck with the sheet names that Prism assigns, or their
order. Rename or reorder 143 your data tables (and graphs...). When
you rename data tables, linked analyses and graphs are
automatically renamed. Delete sheets 143 you no longer need.
You'll see the yellow highlighting 151 in both the navigator and gallery.
Click the push-pin icon to add a color-coded floating note 150 to any
sheet.
Info sheets 147 contain both structured information (lot numbers, page
numbers,....) and unstructured notes. Each info sheet can be linked
to a particular data table.
Some people make the mistake of starting a new project for every
data table, not realizing that a Prism project (file) can have up to 500
data tables, analyses and graphs. You can keep an entire project in
one file.
Common mistakes
Mistake: Repeating an analysis rather than fixing it.
If you want to change your mind about analysis choices, you can do
so 263 . You can change the parameters (options) for the analysis. Or
you can change which data table and which data sets are analyzed.
There is no need to start the analysis again, which leaves an obsolete
copy of the results in your project.
No need to change the fonts, colors, line thickness, etc. etc. for each
graph in a series. Instead, format one graph, and use Magic 317 (Make
Graphs Consistent 317 ) to instantly fix the others.
We call a Prism file a "project", but you can decide how large to
make it. Huge files can be hard to navigate. So it often makes sense
to keep files of reasonable size. You can always merge 123 projects later,
or include 480 a graph from one project in a layout of another.
Mistake: Going through many steps to make the same graph from a new data set.
There is no need to repeat yourself. Once you have a graph you like,
Prism offers several ways to make that graph again with different
data. Start by learning to clone a graph 135 .
We work hard to keep the Windows and Mac versions of Prism very
similar. You can switch back and forth between platforms easily, and files
created on one platform can be opened on the other. But there are a few
differences.
· Object linking and embedding 523 . This means you can paste a Prism
graph into another program and double-click to edit in Prism. We wish
we could allow this to work on Prism Mac, but Apple and Microsoft
simply do not provide the "hooks" Prism would need.
· Excel linking and embedding 233 . This means you can paste Excel data
into Prism, and double click to edit within Excel. We also wish this were
possible with the Mac.
· Import and export EMF and WMF images. These formats are only used
on Windows computers.
· Visual Basic for Applications (VBA) commands 574 to launch Prism and
start a script.
· Color eyedropper. While setting any color, click the magnifying glass
tool, then click on any color on any window to select that color.
· Use an Automator action 569 to launch Prism and run a Prism Script;
File compatibility
Files saved on one platform can be opened on the other. In almost all
respects, this just works and you don't have to think about it. But there
are two exceptions:
· If you use Prism Mac and copy a graph from one project and paste it
onto a layout on another project, Prism can handle this in two ways. By
default it pastes in a form that works only with Prism Mac. If you open
the project in Windows, that graph will be missing from the layout. Use
an option in the Files tab of Prism Preferences to save in a compatible
format. On the Mac, you will notice that the files will be larger and
pasting graphs from other projects noticeably slower (but still only a few
seconds). Now the file will open fine in Windows.
· Tabs within titles or text objects are handled differently on Windows and
Mac. If you include tabs within text objects, the spacing of that text may
change a lot when you open it on the other platform.
perform tasks.
The most versatile way to jump from sheet to sheet is by using the Prism
Navigator tree on the left side of the Prism screen.
The names of all sheets related to the current sheet are shown in bold.
Click on any sheet name to go there.
If you need more room to work, click the button in the lower left corner of
the Prism window to hide the Navigator. Click it again to bring the
Navigator back.
See also:
The gallery shows you thumbnails (small images) of all the sheets in a
section at once.
To view the gallery, click on the folder name in the navigator. Click on
"Family, Data Table", "Info", "Results", "Graphs" or "Layouts".
Another way to view the Gallery is to click its button on bottom toolbar.
Notes:
· Don't confuse the Gallery, which shows you all the sheets in your
project, with the Portfolio which shows you a set of polished graphs
you can use as examples.View the Portfolio from the Welcome dialog.
· From the gallery, double-click on any thumbnail to leave the gallery
and go to that sheet. This is the main purpose of the gallery -- to help
you quickly find a particular sheet and go to it. Alternatively, you can
leave the gallery by clicking on the gallery button in the lower toolbar
shown above.
Some of the tool tips in Prism do more than provide information. They
also provide links to take you to other parts of the project.
Hover your mouse over a symbol on a graph to see the value, and
location of the data point. Click on the name of the data (or results)
table to go right to the correct spot on the data table.
Hover your mouse over any graph on a Layout. A tool tip will appear
with a link to the graph. Click on the link to go to the graph.
Zoom buttons
Use the Zoom buttons located on the lower right side of the Prism screen
to enlarge or shrink the view of the current sheet. You can also use the
Zoom tools to make thumbnails in the Gallery 144 larger or smaller.
On graph and layout sheets, Zooming lets you see more or less of the
sheet, but doesn't change the size of the graph or layout when printed.
Click here to find out how to actually resize graphs 322 and layouts.
On data tables and analysis results sheets, the Zoom buttons not only
change what you see, but they also actually change the size of the font
Prism uses when you print. If you want to get more of your data table on
a printed page, Zoom Out to make the font smaller.
Keyboard shortcuts
Shortcut: Hold the Ctrl key and press M to zoom larger or J to zoom
smaller.
Mac gestures
Starting with release 6.0c (and if you use Lion or later), you can zoom
Prism Mac by pinching with two fingers, but only if that gesture is turned
on in the Trackpad control panel. You can also smart zoom by double
tapping with two fingers. Depending on where you tap, that will zoom you
in to see the entire page, the graph plus all other objects on the page, or
just the graph.
It is possible to use Prism without the Toolbar (top) and Navigator (left),
and this can be useful on computers with tiny screens. But it is easier to
learn Prism with both the Toolbar and Navigator showing. If someone else
using the computer has turned off either, it is easy to bring it back:
· If the toolbar is not showing on your computer, drop the View menu and
choose Main Toolbar.
· If the Navigator is not showing on your computer, drop the View menu
and choose Navigator.
A Prism project can have multiple pages or sheets 100 , divided into five
sections 86 . These are all saved in one Prism .PZF or .PZFX file. Of course,
you can export individual data tables or graphs, but this is not the same
as saving the entire project as one Prism file.
People new to Prism often make the mistake of thinking that each Prism
file can only have one data table and one graph. Not so. Each file can
contain up to 500 data tables, info sheets, analysis results, graphs and
page layouts.
While working with Prism, you can open multiple Prism files. The details
are slightly different for Windows and Mac.
Prism Windows. Multiple open projects in one overall Prism window.
There are two ways to work on a second Prism project.
Use the Window menu to move from one project to another. Use the
Navigator or bottom toolbar to move between sheets in one project file.
Prism Windows. Two open projects in distinct Prism windows.
If you want to open a second Prism window, simply launch Prism again.
To do so, you must double-click on the program icon. Double-clicking on a
Prism file when Prism is running will always open that file in the instance
of Prism already running. To launch a second instance of Prism, double-
click on the program icon. It is not possible to launch more than two
instances of Prism. (Opening two distinct Prism windows is new to Prism
6.)
Once you have a second copy of Prism running, you can open any file
within it. If you open a file that is already open in the other copy
(instance) of Prism, then you won't be able to save any changes unless
you use File..Save As and choose a new file name. Any changes you make
in one instance of Prism will not be seen in the other. Use the second
launch of Prism to view one sheet (say a page of results) while you edit
another sheet (the graph) in the original launch of Prism.
The Windows menu of each instance of Prism will only show files open in
that instance.
Prism Mac before Sierra
There is no concept of a parent or overall Prism window when you use
Prism Mac. Each project file you open or create is in its own window. You
can move each project to whatever display you want. The Windows menu
lists all open Prism projects.
Prism Mac after Sierra
MacOS Sierra introduces the possibiltiy of putting multiple Prism files into
a single tabbed window. The files are distinct (not merged) but they all
will be in one window, with a row of tabs that let you choose which file
you want to view. This can make it easier to be organized, especially if
you work with a small display. Details 106 .
A new feature in MacOS Sierra lets you put several files into one window
and navigate between them using tabs. This is a great way to work on
multiple files at once, especially on a small screen. Apple calls this feature
"Tabbed Windows".
When you create or open multiple files with Prism Mac, it looks like this.
Each project file is in its own window with its own toolbar.
Until Apple released MacOS Sierra, this was the only way to work with
multiple files in Prism Mac. But Sierra lets you put them all into one
Window. Drop the Window menu on any of the Prism files, and choose
"Merge All Windows".
Now there is only one Prism Window, with a set of tabs just below the
toolbar that let you choose which project file you are working on:
Notes:
· To work on another window, click the + button to the right of the file
selection tabs which will bring up the Welcome dialog.
· There are three ways to move a file into its own window and away from
the tabbed window: Control-click on the tab and choose Move Tab to
New Window; grab the tab and drag it away from the joint window; or
drop the Window menu and choose Move Tab to New Window.
· There is no way to have one window with one set of files, and another
window with another set of files. Each file is either in its own window, or
in the single tabbed window.
· Keyboard shortcut: Hold Control and click Tab to move to the next
project file in the tabbed window. Hold shift and control and click tab to
go to the prior project file.
saved in different names. Putting multiple files into one tabbed window
does not merge the files.
· With Prism Mac, you can click the green circle the top left of the toolbar
to run full screen. When you use the View menu to view or hide the top
toolbar, the navigator or the bottom toolbar, those choices will apply
only to the full screen view. You can make different choices for when the
Prism window does not fill the screen.
· Hide the Prism Navigator tree. To hide the navigator, use a command in
the View menu or the leftmost button in the bottom toolbar.
· Hide the main (top) toolbar, and use the menu commands instead. To
hide the toolbar, use a command in the View menu.
· You can also hide the bottom toolbar, but we suggest you don't. If you
hide the navigator, you'll want to use the bottom toolbar to navigate
from sheet to sheet and section to section.
· With MacOS Sierra, you can put multiple Prism files into a single tabbed
window 106 .
6 Prism files
pieces of a project.
Versions and Autosave have been part of OS X since 2011, but are new to
GraphPad Prism 7. Apple thinks these features are so intuitive they don't
· As you work, your file is automatically saved. You don't have to think
about saving. Close a project file when you are done working with it.
Open it when you want to work on it again.
· When you quit Prism, it remembers the files that were open. When you
next launch Prism, those files will open again. This means you won't see
Prism's Welcome dialog unless you closed all open files before quitting.
· If you need to, you can revert a file back to how it was an hour (or a
day...) ago.
· The File..Save command has a new purpose after its first use to specify
a file name and location. After that, use the Save command to save a
milestone you can revert to.
· Two settings in the General tab of Systems preferences affect how your
work flows: whether to keep changes when closing a document, and
whether to open the documents you were working on when you start an
application. These apply to all applications that use the new way of
working, not just Prism. We suggest you keep both options checked.
Note that this page only applies to Prism Mac, not Prism Windows.
Versions and Autosave have been part of OS X since 2011, but are new to
GraphPad Prism. If you use iWork (Pages, Numbers and Keynote), you
are already familiar with these features. Since we know that many Prism
users will have not encountered this new way of working before, we detail
how these features work below. But note that we simply implemented the
framework that Apple created, so have no ability to fine-tune how it
works.
Two setting in the General tab of System Preferences have a huge impact
on how Prism acts when you close a file or exit Prism.
By default, both are unchecked. Both options apply to all applications that
use Autosave and Versions, not just Prism. Most of this page assumes
that you have left both options unchecked (the default), and we urge you
to use Prism for a while this way to get used to the new approach. What
happens when you check these options is explained at the bottom of this
page.
A new mindset
Prism 7 now conforms to the new paradigm that Apple created for
working with files. There are three new big ideas:
· As you work, your file is automatically saved. You don't have to think
about it. Close a project file when you are done working with it. Open it
when you want to work on it again. You'll always be working on the
latest version and don't have to think about saving.
· When you quit Prism, it remembers the files that were open. When you
next launch Prism, those files will open again. This means you won't see
Prism's Welcome dialog unless you closed all open files before quitting.
· If you need to, you can revert a file back to how it was an hour (or a
day...) ago.
Once you get used to it, this is a much better way to work.
The Save command has a new purpose
The File..Save command still exists, but its purpose has changed. When
creating a new project, use this command to specify its name and
location. After that, use this command when you have completed a chunk
of work, and want to save a milestone you can revert to (see below).
No File..Save as command
The File.. Save As was always annoying because it rarely did exactly what
you wanted. So Apple removed it and replaced it with three new
commands on the File menu:
· File..Duplicate. Make a copy of the file (which you can name), and then
edit the copy while leaving the original alone.
· File..Rename. Rename the file you are working on. You'll enter the new
name at the very top of the Prism window.
If you really want to use the old File..Save As command, there are two
ways you can:
· Hold down the Option key while dropping the File menu. Now the Save
as... command appears (instead of Duplicate).
· Click the Save button on the toolbar that drops a menu. That menu
includes Save As.
There is one purpose of Save As that is not available with the other
commands. If you have saved the file in .pzf format and want to switch to
.pzfx. Or the other way around.
Reverting files
A version of your file is saved when you open a file, when you choose the
save command, and occasionally while you work on it. The File.. Revert
To command lets you revert your file back to how it was when you opened
it, to how it was the last time you explicitly saved it, or to browse all
versions to find the one you want to restore.
When you ask to browse all versions, the Revert screen fills your display.
The right side shows a stack of older versions.. Use the buttons near the
bottom right to step through the versions. There is a bug (by Apple) so
sometimes these buttons appear gray/unavailable. Click them anyway --
they work. If you have lots of saved versions, use the time line on the
right side of the window to fast forward to the version you want.
· You are only able to browse prior versions so long as you work on a
single computer, as the information about prior versions is not saved in
the file itself, but rather is saved in hidden files on your computer.
· To delete a version or all old versions: Select that version on the right
side of the revert screen, move the cursor to the top of the screen to
see a menu, drop the File menu, choose Revert to, and you'll see the
new command "Delete this version". Hold Option and that command
changes to "Delete old versions" to delete all versions (except the
current one, of course).
· To save an older version while also saving the current version: Click on
the version you want, move the cursor to the top of the screen to see a
menu, drop the File menu, and choose Restore a copy. If you have
selected the current version (on the left), this command changes to
Duplicate.
· Instead of reverting, you can select data or text (but not graphs or
layouts) from an old version, copy to the clipboard, and then paste into
the current version. In some situations, this is more useful than
restoring an old version entirely.
Rename, tag and move from the top of the Prism Window
Hold the mouse cursor over the file name at the top of the Prism window,
and you'll see a downward pointing arrow. Click that arrow, and you can
rename the file, add tags, move it to a different folder, or lock it.
Notes:
· The full file name is shown, including the extension .pzf or .pzfx. You
cannot change the extension from here. You'd need to use the File..
Save as command (see above; you need to hold down the Option key).
· When you click in the Tags field, a list of current tags (including color
codes) will drop down. You can also tag files in Finder, and search there
for files by tags. A file can have more than one tag. You can also enter a
new tag. Tagging is an OS X feature, not a GraphPad feature.
As mentioned at the top of this page (with screen shot), two settings in
the General tab of the System Preferences dialog affect how Prism saves
files.
Ask to keep changes when closing documents
With the new way of working, Prism doesn't prompt you when you close
a named project with unsaved changes because the changes are
automatically saved. When you open that file again, you'll see the latest
version. If the project is new, and not yet named, Prism will prompt you
to name it or delete it when you close its window.
If you click Save, Prism saves your latest work before it closes the
document.
If you click Revert Changes, Prism reverts any automatic saves so the
document that gets saved is the same as it was the last time you
explicitly saved it.
It reminds you of the old way of working, but it doesn't reset things all
the way back. We recommend not using this option, as it isn't really
helpful.
Close windows when quitting an app
With the new way of working, when you launch Prism the projects that
were open when you exited Prism open automatically. You won't see the
Welcome dialog if any project windows were open when you exited Prism.
The idea is that you open a document when you want to work on it, and
close it when you are done (for now). Launching and quitting Prism has
nothing to do with with when project files are open.
If you check the setting in General tab of System Preferences (see top of
this page) "Close windows when quitting an app", then Prism will save
and close all the projects when you quit. If a project file is not yet named,
Prism will ask you to name it. When you next open Prism, you'll see the
Welcome dialog and no projects will open automatically.
This seems like a useful personal preference. Check this option if you
want to start fresh with the Welcome dialog each time you launch Prism.
Don't check it if you like the idea of that when you launch Prism, it opens
the files you left open the last time you used Prism.
Prism has had automatic backup since version 3. You won't notice any
difference in Prism 7, but now Prism uses the built in automatic backup
feature provided by Apple, instead of using our homegrown method. In
the File tab of Preferences, choose how often you want to save automatic
backups and whether you also want a backup made whenever you view a
different sheet. We suggest saving backups every few minutes and
whenever you view a new sheet.
Two formats
When you save a Prism file, you save the entire project into one file. You
have a choice of two file formats:
· The PZFX format is a bit more secure for long-term backing up. Even
without access to Prism, the data in the file are in plain text that can
readily be viewed.
· The PZFX format is very useful for data interchange. With a bit of work,
someone could write a program to replace the data in a .PZFX file with
new data. When you open that file, you'll see the new data analyzed and
graphed.
· Files saved in PZFX format tend to be much smaller (about 10% the
size) than the same file saved as PZF. We recommend that you use
PZFX files routinely.
· The data (and info and scripts) are in a plain text format at the
beginning of the file. It is possible to create a program that swaps out
the saved data with new data. Open that altered file, and the new data
will instantly be analyzed and graphed.
· Prism 5 was the first version that understand the PZFX format, so these
files cannot be opened by Prism 4.
· Avoid the .PZFX format if you plan to post the file on a web site. Web
browsers sometimes detect the XML format, and try to open the file
with some program that understands XML rather than opening with
Prism. In contrast, .PZF files generally just open in Prism, as they
should.
Choose to use, or not use, the compact format in the File tab of
Preferences. PZFX files are always saved in the compact format.
Zipping the file with standard zipping software can do a much better job
of compression. To really make a PZF file as small as possible, save it
compact and then zip it.
Check the box in the Preferences 153 menu if you want to save in compact
format as a default.
We have not tested opening Prism 7 files in Prism 5 and earlier. It will
probably work ok for many files, but this is not something we have
tested.
The location of analysis results may change between versions
When you open a Prism file on a version different than the version it was
created on, results also may appear on different rows of the results
tables. This means that results tables embedded on graphs may no longer
show the results you care about, and scripts may export the wrong
values. You may have to do a bit of clean up to make the files work as
intended.
Choose automatic backups on the File & Printer tab of the Preferences 153
dialog and Prism will automatically save your work. Choose to save based
on a timer (every few minutes) and/or whenever you switch from one
sheet to another. The backup files are also saved whenever you save your
work, so the backup files can never be older than your project files.
When you quit Prism normally, the timed backup files are deleted. In case
of computer disaster, the backup files will remain on your hard drive.
When you restart Prism, these files will automatically load so you can
continue your work. Note a potential problem: After a crash, the
automatic backup system works fine if you launch Prism as an application.
But if you double-click on the file you were working on, then you open
that saved file, not the backup that Prism saved elsewhere.
If you need to find these files, which have the extension .PZB, their
location depends on which operating system you use:
Note the difference between Save As and Save a Backup. When you use
the Save a Backup command, Prism doesn't rename the file you are
working on -- it simply saves a copy of that project with a new name.
We recommend that you backup to the .PZFX format (rather than .PZF)
because it provides a more secure backup. It is possible (with some
difficulty) to extract your data from these files without Prism, even if the
file is damaged or truncated.
If you share a computer with others and don't want your unsaved work to
be recoverable, you can turn this feature off on the File & Printer tab of
the Preferences dialog.
Manual backups of Prism data tables
You can backup individual data tables using the File Export command. You
can export to plain text tables (delimited with tabs or commas) or to an
XML format. We recommend using XML format, even though it is harder
to view, because it contains more information. XML format includes
details omitted from the plain text formats, including about subcolumns,
excluded values, and floating notes.
Merging
To merge an entire Prism project into another, drop the File menu and
choose Merge. Pick a file (which must not be open), and all sheets from
that file will be appended to your current project. Prism places an asterisk
(*) in front of the name of each merged sheet.
It is also possible to merge just one graph and its related sheets into
another Prism project. To do this, do not use the Merge command.
Instead, copy a graph from one Prism project and then paste it into a
Layout 478 in another project. The Place Graph on Layout dialog lets you
choose to merge the graph and data into the new file. The alternatives
are to maintain a link to the original Prism file so that changes in the
original will update in the current file, or to paste just a picture of the
graph.
Splitting
To save one family of sheets into a new file, choose Save Special from the
File section of the Prism toolbar, and then Save Family As. This will save
the current sheet along with all linked sheets.
To save the entire project with a new name, drop the File menu and
choose Save As.
Once you've saved a copy of the file, or part of it, you'll want to delete
unwanted sheets from the original project. The easiest way to delete
several sheets is to go to the Gallery for a section and click the checkbox
on several sheets. Then press Delete. When you delete a data table, you
have the option to delete all info, results, and graph sheets linked to the
selected data tables at the same time.
Click on the Send by Email button in the Send section of the Prism
toolbar. You can choose to send a Sheet, Family, or Project. Your email
program will open and create a new message with the Prism file attached.
The format of the file depends on what you choose:
You can send a Prism file even to someone who doesn't own Prism. The
recipient can view the file using a free Prism trial. After the trial expires in
30 days) it becomes a Viewer that doesn't expire. It lets anyone view and
print all contents of a Prism file, including analysis choices, but does not
allow any editing or exporting.
6.8 LabArchives
· From the Welcome dialog of Prism 6, you can open a file directly from
LabArchives. There is no need to first go to labarchives.com to download
the file you want.
If you have a LabArchives username and password, Prism will prompt you
to enter it.
But what if you log in to your institutional server, and the logging into
LabArchives is done behind the scenes using an authorization system like
Shibboleth? In this case, you don't have a username and password
specific to LabArchives. But you can get one. After logging into the
LabArchives web browser interface, click on your name in the upper right-
hand corner of the screen and click “LA App authentication". The next
screen will show you a username (email) and password for LabArchives.
Use these when Prism asks you to log in to LabArchives. Prism will
remember this information, so you should only need to do this once.
Saving and exporting are different. When you Save your Prism file you
save it in Prism's file format. This type of file can only be opened with
Prism and it contains data, analyses and graphs.
Prism can save your work in three special kind of files designed to let you
save time by not repeating steps with future data.
· A template is a Prism file that you recycle. Choose a template, add new
data, and all the analyses and graphs will be created automatically.
Templates 138 are an older feature, and we recommend you use example
files or method files instead.
method, you apply the analyses and graphs in the method file to the
data table you have already entered.
There are two ways to close a Prism file without exiting Prism:
Using Windows, distinguish the icon to close a file from the one that
exits Prism.
Using Mac, each open file is in its own window. Close it by clicking the
red circle in the upper left.
Why close Prism files without exiting Prism? Because the next file will
open more quickly if you don't also need to initialize Prism by relaunching
it.
With either platform, if you close the last open file, Prism still remains
open. On the Mac, you'll see it in the dock and its menu will still be active
so you can open another project or create one. On Windows, the Prism
window will remain open.
new data using the same steps and format that you
used earlier.
What is a family?
Prism remembers the links between data tables, info sheets, analyses,
graphs and layouts. Replace or edit the data, and linked analyses and
graphs will update. We call the current sheet with all linked sheets, a
family of sheets.
From anywhere in the program, drop the Family menu to see all the
commands that work with the entire family of sheets related to the sheet
you are now looking at.
The navigator 100 shows you which sheets are linked to the current sheet.
Their names are bold. If you have a large project, you may have to scroll
to see all the linked sheets. Prism, therefore, lets you see them all at
once.
Click the plus sign in front of the Family 100 folder to see a list of all sheets
linked to the current one. Alternatively, click on the word "family" itself,
to see all linked sheets displayed in a gallery 101 . When you move to
another sheet, the definition of family will change.
Click the Link button in the bottom toolbar to jump to any linked sheet.
Drop the File menu, and choose Save Special, and then Save Family
As. Choose a location and file name, and Prism saves just the current
family of sheets into that file. The current full file will remain open, and
Prism will open the new, smaller, file in a smaller Window. Be careful
not to mix these up.
To delete a family of sheets:
Click on the plus sign in front of the Family folder to open it. While
pointing at that folder, right click (control click for macs) and choose
Delete Entire Family from the shortcut menu.
To duplicate a family of sheets:
From any sheet, click the New button on the toolbar, and choose
Duplicate Family. Alternatively, right click on any sheet in the
navigator, and choose Duplicate Family.
Starting from a new data table, create an entire family of sheets with the new
data analyzed and graphed:
From a data table (not linked to any graphs or analyses), click the
Wand 136 (Wizard to ANalyze Data) button. Choose the table you want
to use as an example, and Prism will analyze and graph the current
data exactly as it analyzed the example data.
Merge a graph and its family into a different project:
Copy the graph, go to a layout of the other project, and paste. You'll
get a choice of pasting just the picture, a picture with a link to the first
project, or to merge in that family of sheets into the project that
contains the layout.
Possible ambiguities:
· A graph can contain data from two more more data tables. If you start
from the graph (and save or duplicate its family), both (all) those tables
are part of the family. If you start from only one of those tables, the
other table(s) are not part of the family.
· A layout can contain many graphs. If you start at the layout, the family
contains all those graphs and all the data and results on those graphs. If
you start from a data table, the family includes the layout, but not the
other graphs (and related sheets) that are also on that layout.
· If you start from a data table, all analyses are part of the family. But if
you go to a graph that includes that table, only analyses plotted on that
graph (directly, by pasting a table, or by using analysis constants) are
part of the family. Other analyses that are not actually on that graph are
not part of the family defined when you start on a graph.
· Don't confuse two uses of the word family. The term is also used when
Prism does multiple comparisons after one- or two-way ANOVA. The
multiple comparisons calculations are done so the probability alpha
applies to the entire family of comparisons. This use of the word family
in multiple comparisons has nothing to do with a family of sheets.
If you are starting from a data table, Prism will ask you to assign a name
to the new (duplicated) data table. If the table name is a subset of the
results and graph sheet names, these will be substituted. Suppose that
the data table is "October data", and is linked to "Nonlin fit of October
data" and "Graph of October data". If you start from the data table and
choose to duplicate family, you can enter "November data" as the name
of the new data table. Prism will then rename the linked sheets to "Nonlin
fit of November data" and "Graph of November data". It can only do this
because the exact name of the data table is part of the sheet names.
If you are starting from another kind of sheet, Prism will ask you to enter
a prefix that will be prepended to the names of all the duplicated sheets.
In the example above, you could choose to preface all the newly created
sheets with "* ". These new sheets will be "* October data", "* Nonlin fit
of October data" and "* Graph of October data". Of course, you can
rename the sheet titles however you want.
Prism will duplicate the current sheet and all linked sheets. Replace the
data on the new data table, and all analyses and graphs will update.
Cloning means to duplicate a graph along with data tables and results
included on that graph. Enter a new data set, and it will instantly be
graphed the same way. Clone a graph when starting a new project from
the Welcome dialog. Or drop the New button and choose Data Table (+
Graph).
Once you choose the graph to clone, Prism will show a dialog where you
can choose which data and titles to delete. By default, Y values are
deleted, but X values and titles are kept.
Q&A:
· Are analyses also cloned? It depends. Cloning duplicates regression
lines and curves,and results tables that are included on embedded
tables. Other analyses that are linked to the original data table, but are
not included on the graph, are not cloned. For example, say you have a
data table, analyzed with a t test, and graphed. If you clone the graph,
the t test analysis will only be included if you have pasted a portion of
the results onto that graph.
· To save the current graph as an example to clone in the future, drop the
File menu and choose Save Special ...Save as Example. Start from a
completed graph. Don't erase the data before saving the example, as
you can ask Prism to remove the data when you clone the example.
· Can you clone layouts? No. But you can go to any layout, and duplicate
that sheet with its family 134 .
7.4 Use Wand to analyze and graph just like another table
If you've already entered data onto an empty data table, it is too late to
clone, too late to duplicate a family, and too late to open a template -- at
least without copy and pasting data.
Prism provides a great tool to solve this problem: the Wand (Wizard to
ANalyze Data). Use it to analyze (and graph) the new table just like you
already analyzed (and graphed) an existing table.
1. Start from the data table (with data) you wish to analyze (and graph).
2. Click the magic wand button on the toolbar, to bring up the Analyze
Data dialog.
3. Choose the existing data table (in the current project) you want to use
as an example.
4. At the bottom of the dialog, optionally enter the prefix you wish to
appear in front of the new sheet names.
That's it. The new table (where you started from) will be graphed and
analyzed the same as the example table.
Notes:
· You can also choose to use the Wand to apply a stored method file. To
save a method file, go to the data table that has been analyzed and
graphed, drop the File menu and select Save Special… Save Method.
Where is it saved? 581 To apply that saved method file, click the Wand
button (or the Analyze button), and then choose Use Saved Method at
the top left of the Analyze Data dialog.
· The same icon is used for both the Magic 317 and Wand features. Magic
(Make Graphs Consistent) changes the appearance of selected graphs
so they look like an example graph. You can only use Magic when you
start from a graph, or a set of graphs selected in the gallery. The Wand
analyzes and graphs a new table the way you have already analyzed
and graphed another table. You can only use the Wand when you start
from a data table with data.
7.5 Templates
What is a template?
A template is a Prism file that you recycle. Choose a template, add new
data, and all the analyses and graphs will be created automatically.
Templates are an older feature, and we recommend you clone a graph 135 ,
or use the Wand 136 , instead.
· A template has the extension .pzt. When Prism opens such a file, it will
"forget" the file name and open the file without a name. Then it will
prompt for a file name when you save. This prevents accidentally
overwriting the template.
· When you save a template, you can enter some brief instructions that
will appear when the template is opened. This can be useful when you
are creating a template for others to use.
· A list of templates will appear on the Open a File tab of the Welcome
dialog.
Creating a template
1. Create a Prism project with only the sheets you want to see when a
template is opened.
2. (Optional) Delete the portions of the data that will change with each
experiment (for example, delete the Y values but leave the X values
and the column titles). Don't delete entire columns, just the values in
those columns by selecting and then using the Delete key.
3. Go to the data table you want to see when the template is opened,
drop the File menu, choose Save Special, and finally choose Save
Template. You'll also find this command if you drop the Save menu
from the toolbar.
To use a template
You'll see a list of templates on the Open a File tab of the Welcome dialog.
When you open a template, any description of instructions you entered
will appear in a popup dialog.
The instructions above assume that you save the template from Prism,
and open it via a link on the Welcome dialog. But you can rename any
Prism file to have a .pzt extension, and open it from Windows Explorer or
Mac Finder by double-clicking on it. The difference is that when Prism
opens a .pzt file, it instantly forgets the file name so will prompt you for a
file name when you first save it.
When you want to analyze and graphs a set of experiments the same
way, you can save your example as either a template or method. In
either case, start from the data table, drop the File menu, choose Save
Special..., and then choose Save Template or Save Method.
· Open the template from the Welcome dialog, enter or import data, and
then look at (and fine-tune) the results and graphs.
· Methods are different. First create a data table and enter or import
data. Only then do you click the Wand button to apply the method to
your data.
Perhaps this distinction will help: A template is a noun - a place you start
from. A method is a verb - something you apply to data you have already
entered
Simply drag-and-drop within the Navigator 100 or Gallery 144 , or pull down
the Edit menu and choose Reorder
Renaming a sheet
Prism initially assigns each data table a generic name such as “Table 1.”
It is easier to manage large projects if you give the sheets more
descriptive names.
Click on the name of the sheet in the Navigator, and press F2 (or wait for
a second and then click again) to rename the sheet. Or right-click on the
sheet name and choose Rename sheet.
You can also rename using the bottom toolbar. Click on the sheet name in
the bar at the bottom of the Prism window and edit it. Don’t try to delete
the current sheet name from the drop down in the bottom toolbar. If you
delete the name, Prism replaces it with a default name. Instead, double-
click on the sheet name in the bottom toolbar to select it, and then begin
typing to replace it.
Deleting sheets
Select a sheet in the Navigator 100 and click the Delete Sheet(s) button
on the Sheet toolbar. You can select one or more sheets in the current
section to delete.
To delete several sheets,go to the Gallery 144 and select several sheets.
Then click the Delete Sheet(s) button or press the Delete key.
Because sheets are linked, remember that deleting one sheet may affect
others. When you delete a data table or results sheet, you also remove
those data and results from all graphs. The graphs themselves remain;
they just have fewer (or no) data plotted on them. Any results sheets
based on the deleted data become orphaned (you can still view and graph
orphaned results but can’t change the calculation parameters). When you
delete a graph sheet, you also delete that graph from all page layouts,
but you will not delete the page layout sheets themselves.
Use the Zoom buttons on the bottom toolbar 100 to make the Gallery
thumbnails larger or smaller.
· Export 505 .
When you edit data, Prism automatically recalculates the analyses and
updates the graphs.
Click the Freeze button in the Sheet section of the Prism toolbar to freeze
a results sheet or graph. The Prism bottom toolbar will indicate that the
sheet has been frozen and won't update with any changes to the data.
Click again to unfreeze.
· Use Info sheets to record experimental details and constants you want
to use in analyses.
· Each project can have up to 500 info sheets. Each Info sheet can be
linked to a particular data table, but doesn't have to be.
To create an info sheet, click the New button on the toolbar, or drop the
Insert menu. Choose New Info. Select whether the new sheet should be
linked to a particular data table, or a global Info sheet for the project.
Select all or part of the Info table, and copy to the clipboard. Then go to a
graph or layout and paste. This will be a live link, updated if you edit the
info table. Double-click on the embedded table to adjust its properties.
While entering text, use the right-click menu or click on the Inset Info
Constant button in the Write section of the Prism toolbar. Prism will
maintain a link to the constant, updating the text when you edit the info
table.
Click on a row, then click the right mouse button. From the shortcut
menu, choose to move the selected row up or down, or choose to delete
that row. Or use the buttons on the Change toolbar.
Click this button in the Change toolbar to bring up a dialog to select which
data table the Info Sheet is linked to.
Use the Info Sheet tab of the Preferences 153 dialog. You can also choose
whether or not Prism will automatically create a linked Info sheet to
accompany every new data table you create
To import info constants into an Info sheet at the same time you import data:
You can "hook 267 " Info sheet constants as constraints in nonlinear
regression or as constants when transforming.
You can also Hook to Info sheet constants to create custom ticks 437 , set
axis limits 427 , etc.
You'll know when it’s possible to insert ("hook") an info constant, because
you'll see the fishhook icon next to the number entry field.
When you see this button in a dialog, click on it to bring up a list of all
Info and Analysis constants that can be hooked.
Floating notes will show on the sheet but not when a sheet is printed, or
exported or copied to another application, such as Word or PowerPoint.
Use different color notes for different things. For example, have one
colleague review all yellow notes, and another one review all the blue
notes.
- Click the Minimize button on a Floating Note to collapse the note. Or just double-
click on the top border of the note.
- Drop the Note menu using the button in the upper left corner of a note and use
the Create Link to add a hyperlink to a website or other document.
It is easy to color code selected cells in data or results tables. This would
be useful to highlight the key results, to mark surprising results you wish
to review further, or to mark suspicious data entries that you wish to
double check. To do so:
1. Place the insertion point in the single cell you wish to color code. Or
drag to select a range of cells.
2. Click the paint bucket icon in the Change area and choose a color. Or
drop the Change menu and choose "Cell Background Color".
Note the distinction between two related features. You can highlight
(color code) sheet names in the navigator, and you can also color code
the background of selected cells in data and results tables.
Select a sheet in the Navigator, or several sheets in the Gallery 144 , and
click on the Highlight button on the Sheet toolbar. Click on the highlight
button itself to use the same color as last time, or click the arrow to
choose a color.
8.9 Preferences
Overview
Use the Preferences dialog to set Prism's defaults for various settings.
Note that some preference settings are made elsewhere in Prism. The
import 241 (paste special) dialog, the export graph 506 , and the diagnostics
tab of nonlinear regression, for example, all have check-boxes to make
the current settings default for the future.
Also note that on Macs, there are two settings in the General tab of the
System preferences 111 dialog that have a big effect on how Prism works.
To open the Preferences dialog, drop the Edit menu (Windows) or Prism
menu (Mac) and choose Preferences.
Most of the settings are self explanatory. The rest of this page discusses
the settings that are not entirely self-explanatory.
· Compact format for PZF files. All this format does is leave out analysis
results (and instead recomputes them when you open the file). IT
makes the file a bit smaller, but PZFX files are way smaller.
· Do not read page size when opening files. Prism tries to track the page
size of your printer, so the margins are realistic on screen. With some
printers, Prism gets confused and you can check this option, which is
rarely needed.
· How to paste a graph from another project (Mac). Choose whether the
pasted image needs to be Windows compatible (which increases file
size, and on slow computers may noticeably delay pasting).
Set defaults for axes, error bars, symbols and lines, color schemes, and
graph and data table fonts. These defaults apply to newly created graphs,
but don't change existing graphs.
Internet tab (Windows only)
Most of the tab is a relic from old versions. It sets up a feature that lets
you send a graph to a web server via FTP. We can't imagine why anyone
still uses this. If you do, please let us know!
The bottom part is an option for Prism to automatically check for updates
and let you know if there is one.
A Prism project file can have up to 500 data tables, info tables, analyses,
graphs and layouts. Lump lots of data into one file? Or split into many
files? The choice is yours. If you prefer to create large projects, use these
tools to navigate:
Use the bottom toolbar
Ping pong between any two sheets. Click once to go back to the previous sheet. Click
again to return.
Pick from a menu of sheets linked to the current sheet, or use the five tabs to jump
between sections.
At the top of the Navigator tree, there is a section that organizes related
sheets as Families 130 . When you select a sheet in any section of the
Navigator, all the related sheets -- data tables, graphs, info sheets,
results, and layouts will be grouped and shown in the Family section.
The screenshot is for Windows. If you use a Mac, you'll see triangles (pointing up or down) rather
than + and - signs.
This folder is especially useful when working with a project with lots of
sheets. It lets you move between related sheets without scrolling the
Navigator. The keyboard shortcut to open the Families folder is Ctrl-L.
Use the link button on the bottom toolbar to move between sheets in a
family. Note below that each type of sheet is marked either (D) for data
table, (R) for results, and (G) for Graph.
Prism lets you chain analyses. For example, you can first transform data,
then normalize the transformed data, then fit a curve to the normalized
results, and finally run a normality test on the residuals from the curve
fit. To make it easier to view analysis chains, Prism can show an additional folder in
the Navigator 100 labeled Data with Results.
The default is to show the Data with Results folder whenever a results
sheet is analyzed. To change this default, click on the Prism section of the
Toolbar and select Preferences, and then click this option in the View tab.
Click Find from the Edit menu (or click on the binoculars icon in the
bottom toolbar, or use the control-F shortcut key) to search for sheets in
a project containing specific words or text. The results will appear in the
Search Results folder which you'll see in the Navigator. Prism searches for
the text you entered in titles, floating notes, sheet names, and more. It
only searches within the current file, and does not search within other
files that might include the search text.
There is no way to empty the search results folder. Just use the Search
command (or button) again to search for something else.
Navigator tool tips
If you've got lots of sheets in a project, and you're not sure which one
you are looking for, hover your mouse over any sheet name in the Prism
Navigator 100 to display a thumbnail of the sheet.
· Highlight 151 the sheet name in a color you choose. Use the highlight
button in the toolbar, or right click on the sheet in the navigator and
highlight from there.
· Move 143 them to the top of the list in the navigator. Just drag and drop.
If a file is large, consider breaking it into smaller projects. You can save
any family of sheets as a new project. Find the command "Save Family
As" on the File menu.
Merge Prism files
If you later want to merge two files, that is straightforward. Use the
Merge command on the File menu.
You can place a graph from one project onto a layout of another project.
There are two ways to do this:
· Copy the graph from one project, go to the layout of another project,
make sure no placeholders are selected, and paste. Prism will give you
three choices:to paste a picture, paste a hot link, or to paste merge, so
the entire family of the pasted graph is merged in.
9 Data tables
that you pick the right one for your data and
experimental design.
The format of a Prism data table determines what kinds of graphs you can
make and what kind of analyses you can perform. Choosing the right kind
of table for your data is really important. Take the time to get this right.
Don't try to choose the right kind of table for your data by thinking about
the kind of graph you want to make. That approach will end up confusing
you because there isn't a one to one match between table and graph
types. For example, you can make a bar graph from four kinds of
tables 176 , and a graph showing points (a scatter graph) from three kinds
of tables 179 . You need to choose a table based on the organization of the
data and the analyses you wish to perform.
You are not limited to one data table per file. Store an entire project in
one file, creating as many data tables as you need.
Prism can plot error bars from raw data. You don’t have to enter SD or
SEM values, although you can.
The format of a Prism data table 159 determines what kinds of graphs you
can make and what kind of analyses you can perform. Choosing a data
table format lets Prism create a data table that fits your data and makes
it easy to create the kind of graph you want and perform the appropriate
analyses. You can always change 189 the format of a data table.
Column tables 167 - Use column tables if your groups are defined by one
scheme, perhaps control vs. treated, or placebo vs. low-dose vs. high-
dose. Each column defines one group.
Contingency tables 171 - Contingency tables are used to tabulate the actual
number of subjects (or observations) that fall into the categories defined
by the rows and columns of a table.
Survival tables 172 - Survival tables are used to enter information for each
subject. Prism then computes percent survival at each time, and plots a
Kaplan-Meier survival plot (and also compares survival with the log-rank
and Gehan-Wilcoxon tests).
Parts of whole tables 174 - A Parts of whole table is used when it makes
sense to ask: What fraction of the total is each value? This table is often
used to make a pie chart. The example below shows the number of
students who received grades of A, B, C, etc. The sum of the five values
in the table (61) is the total number of students.
9.1.3 XY tables
· Linear regression
· Nonlinear regression
· Smooth curve
§ The first optional column is for row titles 198 to label individual
subjects or samples.
§ The next column ("X") is for X value. There is only one X column per
table. If you choose to enter X error values, then there will be two
subcolumns (one for X, one for the error value).
§ Blue italics values are excluded 203 from graphs and analyses.
§ An XY data table can contain 256 data set columns and as many as
256 sub-columns.
Error bars
Prism will compute and graph error bars 213 from replicate values placed in
side by side subcolumns as shown (triplicate values) in the table above.
Specify the number of subcolumns when you create a new XY table.
You can also choose to create subcolumns formatted for entry of error
values computed elsewhere.
Prism can also plot horizontal X error bars 212 . You must enter X error
values directly; Prism cannot compute them. You can only enter one
value for X errors, so X error bars are always symmetrical.
Each data table has a single column for X and up to 104 sets of Y values.
What should you do if you have different X values for each set of Y
values? There are two ways to solve this problem.
The best approach is to stagger the data entry. You don't have to start
entering data in the first row. This example shows three data sets with
different X values.
In a column data table, each column defines a group. These groups are
defined by one scheme, perhaps "control vs. treated." You can have more
than two groups in a one-way table, for example "placebo vs. low-dose vs.
high-dose.
Analyses performed with Column data
· t test (one-sample, paired and unpaired)
· Mann-Whitney
· Wilcoxon
· Kruskal-Wallis
· Friedman
· Bland-Altman
· ROC curves
· Prism does not use a separate column to enter the grouping variable.
Instead the groups are defined by columns.
· If data are not matched or paired, rows have no special meaning. Enter
data in any order you like.
· Blue italics values are “excluded” (ignored by analyses and graphs). Use
Exclude Values.
Error bars
For one-way tables, Prism will calculate error values and create error
bars automatically. Enter all the values for a data set in a single column.
· Two-way ANOVA
· Three-way ANOVA
Notes:
· Unlike other programs, Prism does not use separate columns to enter
grouping variables. Instead, groups are defined by rows and columns.
· The rows categorize the data by one grouping variable (men vs.
women).
· Blue italics values are “excluded” (ignored by analyses and graphs). Use
Edit… Exclude values.
· A grouped data table can contain 256 data set columns and as many as
256 sub-columns
Error bars
Prism will compute and graph error bars from replicate values placed in
side by side subcolumns as shown (duplicate values) in the table above.
Specify the number of subcolumns when you create a new table.
You can also choose to create subcolumns formatted for entry of error
values computed elsewhere.
· Chi-square test
The three rows represent three different treatments. The two columns
represent two alternative outcomes. Each value is the actual number of
subjects (in this case; in other cases, the values may represent number
of objects or events).
On a survival table, you enter information for each subject. Prism then
computes percent survival at each time, and plots a Kaplan-Meier
survival plot (and also compares survival with the log-rank and Gehan-
Wilcoxon tests).
· Kaplan-Meier
· Log-rank
· Wilcoxon-Gehan
Each row represents a distinct subject. (You can optionally use row titles
to identify the subjects.) Each column represents a treatment.
Notes:
· Enter elapsed time (usually in days) as X (Prism does not let you enter
dates).
· Enter ‘1’ for a subject if the event (death) occurred at the time entered
into X.
· Enter ‘0’ if the data are censored. Data are censored when you simply
don’t know what happened to the subject after that time, or do know
but can’t use the information.
Error bars
When Prism plots a survival curve, it can include SE or 95% CI error bars.
These are computed as part of the Kaplan-Meier method for creating the
survival curve. You can't enter error values directly -- they are computed
from all the data.
A Parts of whole table is used when it makes sense to ask: What fraction
of the total is each value? This table is often used to make a pie chart.
Analyses performed on parts of whole data
· Fraction of total
The example below shows the number of students who received grades of
A, B, C, etc. The sum of the five values in the table (61) is the total
number of students.
You can enter data onto as many columns as you want. But note that only
the values entered into column A will be automatically graphed. Use
New...Graph to plot other columns.
Click the new button from the Sheet section of the Prism toolbar (or
drop the Insert menu) and choose New Data Table and Graph.
The New Data Table and Graph dialog looks a lot like the Welcome dialog
(with fewer choices). Choose which of six kinds of data tables 162 you
want, and whether to start with an empty table or use sample data. If
you choose to create an XY or Grouped table and are not using sample
data, you'll also need to decide about subcolumn format.
Prism offers six distinct types of data tables. Four of these can be used to
create bar graphs, as detailed below. If you want to create a bar graph, it
is important to choose the kind of data table that matches your data.
Column graphs
Enter data for each group into its own column, stacking replicates (if
any). Each column then becomes a bar, and you'll be able to control the
appearance of each bar individually.
Note: If you want to create a graph like the one above, but you've already
calculated your mean and error values, start instead with a Grouped table
(see below) and use only the top row.
Grouped graphs
Contingency tables
A contingency table is like a grouped table but doesn't have subcolumns
for replicate values.
XY Bar graphs
Both the kinds of bar graphs mentioned above place the bars on the
graph in order, with equal spacing. In some cases, you want to position
each bar along the X axis with the position determined by an X coordinate
you enter. To make this graph, choose the Spikes graph in the New Graph
dialog.
Settings in the Format Graph dialog also let you make this graph.
Uncheck the option to "Show Symbols" and check the option to "Show
bars/spikes."
Prism offers six distinct types of data tables. Three of these can be used
to create scatter graphs.
XY scatter graphs
With an XY data 165 table, each X value can have several replicates for
every data set. When you make your graph, you can choose to show all
the replicates on your graph instead of error bars. Start with an XY table
if you want to show all your replicates for each X value.
If you want to compare groups and show every data point along with lines
for mean and error for each group, start with a column table. All the data
values for each group should be entered in a single column.
To make the graph above, start with a Column table 167 and enter all the
data points for each data set in a column. Prism will create error bars
from all the data points in each column.
If you instead want a graph that shows only the mean and error for each
data set, double click on any data point to open the Format Graph dialog
and choose "One symbol per column" and the type of error you want to
show.
If you want to create a scatter plot comparing groups by more than one variable, enter data on a
Grouped data table with side by side replicates.
Prism offers three ways to create graphs that combine points and bars.
XY
Create an XY graph and then tell Prism to plot a data set (or several) as
spikes or bars. Detailed instructions to create the graph below.
Grouped
Create a grouped bar graph, then tell Prism to plot one data set (or
several) as points. Detailed instructions for creating the graph below.
1. Place your insertion point in the first cell of the series. Alternatively
select the entire range that you want to contain the series.
2. Drop the Insert menu and choose Create Series... Or click the
corresponding button in the Change portion of the toolbar.
3. Enter the number of values you intend the series to have. If you
preselected a range of values, Prism will default to the number of rows
in that range. Also enter the first value, and the rule to increment other
values.
Notes:
· If you select an entire column before bringing up this dialog, the dialog
will default to creating a series with 500,000 values. This can take a
while. Adjust the value a reasonable number.
3. Click the Table Format button at the upper left of the table, or drop
Change and choose Format Data Table.
Notes:
· On older versions of Prism, you were able to define this series at the
time you created the table. With Prism 6, you can no longer do this. You
must first create the table, and then use Format Data Table to define
the series.
· When you define the X axis to be a series, Prism will delete any X value
you had entered into the X column without asking for confirmation.
· With this method, the entire X column becomes a nearly infinite series.
You cannot tell Prism where to stop the series.
You can have up to 500 data tables in your project. Each data table can
contain:
· Any number of rows (limited by RAM and hard disk space). We've tested
hundreds of thousands of rows.
If your table has more columns or subcolumns that Prism 5 can handle,
the table will simply not be included when you open the file in Prism 5.
The choice you make about how to format your data table is important.
If you don't choose the appropriate kind of data table, you won't be able
to make the kind of graph you want or perform the analyses you have in
mind.
For XY tables Prism uses subcolumns replicate data. If your table needs
more subcolumns you can always change the number of subcolumns
(replicates) in a data table. Or, if you entered already-calculated error
values but your table is labeled wrong (say you actually entered SD, but
the table is formatted for SEM, you can make the switch.
To change the format of a data table, Click on the Table format area in
the upper left of a data table.
If you want to change the type of error bars displayed on your graph,
you don't need to change the data table. Instead, double-click on any
data point on your graph to open the Format Graph 351 dialog and use the
drop-down menu to change the error-bar format.
From the Welcome or New Table dialogs, you can choose to format the X
column for entry of dates or elapsed time instead of X values.
If you forgot to format the table when you created it, use Change..Format
data table to format the X values for dates or times.
If you chose to enter dates, you can enter dates in almost any format.
Prism will figure out what you meant and display the date on the table in
a standard format (that you cannot change). The only ambiguity is if you
enter a date like 1/2/2014 or 1-2-2014. Is that January 2 or February 1?
The convention depends on where in the world you live. Make your choice
on the bottom of the View tab of the Preference dialog.
If you enter a year with two digits, Prism Windows follows the choices
made in the Date tab of the Additional Settings dialog, accessed from the
Format Tab of the Regional settings control panel.
When Prism shows a ? in the data table
If Prism doesn't understand the date or time you entered, it will show ? in
the data table. This is treated as a missing or blank value for graphing
and analyses. When you click on the cell to edit, Prism will show you
exactly what you entered for editing.
Converting dates or times to values for analyses
When you analyze these data and make some kinds of graph, Prism
converts the dates or elapsed times to values. Choose the units and (for
dates) starting date on the Format Data Table dialog 189 . Note that the
choice of months is not offered, because months have different lengths.
Years also have different lengths, so if you choose years, Prism does the
conversion based on 365.25 days per year.
The Format Axes dialog will present several choices for numbering
format.
Click on the column title and edit. You can bold, italicize, and underline
text in titles and can insert superscripts, subscripts and Greek letters.
To add an additional line of text to the title click the Line Break symbol or
press the key combination Shift-Enter to move to the next line.
Autocomplete
Prism uses autocomplete to help you. It tries to guess what you are
typing, and let's you choose from a list of titles you've used before. If you
find this more annoying than helpful, turn it off in the View tab of the
Preferences dialog.
Prism 7 lets you you change these labels. Use Change... Format data
table and go to the third tab. You'll see the subcolumn tab of the Format
Data Table dialog.
If you check the option on top of the dialog, the titles you enter will
appear on the data table.
Notes:
· Choose between entering one title for every subcolumn for every data
set column (as above) or just a single set of subcolumn titles that apply
to all data sets. If you chose the latter choice with the example above,
you'd enter only three subcolumn titles and these would be repeated for
each data set.
· Controls are provided in the toolbar of the dialog to enter greek letters
and math symbols (etc.) and to enter super and subscripts.
· If the titles you enter are very long, you may wish to uncheck the top
option. This will leave the table labels as "Y1", "Y2", etc.
To hide the row title column, click the X at the top of the Row title
column.
You may enter two or more lines of text for each row title. This is new to
Prism 6. You may also now bold, italicize and underline text in titles (you
could always use superscript, subscripts and Greek).
To add an additional line of text to the title click the Line Break symbol
or press the key combination Shift-Enter to move to the next line.
Autocomplete
Prism uses autocomplete to help you. It tries to guess what you are
typing, and let's you choose from a list of titles you've used before. If
you find this more annoying than helpful, turn it off in the View tab of
the Preferences dialog.
Reordering XY data
If you choose connecting lines on an XY graph, Prism connects the points
in the order they appear on the data table. If the data are not sorted on
the data table, the connecting lines will jump back and forth across the
graph. If you want the connecting lines to go from left to right across the
graph, first sort the data table by X values. Click on the Sort Rows
button on the Change toolbar, and sort by X value.
This example shows the effect of reversing the order of the rows in the
data table.
This example shows the effect of reversing the plotting order of data
sets. Note that this example starts with the same graph as the prior
example.
Decimal format
Notes:
· When you place the insertion point (click the mouse cursor) in a
particular cell in a data table, Prism expands that value to show all
digits. When you move to another cell, the value reverts back to the
selected numerical format.
· It is not possible to change decimal format of a contingency table or
the Y values in survival tables, since all values must be integers.
· You can only change numerical format for an entire column (or
several columns). It is not possible to change the numerical format
of individual cells.
Decimal separator
In some parts of the world, a period (point) is used as the decimal
separator. In other parts of the world, a comma is used.
When entering data into Prism, you can use either a period or a comma
to mark the decimal point.
When exporting data, Prism uses the separator you choose in the Export
dialog. When pasting data, Prism uses the separator set in
Preferences 153 .
When importing data 241 , you can specify how a comma is interpreted (as
a decimal separator, as a thousands separator, or to separate columns
of data).
If a value is too high or too low to be believable, you can exclude it.
Excluded values are shown in blue italics on the data table, but are not
included in analyses and are not shown on graphs. From the point of
view of analyses and graphs, it is the same as if you had deleted the
value, but the number remains on the data table to document its value.
To exclude selected values, Click the Exclude button from the Change
toolbar.
Choose on the File & Printer tab of the Preferences dialog how excluded
values should appear when copied to the clipboard. Your choices are to
copy a blank value (as if missing), a regular number (ignoring the fact
that the value is excluded) or a value followed by an asterisk (to match
what you see). Prism offers the same choices on the Export dialog to
control how excluded values are exported to .TXT or .CSV files.
Highlighting values
Select a cell or a range of cells. Then click the paint bucket icon in the
Change part of the toolbar, or drop Change and choose Cell Background
Color. You can choose from eight colors to mark values that you want to
double check, or to color code different aspects of an experiment.
To remove a data set from a Results sheet, click on the Analyze different
data sets button from the Change toolbar, and uncheck the data sets you
do not wish to include in the analysis.
When entering data, simply leave a blank spot for any value that is
missing. Excluded values 203 are treated exactly the same as missing
values.
Prism never ever treats an empty cell as if you had entered zero -- it
always knows that is a missing value. It will analyze the data if it can, and
leave analysis results blank when it cannot.
The details of how Prism handles missing values differs for various
statistical tests.
Survival curves
Comparison of survival curves does not require equal sample size. If
data are completely missing for any subject, simply don't enter data for
that subject. But before deciding to leave data out, read about
censoring which happens when you know the subject survived up until
a certain time, but don't know what happened after that (or you know,
but can't use the data because the experimental protocol wasn't
followed). Prism handles censored data fine. Don't omit those subjects,
enter the duration that they survived on the experimental protocol and
mark that duration as censored.
· The Tab key moves to the right. Shift-tab moves to the left.
· The Enter key moves to the right to the next subcolumn, if there is one;
otherwise it moves down a row and to the first subcolumn. Shift-Enter
does the opposite.
· The DEL key erases all selected values, but the data set will still be
linked to analyses and graphs. When you enter new data, Prism will
update the analyses and graphs. To delete a block of data completely,
select Delete from the Edit menu. Other data on the table will move to
fill the gap.
9.3.13 Rounding
How many digits can Prism store?
Prism 7 now stores numbers with double precision, so it can store about
fourteen significant figures. If you enter the value 99.2492427 into a
data table, Prism 7 stores it as 99.2492427
By default, Prism tries to show all the digits you entered, and increases
the number of digits after the decimal place to accommodate the extra
digits. But you can use the Format Numbers dialog 202 to show fewer
digits. And the results of analyses can be rounded as well.
What you see on the Prism data table matters only when you look at that
table. Notes:
· Choosing to show fewer digits on the data table won't affect how Prism
analyzes or graphs your data. Analyses are always based on all the
digits you entered (up to the number Prism can store).
· When you save the Prism file (as .PZF or .PZFX), Prism saves all the
digits you entered (up to fourteen), so the values are not truncated
when you save.
· When you copy data from a Prism data table to the clipboard, Prism
copies the values you entered, not only the digits you see on the table.
· When you export a data table as a .TXT or .CSV file, Prism exports the
digits you entered, and not only those that appear on the table.
Rounding
There are two issues that can make rounding seem erratic.
The Format Decimal Format dialog rounds down what you see on the
table. But when you go to edit the data, all the digits you entered (or
imported or pasted) are still present. Prism doesn't yet offer a way to
truly round the data you see on the data table. But you can create a user-
defined function that will create a results table that is rounded.
The transform multiplies the Y value by 10^2 or 100, adds 0.5, and then
rounds down to the lower integer (that is what the Floor function does).
Then divide by 100.
The results will appear on a results page. They will actually show one
more digit than you want and that last digit will always be zero. You can
use Format Decimal format to show one fewer decimal place.
These guided examples of common analyses will get you off to a great
start! CLICK HERE >
The example above has a red border, because the data were copied and
then paste-linked from a red results table. If the values were pasted from
a green results table, the border would be green. Green results tables are
organized like data tables and (unlike red results tables) can be graphed
or further analyzed. If the data were paste linked from another data
table, the border is blue. If the data are linked to an external file, the
border is black.
formatted.
When you collect scientific data, you almost always collect replicate data
and plot the individual replicates or an error bar showing the variability of
the data. One of Prism's strengths is its ability to handle replicates and
error bars.
· If you enter replicate values, Prism can plot either individual replicates
or error bars.
· You do not need to enter SD or SEM values in order to plot them. Prism
can plot error bars from raw data.
· You do not need to run a Row Means analysis to plot error bars. Prism
can plot error bars directly from raw data.
See also:
Key concepts: SD
Key concepts: SEM
SD and SEM are not the same
When to plot the SD vs. SEM
The example below is set for entry of two replicate values (duplicates) but
you can choose any number from 2 to 256.
This next example is set up to enter mean, SD and N for each point.
From the Welcome (or New Table) dialog, you can choose to enter X error
values.
Note that you cannot enter replicate X values and ask Prism to compute
error bars, nor can you separately define the right and left error values.
You only can enter an error value (which could be a SD or SEM) and Prism
will use that value to plot symmetrical horizontal error bars.
If you enter error values computed elsewhere, these are plotted on the
graph.
When creating a Column table, you have no choices for data entry. Learn
more about Column tables. 167
If you want to create a column graph but enter error values computed
elsewhere, you can't do so on a Column data table. Instead, create a
Grouped 216 table, but enter data only into the first row. A Grouped graph
from data entered on one row looks like a Column graph.
The changes you make on the Change Type of Graph dialog apply to all
data sets on the graph.
When you create a grouped graph, the Plot drop down menu gives you
many choices for error bars. Return to this dialog using Change..Graph
Type.
If you choose to enter error values directly, Prism will plot the values you
entered (except that when you enter %CV, Prism plots the SD). If you
entered Mean, SD (or SEM) and N, you can choose (on the graph) to plot
the error bars as SD, SEM or 95% CI.
The changes you make on the Change Type of Graph dialog apply to all
data sets on the graph.
If you plot the 95% confidence interval as confidence bands, you can fill
or color the area between the bands.
When you fit a line with linear regression, or a curve with nonlinear
regression, you can choose to plot confidence or prediction bands.
Confidence bands tell you how precisely you have determined the line or
curve. Prediction bands tell you about the scatter of the data.
Learn more about choosing and graphing prediction and confidence bands
from linear regression and nonlinear regression.
If you only want to make a graph showing mean and SD or SEM, you only
need to enter those values into the data table. It is ok to omit the "n"
values, or to format the table for entry of Mean with SD or SEM, but
without n.
· You want to be able to switch between graphing SD, SEM and CI error
bars. If you enter SD or SEM with n, Prism can switch between plotting
whatever form of error value you want to show.
· You want accurate nonlinear regression. If you enter the data as mean
and SD or SEM, Prism will fit the means, and ignore the values you
enter as SD or SEM. If you also enter n, Prism can account for scatter
and sample size, and the curve fit will be the same as if you had entered
raw data.
The difference between entering +/- errors and upper/lower limit errors
· When you format the subcolumns for entry of +/- error values, the
values you enter are interpreted as distances. These are added to (or
subtracted from) the value you enter as the mean to compute the end
point of the error bars.
What if I want to enter the median and quartiles, or some other kind of
error bars?
If you choose to enter the mean, with sample size (n) and SD, SEM or %
CV, you really ought to enter those exact values. Otherwise, analyses
might be incorrect.
If you choose to format the data table with subcolumns for Mean and SD
or SEM or %CV without sample size, or for mean with +- error values or
upper and lower limits, no analyses will be possible. Or rather Prism will
only analyze/fit the means and ignore the error values. If you choose to
format the table so the subcolumns are labeled mean, upper limit, and
lower limit, you are welcome to enter into those subcolumns the median
plus the 25th and 75th percentiles, or whatever values you want. Prism
will plot the error values you enter, and it is up to you to keep track of
how these values were actually calculated, and to label your graphs
accordingly.
XY graphs
Choose Individual Replicates when you pick the error format for your
graph.
Column graphs
Grouped graphs
Choose the Plot individual values tab. Then choose any of the graph
types.
In the Welcome dialog, or using Format Data Table, you can ask Prism to
create a subcolumn for X error bars.
Prism will then plot each point with horizontal X error bars sticking in
each direction a distance defined by the X error value you entered.
A choice on the third tab of Format Graph, lets you flip any Column or
Grouped graph 318 so the Y axis is horizontal. If you plot symbols or bars
with error bars, those error bars will now be horizontal, but that
horizontal axis is now the Y axis.
Depth charts
import.
When you copy and paste from an Excel file or text file into a Prism data
table, you have the option to paste only the values or to keep a live link
to the original file so that whenever the original file is changed and saved,
your Prism graphs and analyses will update. A third option is to embed
(keep a copy) of the original file inside your Prism file.
When you link, you keep a connection to the original file. When you
embed, you make a new copy.
Linking and embedding works differently with Prism for Windows and
Prism for Mac:
· With Prism for Windows you can link or embed to Excel and text files.
· With Prism for Mac, you can't embed files or link directly to Excel
spreadsheets. You can, however, link to text files.
Double-click anywhere on the data object to open the Import and Paste
Special Choices dialog. Go to the Placement tab. Find the spot where you
can change the top-left corner of the object. Enter a row number and
choose a column for the new location.
It is not possible to move a data object to a different data table. You can
delete it from the table where it doesn't belong. Then go back to the
source data, and copy-paste special to the correct location.
You will see a list of each linked data range on the current Prism table,
with the region you selected already highlighted. Press Change Source,
to bring up a second dialog.
At the bottom of that dialog, you'll see the current linked range in the
field labeled Item Name. With Excel, this range is formatted in an
unfamiliar way. R1C3:R100C4 goes from the first row of column 3 to the
100th row of column 4.
Note that you can also use this dialog to link to a different Excel file.
within an object, click on a cell near the object, but outside of it. Then
use the arrow keys to move the selection inside the object to the desired
cell.
If the linked text or Excel file is in a folder different than the Prism file:
Prism stores the absolute location of the linked file. If the file no longer
exists at that location, Prism won't be able to find it, so will not be able
to update the data table. However, the imported values remain in the
Prism file. You won’t lose any data if the link is broken.
If the linked text or Exel file is in the same folder as the Prism file: It
matters whether the Prism file is stored in the .PZF or .PZFX format. If
.PZFX, the link is absolute. If the folder has a new name, or has been
moved to a new location, the link is lost. If the Prism file is in the .PZF
format, then Prism will maintain the link to the Excel or text file that is
in the same folder as the Prism file. It doesn't matter if the folder has
been renamed or has been moved to a new location.
The pasted results are linked, which means if the original data are edited
or replaced, the linked values will update too.
Example 1
You use nonlinear regression to fit a dose response curve to six data sets.
The first three (A-C) are for one drug; the other three are for another (D-
F). The logIC50 values are all on one row. You know tha the first three
are for one drug, and the next three for another. But Prism doesn't "know"
this. The results look like this:
You want to put the logIC50 values into a new data table. The first three
should go into column A, and the next three into column B. Follow these
steps:
4. Put the insertion point into Column A and Row 1 of that new table.
5. Choose Paste Special from the Edit (or right-click shortcut) menu.
6. On the first (Source) tab of the Import and Paste Special dialog,
choose to insert and maintain link.
8. Click OK.
9. Note that the values pasted into the data table are surrounded by a
red box. You cannot edit these values on the data table. If you double
click within that box, you'll return to the Import and Paste Special
dialog (where you can unlink the values, if you want to).
10. Prove to yourself that the values are linked. Change the data being
analyzed, and notice that the summary values that you paste-linked
automatically update.
Example 2
You have two sets of column data that you wish to plot both as a bar
graph showing the mean and SD, and also has a scatter graph showing
individual values.
Another approach is to put the bars and symbols side by side. Follow
these steps.
When a block of data on a Prism data table is linked to another file (or to
embedded data), the data in Prism is included in a box. The box is black if
the linked data came from outside of Prism, red or green if the data were
paste linked from a Prism results table, and blue if the data were paste
linked from a Prism data table.
You cannot edit these values in the data table because they are linked to
another source.
To unlink the data block, and allow editing, right click anywhere within
the block of data, choose the Data Object command and then choose
Unlink.
Select a range of Excel data and copy it to the clipboard. Switch to Prism,
move the insertion point to the cell of a data table that will become the
upper-left corner of the pasted data.
If you use the Paste button or the keyboard shortcut (Ctrl+V), you will
only paste the values that have been copied to the clipboard. If you
choose Paste Special , you have several options about how you want to
insert your data into Prism's data table. These options are only available
with Windows. If you use a Mac, skip to this page 235 .
· Paste Data -- Prism retains no link back to the Excel spreadsheet. The
values will be pasted and nothing else. The only advantage to using this
method is that it is conceptually simpler.
· Paste Embed -- You paste the selected data into the data table and also
paste a copy of the entire spreadsheet file into your Prism project, even
if it has several worksheets (tabs). This means you don’t need to worry
about keeping track of multiple files. You don't need to save the
spreadsheet file separately (except as a backup). You can open Excel
from within Prism, edit the data, and instantly update the analyses and
graphs in Prism. A potential disadvantage of embedding is that you will
end up with multiple copies of the same data. Another disadvantage is
that your Prism files will be larger.
· Paste Link -- Paste the values into the data table but also create a link
back to the Excel file. The link has two functions. It lets you trace (and
document) the source of the data so you stay organized. It also is a live
link. If you edit or replace the data within Excel, Prism will update the
analyses and graphs. The Paste Link command is available only if you
paste from an Excel workbook that has a name. If you are working with
a new Excel file, you must first save it (thus giving it a name) before
you can link its data into Prism. When in doubt, use Paste Link.
You can copy a portion of an Excel spreadsheet and paste it onto a Prism
graph 339 or layout 478 . This pastes a picture only, and Prism cannot analyze
or graph the values since the values are not in any Prism data tables.
There are some limitations, and we don't recommend it, but you can
import 239 Excel files directly into Prism for Windows.
Select a range of Excel data and copy it to the clipboard. Switch to Prism
and move the insertion point to the cell of the data table that will become
the upper-left corner of the pasted data.
· Paste -- You will only paste the values that have been copied to the
clipboard.
· Paste Special -- You'll then see the Import and Paste Special 241 dialog,
where you can choose whether to filter and rearrange the data (such as
transposing).
Note that it is not possible to Paste Embed or Paste Link to Excel files on
the Macintosh. Those features only work on Windows.
Pasting an Excel table as a picture
You can copy a portion of an Excel spreadsheet and paste it onto a Prism
graph 339 or layout 478 . This pastes a picture only, and Prism cannot analyze
or graph the values since the values are not in any Prism data tables.
Prism Mac cannot import Excel files. Either copy and paste, or save from Excel in CSV
format and import that text file 238 into Prism.
Excel Windows 2003 and later, and Excel Mac 2008 and later, copy data
to the clipboard in two formats: Plain text, and HTML.
By default, Prism (starting with 5.0c and 5.03) pastes the HTML format.
This will allow much better pasting of Greek letters, superscripts and
subscripts, titles with punctuation, numbers with nonstandard thousands
or decimal separators, and more.
There is one change. If the Excel cell contains a mixture of numbers and
letters (or punctuation), prior versions tried to extract and paste just the
numbers. Prism Mac still works that way, but Prism Windows won't paste
anything into numerical parts of a Prism table if the corresponding Excel
cell mixes numbers and text.
This new style of pasting only applies when pasting values. Paste-
embedding and paste-linking (Windows only) are not changed.
Pasting from the HTML rendition is the default. If it causes any problems,
use Edit..Paste Special and check the new option on the first tab:
With this option selected, pasting will work just as it did with earlier
releases of Prism (before 5.03 and 5.0c). We don't anticipate any
situations where this will be helpful, but are providing this option just in
case. The option is available when you insert data only, and will be
unavailable if you choose to paste link or paste embed.
Numbers, a Mac spreadsheet that is part of iWork, does not copy the
HTML so pasting from Numbers always uses the older style.
Two ways to paste from the HTML (older Excel versions only)
It gets even more complicated.
The HTML clipboard version from some versions of Excel contains each
number twice. One copy shows exactly what you see in Excel. The other
copy shows the value with more decimal places (ignoring formatting or
rounding within Excel). By default, Prism pastes the values as they are
shown in Excel. But the Paste Special dialog lets you use the other
version if you prefer:
This option is not available if you copy from Excel Windows 2007 or later,
as the necessary information is not copied to the clipboard.
From the Welcome dialog (or the File Open command), you can only open
Prism project files.
To import data from a file, you must first create a new Prism project (or
open an existing one), go to a data table, position the insertion point, and
then choose the Import command or click the Import button.
Prism will show you the Import and Paste Special 241 dialog, which will let
you choose whether to embed or link to the original file (Windows only),
and also will let you filter and rearrange the data as you bring it into
Prism.
Formats Prism can import
Prism can import text files delimited with tabs (.txt files) or commas (.csv
files).
When thinking about whether Prism can import your files, distinguish the
format of the file from the arrangement of the data. Prism arranges data
differently than do most programs that can export .csv or .txt files.
To create a text file from an Excel spreadsheet:
Go to the worksheet (tab) you want to export, then pull down Excel's File
menu and choose Save as. Select either Text (tab delimited) or CSV
(comma delimited) format.
To import data into Prism (Windows only) directly from an Excel XLS file:
Start from a data table and move insertion point to the cell that will
become the upper-left corner of the imported data. Click on the Import
button on the Prism toolbar. Whenever possible, transfer data from Excel
using copy and paste.
Prism will show you the Import and Paste Special 241 dialog, which will let
you choose whether to embed or link to the original file, and also will let
you filter and rearrange the data as you bring it into Prism.
Limitations to importing Excel files directly
When importing data from an Excel file, Prism launches Excel invisibly,
and asks Excel to send the data to Prism. So Prism can only import an
Excel file if Excel is available on the same computer as Prism. Because
Excel has to be launched in the background, importing an Excel file can
take much longer than importing a text file.
You'll rarely want to import an entire Excel spreadsheet into Prism. While
you can restrict rows and columns to import on Prism's Import dialog, in
most cases it is easier to simply copy and paste the appropriate range.
Importing text files
Start from a data table and move insertion point to the cell in the data
table that will become the upper-left corner of the imported data. Click on
the Import button on the Prism toolbar.
Prism will show you the Import and Paste Special 241 dialog, which will let
you choose whether to embed or link to the original file (Windows only),
and also will let you filter and rearrange the data as you bring it into
Prism.
Note that it is not sufficient that a file have an extension (file type) such
as txt or dat. The file has to be a plain text file. If a file is not a text file,
renaming it to have a different extension won’t help.
The Import and Paste Special Choices dialog comes up when you import
a file into a Prism data table, or use the Paste Special command. It has
five tabs:
Source 242 - The Source tab lets you choose between importing or pasting just the
View 244 - The View tab lets you see the contents of the imported or pasted file,
Filter 245 - The Filter tab let you choose which portions of the data file to import.
Placement 249 - The Placement tab lets you rearrange your data as you
Info & Notes 250 - The Info & Notes tab gives you tools to import
sections of a text file directly into a Prism Info sheet 147 .
you must first save it (thus giving it a name) before you can link its data
into Prism.
Embed as OLE object (Windows only)
Paste or Import the selected data into the data table and also paste a
copy of the entire original spreadsheet or text file into your Prism project,
even if it has several worksheets (tabs). This means you don’t need to
worry about keeping track of multiple files. You don't need to save the
spreadsheet file separately (except as a backup). You can open Excel
from within Prism, edit the data, and instantly update the analyses and
graphs in Prism. A potential disadvantage of embedding is that you will
end up with multiple copies of the same data. Another disadvantage is
that your Prism files will be larger.
Automatically update Prism when the data file is changed
If you check this option, whenever you look at a Prism data table, graph,
results sheet, or layout, Prism will update the sheet if the linked Excel file
has been changed.
Excel options
When there is a choice, Prism (starting with 5.0c and 5.03) pastes the
HTML format by default. This allows much better pasting of Greek letters,
superscripts and subscripts, titles with punctuation, numbers with
nonstandard thousands or decimal separators, and more. You can specify
that the text format be used by checking the option:
This option is not available if you copy from Excel 2007 or Excel 2010, as
the necessary information is not copied to the clipboard. It is also not
available when importing from Excel.
Note a bug in Excel Mac 2008. In very rare cases, using that option will
result in the wrong values being pasted into Prism.
The View tab of the Import and Paste Special dialog lets you see the
contents of the imported or pasted file, divided into columns. This can be
especially helpful if you are importing from a comma- or tab-delimited
text file. A quick look can let you know if your columns are formatted
correctly.
Once you've pasted linked data into a Prism data table, click Open File to
open and edit the data file. If it is an Excel file, you'll open Excel. If it is a
text file, you will open a text editor.
Choices on the Filter tab of the Import and Paste Special dialog let you
choose which portions of the data file to import.
When entering data, you can leave any parts of the data table empty.
Prism automatically figures out what to do with missing values. When
Prism imports a text file, it handles missing values automatically. Some
other programs use a code, say 99, to denote missing values. If you
import data from such a program, enter that code value into the Filter
tab.
First, select the range of rows to import, by specifying a first and last
row. If the file you are importing includes <notes>, <info>, or <title>
information (see Paste Special Notes 250 ), start counting rows at the first
row after this structured information. Row “1” is the first row with data,
not the first row in the file.
Then, select rows you want to skip. The filters work by checking whether
a designated column meets a criterion you specify. You can apply any
combination of three kinds of filters:
When defining the criterion, you can use any column in the original file.
You are not restricted to using columns you choose to import. Compare
by checking whether the value of each row in the column is less than or
equal to (<=), less than (<), equal to (=), greater than (>), greater than
or equal to (>=) or not equal (<>) to the value you enter. It is not
possible to define criteria that compare values in two columns, only to
compare the value in one column with a value you enter in the dialog.
Decimate
If your data file is huge, you can decimate the data as you import.
"Decimate" literally means to keep one of every ten rows of data, but you
can tell Prism to skip any number of rows between the rows it imports. It
imports a row, skips the number of rows you specify, and then reads
another row. This is useful with any large file as a way to make Prism files
smaller and accelerate analyses and graphing.
Columns
Choose a range of columns you wish to copy by specifying a first and last
column. Or choose to read a column, skip one or more columns, and then
read another column.
Unstack
Example: In this sample of an indexed data file, you may want to import
only the data in column 2 and use the values in column 3 to define the
two groups.
Row # Col. 1 Col. 2 Col. 3
1 12 123 5
2 14 142 6
3 13 152 5
4 12 116 6
5 11 125 6
6 15 134 5
Check the box to unstack data only from column 2 using column 3 to
identify the groups. Prism will automatically rearrange the data, so they
look this like:
Row # First Col. Second Col.
1 123 142
2 152 116
3 134 125
The column that identifies the groups must contain integers. The lowest
integer defines the group whose values will be placed in the column
containing the insertion point in Prism. If there are gaps in group
numbers, Prism will leave blank columns.
The Placement tab of the Import and Paste Special dialog lets you
rearrange your data as you import/paste into Prism.
Names
The Placement tab lets you choose whether you want to rename the Prism data
table and/or the Prism column titles with the name of the imported file, or with text
imported from a specified row in that file.
Top-left position of inserted data within Prism
Specify the top-left corner of the data object in Prism. Normally, this will be the
position of the insertion point when you choose the Import or Paste Special
command, but you can change it.
Row and column arrangement
Prism can rearrange the data as it imports. If you choose Transpose, the first row
in the source will become the first column in Prism, the second row in the source
becomes the second column in Prism, and so on.
If the data source has one column (or row) and you want to organize the data
according to your experimental design, choose By rows or By columns.
Example: Your data source has thirty values in a single column, and you wish to
import into a Prism data table formatted for entry of triplicate Y values. The first
three values in the source data are the first three replicates. The next three values
in the source are the next set of replicates, so they belong in row 2 of data set A.
After filling five rows, you want to start inserting data into data set B. Choose to
import by rows, placing three values on each row and starting a new column after
five rows.
If the data source has the values from 1 to 30 in order, Prism will organize the data
as shown below. When placing values side-by-side, Prism does not pay attention to
the difference between subcolumns and data sets. It just puts values in the next
cell to the right, and it doesn't care whether that cell is another replicate, a SD, or
the beginning of a new data set. In the example above, Prism placed three values
side-by-side because you entered 3 into the dialog. The data table was formatted
separately for triplicate values.
You can import this information at the same time you import your data.
Use the Info and Notes tab of the Import and Paste dialog to specify the
rows you want to import and which Info sheet you want them to flow into.
If you can control the format of your text file, you can include info
constants and notes in structured sections at the beginning of the text
file. Mark the sections in the text file with <> variable names as below.
· Mark any values that you want to import to Info Sheets for use as
constants as:
<Info>
· Mark sections you want to go into the freeform Notes areas of an Info
Sheet as:
<Notes>
Example:
<Info>
Experimenter, “Dave Jones”
”Lot number”, 564A
Enter a tab after the <Info> variable name to separate it from its value.
If the variable name is two words, use quotation marks around the name.
The three blocks (info, notes, and title) can appear in any order, but they
must appear at the beginning of the file. You can include one, two or all
three of the blocks,
If the info variable already exists, its value will be updated. If the variable
doesn’t exist, it will be appended to the Info table.
Notes will be appended to the notes (right) portion of the Info sheet. It is
not possible to include Greek letters or super or subscripts when
importing.
The title you enter, if you include a <title> section, will replace the data
table sheet title.
When you use the View 244 tab to preview your data, the preview will not
include the structured information enclosed in notes, info, or title tags.
When you enter information on the Filter tab about which rows to include,
start counting with the first row following the structured information.
· Number format can depend on where you (or whoever made the text
file) lives. A comma sometimes denotes thousands, but sometimes is a
decimal separator. If you have trouble, use Edit..Paste Special (instead
of plain Paste). The first tab of Paste Special dialog 242 (which is identical
to the Import dialog) lets you define what commas mean.
The best way to take Prism data or results and place into Word or
PowerPoint is via copy and paste. If you do this, you'll find that Prism's
tables remain tables when they are pasted. In contrast, if you export then
import a text file, Word or PowerPoint won't format it as a table.
From any data or results table, click on the Export button on the toolbar
and choose the format you want to use.
Exporting format
You also can specify how to handle excluded values, how to format the
decimal separator, and whether or not to include column titles.
CSV
CSV stands for "comma-delimited values", but the name is a bit
misleading. In the Preferences dialog 153 , you can choose to export values
using a period (point) or a comma as the decimal separator. If you choose
a period (the US standard), then adjacent columns are indeed separated
by commas. If you choose a comma, then adjacent columns or
subcolumns are separated by semicolons.
This is a very standard format useful for moving blocks of data into
spreadsheet or word processing programs like Excel and Word.
When Prism exports into CSV, it doesn't distinguish row titles, X columns,
Y columns, and subcolumns. All the values are simply exported. Special
characters (Greek, subscripts, etc.) are lost from column and row titles.
TXT
This format is very similar to CSV, except that adjacent columns are
separated with tabs. These are common formats and can be used to
This is a very standard format useful for moving blocks of data into
spreadsheet or word processing programs like Excel and Word.
When Prism exports into TXT, it doesn't distinguish row titles, X columns,
Y columns, and subcolumns. All the values are simply exported. Special
characters (Greek, subscripts, etc.) are lost from column and row titles
XML
When you export in XML format, the exported file includes all special
formatting, including Greek characters, subscript, superscript, subcolumn
formats, etc. When you later import an XML file, everything should just
work. There is no ambiguity about whether commas are for decimals,
thousands, or used to separate columns.
If your goal is to export a data table from Prism on one computer and
import it into Prism on another computer, you should choose XML format.
While Prism follows standard XML conventions, the way Prism denotes
unique features such as subcolumn, excluded values, floating notes, etc.
is unique, so files can't be directly imported into other programs without
some extra work.
An XML style sheet is appended to every XML file exported. That means
you can view the file in a web browser, and see your data tables as
tables. But don't be fooled by this. You are not viewing the XML file, you
are viewing a rendition of the file as specified in a style sheet that is
appended to every XML file Prism exports. To see what the file actually
looks like, use a text editor.
· Just like all other values, ignoring the fact that they are excluded.
· Followed by an asterisk.
· A comma (1,23).
These choices are set in the export dialog when you export to a file. When
you copy the data to paste somewhere, these choices are set in the File &
Printer tab of Preferences dialog.
Prism uses the term analyze to include not only statistics and regression,
but also for data manipulations such as transforming, removing baselines,
and normalizing. In all cases, the results of each analysis appears on new
results pages and the data table does not change. For example, if you
transform data, the transformed values do not replace the data. Instead
they appear in a new results page.
Data and analyses are linked. If you change the data, the analysis results
will update automatically as will any linked graphs. This works even if you
have chained several analyses 262 . You can stop the automatic updating by
freezing 264 the results sheet.
Start from any data table or graph, and click the Analyze button. There is
no need to first select a range of data to analyze. Prism completely
ignores any selection you made before clicking the Analyze button.
The analyze dialog appears. Make sure the drop down on the upper left is
set to 'Built-in analysis', and choose an analysis.The right half lets you
choose to only analyze selected data sets. It is not possible to choose a
subset of rows to analyze -- Prism always analyzes them all.
Each analysis dialog has its own parameters dialog. Read about individual
analyses to learn about the choices.
If you are unsure about an analysis choice, click the Learn button in the
dialog to go to the appropriate help topic.
Prism places the results into new sheets in the Results section of your
project. These results sheets are saved with the project. At any time in
the future, you can view the results, review the analysis choices, and
change those choice to update the results.
Click the Analysis checklist button in the Interpret section of the Prism
toolbar to review a list of questions that will help you interpret the
results..
Prism comes with three on-line guides (help). The details of statistics and
regressions are explained in the other guides.
Results tables that can be analyzed further have green grids (in contrast,
tables with red grids cannot be analyzed further). There is no limit to how
many analyses can be chained together in this way. When you edit or
update the data table, Prism will update the entire analysis chain
automatically.
At the top of the Navigator, you'll then see The Data with Results folder,
which shows results nested underneath the corresponding data tables.
Analysis chains appear as a series of indented sheet names. In this
example, data are first transformed, then normalized, and then fit with
nonlinear regression.
You can choose to hide the Data and Results folder or to show it always,
even when you haven't created any analysis chains. These choices are in
the View tab of the Preferences dialog.
After reviewing the results of an analysis, you may want to change your
choices. Click on the Results button in the upper left corner of any
analysis results sheet to open the Analysis Parameters dialog.
If you have already entered data on a new data table, and want to
analyze it exactly like you have already analyzed an existing table, use
the Wand 136 (Wizard to ANalyze Data).
If you haven't yet entered the new data, consider duplicating the current
data table with its family 134 . Then edit or replace the data on the new
table and the analyses will update.
If you have placed parts of the analysis on the graph (regression lines or
curves, or copy and paste a results table), then you can clone that
graph 135 .
If the results table has a green grid (but not a red grid), you can make a
graph of the results. To do so, click the New button and choose Graph of
existing data. Then choose the graph you want to make.
Frozen results
When you edit or replace data, Prism automatically recomputes all linked
analyses. If you don't want a particular results sheet to automatically
update when the data are changed, click the freeze button in the Sheet
section of the toolbar.
For example, you might want to freeze the results before excluding some
outliers from the data, then repeat the analysis and have both copies of
the results in the project. Click again to unfreeze. When a sheet is frozen,
you'll see the word "frozen" in the status bar at the bottom of the
window, as well as a floating note that explains.
Note the distinction between deleting a data table and deleting all the
values. If you erase all values from a data set using the DEL key, Prism
still maintains the links between the data and the results. Since there are
no data, the results become blank. If you delete the entire data table, the
results remain and the sheet becomes orphaned.
Orphaned results
Prism hot links data to analysis results. If you change the data, Prism will
automatically recalculate the analyses and update the results. It will also
update any result tables that have been embedded 265 onto graphs, or any
result values that have been hooked 267 to dialogs.
If you delete a data table, any linked results become orphaned, and you'll
see the word 'orphaned' in the bar at the bottom of the Prism window.
When the data table is gone you can view and graph the orphaned
results, but you cannot change the analysis parameters.
How does Prism deal with empty data sets (or data sets with all values
excluded) when analyzing data. If you selected to analyze all the data
sets (the default), then Prism leaves empty data sets in results if there
are empty data sets in the data table. That way the data set letter is
always the same for data tables and results. In other words, the results
for data set E are always in results data set E. But if you selected to only
analyze selected data sets, Prism only analyzes selected data sets that
are not empty. If you choose to analyze data sets C, D and E, but data
set D is empty, then the results for data set C will appear in column A of
the results table, and the results for data set E will appear in column B in
the results table.
You can copy any part of a results table and paste it onto a Prism graph
or layout. The link will be 'live' so will be updated if you edit the data or
change the analysis parameters.
Double-click on the embedded table to bring up a dialog that will let you
adjust its font, borders, etc.
When you copy a single cell of a table, you have two choices when
pasting:
· Paste as text -- First use Prism's Text tool to create a text box, then paste a
results sheet value into it. In this case, the value you copied will be pasted as text,
with no link back to the analysis.
· Paste as link -- If you paste directly onto the graph or layout (not while entering
text), you will paste an embedded table (with one cell) that will automatically
update when the results are recomputed.
Prism lets you 'hook' the results of some analyses into the parameters
dialog of other analyses, and graphing dialogs. The value you hooked is
then used in that analysis, or graphing choice. The link is live, so will be
updated if the data are edited or (in some cases) if the analysis
parameters changed.
Bland-Altman · Bias
· Lower 95% confidence limit of bias
· Upper 95% confidence limit of bias
Column · Mean with lower and upper confidence limits
statistics · Geometric mean with lower and upper
confidence limits
· 25th and 75th percentile
· Median
· Lowest value
· Highest value
· QC control lines (Mean plus or minus 1, 2 or 3
SD)
· SD
· SEM
· N
· Sum
How to hook?
You'll see hook icons on many dialogs next to fields requiring you to
enter a value. Click, and then choose an analysis constant (or info
constant) on the Hook dialog.
Example 1. QC plot
First double click on the Y axis to bring up the Format Axes dialog on the
Y axis tab. The bottom of the dialog lets you add additional ticks and grid
lines. To hook the grid line to an analysis constant requires drilling down
to the details dialog, so click the '...' button.
You don't want to enter a Y coordinate on the Format Additional Ticks and
Grids dialog. Instead you want to hook an analysis constant. Click on the
hook button, and then choose the value you want to hook (first the mean,
later the two QC control lines). Back on the Format Additional Ticks and
Grids dialog, choose the kind of line (solid or dotted) and its color. You
can also choose to fill between the two QC lines
This example shows grid lines at the EC50 along with its 95% confidence
interval. All three values were 'hooked' from the nonlinear regression
using Additional ticks and grid lines 437 , so will move if data are edited or
replaced.
1. Place the insertion point in the single cell you wish to color code. Or
drag to select a range of cells.
2. Click the paint bucket icon in the Change area and choose a color. Or
drop the Change menu and choose "Cell Background Color".
Note the distinction between the feature explained here with a related
feature that lets you highlight (color code) sheet names in the navigator.
programming or scripting.
To simulate a family of XY data sets with random error, start from any
data table or graph, click Analyze, open the Simulate data category, and
Equation tab. You can choose to use Y values from the data table you
are analyzing, and then add random scatter. More often, you will choose
an equation on this tab.
Parameter values tab. On top of the tab, choose how many data sets
you wish to simulate, and how many replicates each data set will have.
The main part of the tab is where you enter the values of each parameter.
If you choose to simulate more than one data set, then you can choose to
enter a parameter value just for one data set, or to enter a parameter
that applies to several, or all, curves. Choose the data sets on the top
part of the dialog, and enter the parameter values for that data set (or
that group of data sets) below.
To simulate a set of Column data sets with random error, start from any
data table or graph, click Analyze, open the Simulate data category, and
then select Simulate Column Data.
Experimental design tab
Choose the number of data sets. In most cases you will want to specify
the population mean of each column. But you can instead enter the mean
and SD of a set of means, and have each column mean chosen randomly.
For each column you wish to simulate, specify how many rows it will
have, its column mean, and its title. Note you can hook the column mean
to a results computed in another analysis.
Random error tab
Choose among several methods for generating random scatter and also
adding outliers.
Choose the total sample size, the grand total of all four cells in the
contingency table.
· Experimental. You assign treatments that define the row, and then
tabulate outcome that defines the columns.
1. Simulate a data table using 278 one of Prism's simulation analyses.. Note
that these simulations include random scatter, so will produce new
results when they are updated.
3. From that results page, click Analyze and choose Monte-Carlo analysis.
This analysis will repeat the simulations many times, and tabulate
selected results. The Monte Carlo analysis will only be available for
analyses that create analysis constants. Note that linear regression
does not, but you can fit a straight line with the nonlinear regression
analysis.
The explanations below explain the basic ideas of the Monte Carlo
analysis. Follow the example to learn the details. 282
Simulations tab
How many simulations?
How many simulations should you run? If you make only a few
simulations, the results will be affected too much by chance. Running
more simulations will give you more precise results, but you'll have to
wait longer for the calculations to repeat. When just playing around, it
might make sense to use as few as 100 simulations so you can see the
results instantly. When trying to polish simulation results, it can make
sense to use as many as 10,000 or 100,000 simulations. A good
compromise is 1000.
Append?
If you go back to run more simulations, check an option box to append
the new simulations to the existing results, rather than start fresh.
Random seed
The choice of random numbers used in a series of simulations depends on
the random number seed used to generate the first set of results. By
default, Prism picks this seed automatically (based on the time of day),
and presents this seed in a floating note superimposed on Monte Carlo
results.
If you want two or more Monte Carlo analyses to use precisely the same
data sets (so you can compare two ways of analyzing those data), enter
that random seed on the Simulation tab.
Prism lists all of the analysis constants generated by the analysis. Check
the ones whose values you want to tabulate.
You cannot change the set of values included in this list. Let us know if
there are parameters missing, and we can add them in a future version.
If you skip this Hits tab, Prism will tabulate the selected parameters
(different columns) for each simulation (rows).
Prism can also reduce the results down to a single number -- the fraction
of the simulations that are "hits". Define a hit to be when a value
tabulated by the analysis equals a certain value, or is within a specified
range. Click New...Graph of existing data from this table, and choose a
parts-of-whole graph to create a pie graph of the fraction of hits vs. not
hits.
Prism can also tabulate the selected parameters only for simulations that
are hits, and/or for only the simulations that are not hits. Choose any or
all of these options (Hits, Not hits, All simulations) at the bottom of the
Hits tab. Each option you check will create its own results table.
Overview
When you fit a curve with nonlinear regression, one of the most important
set of results are the 95% confidence intervals of the parameters. These
intervals are computed from the standard errors which are based on
some mathematical simplifications. They are called "asymptotic" or
"approximate" standard errors. They are calculated assuming that the
equation is linear, but are applied to nonlinear equations. This
simplification means that the intervals can be too optimistic, too narrow,
so your true confidence level may be less than 95%.
How can you know whether the intervals really do have 95% confidence?
There is no general way to answer this. But for any particular situation,
you can get an answer using simulations. This page explains how to do
this with Prism using the Monte Carlo feature new to Prism 6. We will
simulate a dose-response curve and ask how accurate the 95%
confidence intervals are for the Hill Slope. Christopolous suggested that
the distribution of HillSlope can be asymmetrical, and suggested fitting
the logarithm of the HillSlope instead (1).
· X values tab. Start at X=-9, increment each X value by 0.5, and stop
when X equals or exceeds -3.0.
· Parameter values tab. Choose to simulate one data set, with 3 replicate
values. Set Bottom=250, Top=5000, logEC50=-6, and HillSlope=0.5.
1. From the graph, click Analyze and choose Nonlinear regression. Or click
the nonlinear regression shortcut button in the Analysis part of the
toolbar.
3. Click OK, and Prism will fit the model to the data and graph the curve
on the graphs.
Start from the nonlinear regression results, click Analyze and choose
Monte Carlo simulation.
On the first (Simulations) tab, choose how many simulations you want
Prism to perform. For this example, enter 1000.
On the third (Hits) tab, define a criterion which makes a given simulated
results a "hit". For this example, we'll define a hit to mean that the
confidence interval brackets the true value of 0.5 (set in the simulation).
So a hit is defined when the lower limit is less than or equal to 0.5 and
the upper limit is greater than or equal to 0.5.
Click OK and Prism will run the simulations. Depending on the speed of
your computer, it will take a few or a few dozen seconds.
The results of the simulations are shown on only one page (since we
unchecked all the options for reporting individual simulations in the Hits
tab above.
The fraction of hits is 0954. For this example, we defined "hit" to mean
that the confidence interval included the true parameter value. In other
words, the confidence interval for HillSlope included the true value (used
in the simulation) in 95.4% of the 1000 simulated data sets. (Since these
results depend on which random numbers Prism generates, you'll get
somewhat different results when you try to follow this example). This is
what you'd expect for 95% confidence intervals, and means that you can
trust the confidence intervals in experiment of this design. There is no
need to consider fitting the logarithm of the Hill Slope instead, for this
experimental design
If the confidence interval for this result is too wide for your tastes (it
ranges from 93.9% to 96.6%), go back and run this Monte Carlo
simulation with many more iterations (perhaps 10,000).
By varying the choices in the simulation dialog, and seeing the effect via
Monte Carlo simulations, you can design better experiments.
Another example
Reference
This page will help people who set up these scripts in earlier versions of
Prism, or people for whom the Monte-Carlo analysis doesn't provide
enough versatility. But for most people, the Monte Carlo analysis 280 will
be much easier.
1. Create a Prism project that simulates 278 a data set with random scatter
(the first results sheet), and then fits a curve through these data with
nonlinear regression (second results sheet).
2. Make sure that the first data table is empty and formatted for column
data (no X column; no Y subcolumns). The script will write the results
into this table. You can click in the upper left corner of the data table to
change its format, if needed.
3. Click the Prism button at the left of the toolbar, and choose Run Script.
Then choose New Script, and start with a blank page.
6. After the script is complete, go to the first data table and view the
results.
The first line of the script specifies which data table will hold the results.
Then the script loops 100 times. With each loop, it goes to the first
results page (with the simulation) and regenerates with new random
scatter. It then goes to the second results page (with curve fit results)
and writes the value in the fifth row of the first column into the data
table, and labels that column "logEC50". You will want to adjust the row
number, and perhaps add additional lines to output additional results. You
may also want to loop more than 100 times.
Prism uses the time of day when calculating the first random number, so
you will get a different series of random numbers every time you run the
program.
If you choose relative error, Prism first calculates a random number from
a Gaussian distribution with a mean of zero and with a SD equal to the
percent error you enter. It then multiplies that percentage times the ideal
Y value to yield the actual random value that is added to the Y value.
Prism also can generate random numbers from a t distribution with any
number of degrees of freedom (df). This lets you simulate wider scatter
than Gaussian. If df is low, this distribution is very wide. If df is high
(more than 20 or so), it is almost indistinguishable from a Gaussian
distribution. If df=1, the distribution is extremely wide (lots of outliers)
and is identical to a Lorentzian distribution, also known as the Cauchy
distribution. Prism uses this equation to generate random numbers from
the t distribution with df degrees of freedom:
· With Gaussian rand values, the simulated values equals the calculated
ideal value plus a random error drawn from the Gaussian distribution
(with a mean of zero and a SD you specify). We say "plus" the random
value, but this random value will be negative half the time (on average),
so the simulated value will be less than the idea value half the time.
· The RAND() function generates a random value between 0.0 and 1.0. It
is an Excel function, at least in name, so it needs to be followed by
parentheses, but the function has no inputs so there is nothing between
the parentheses.
Multiple by the standard deviation and add a mean, and you'll have
random numbers drawn from a Gaussian distribution with that mean and
SD. For example, use this formula to sample from a Gaussian distribution
with a mean of 100 and a SD of 15:
=(NORMSINV(RAND())*15)+100
Before graphing or analyzing your data, you may first need to do some
calculations to get your data into proper form. While Prism is not as
versatile as Excel, Prism can do many data manipulations. From the data
table, click on the toolbar, and choose transform 292 , normalize 299
Results will appear on a new results page with a green grid. While you
cannot edit these results, you can create a new graph (use the New...
button) and do further analyses (click the Analyze button). To help keep
track of chains of analyses, use the Data+Results folder 154 in the
Navigator.
What if Prism doesn't offer the tools you need to manipulate your data? In
that case, you should enter and manipulate your data in Excel, and then
copy and paste the results into Prism for further analysis and graphing.
Note that you can paste a link to the Excel file or you can paste embed
the entire Excel workbook within Prism. Learn more about linking and
embedding data from Excel 233 .
When you transform data with Prism, the data table isn't changed.
Instead, Prism creates a new results table with the transformed values.
Standard functions
Choose a function
Choose from one of these functions for transforming Y values (analogous
functions are available for X):
Function Comments
Y= Y * K Enter K in the box provided.
Y=Y+K “
Y=Y–K “
Y=Y/K “
Y = Y squared Enter K in the box provided.
Y=Y^K
Y = log(Y) Log base 10
Y = -1*log(Y) Natural logarithm (base e)
Y = ln(Y) Ten to the Yth power (inverse of log).
Y = 10 ^ Y
Y = exp(Y) eY (inverse of ln)
Y = 1/Y
Y = sqrt(Y) Square root.
Y = logit(y) ln(Y/1-Y)
Y = probit(Y) Y must be between 0.0 and 1.0. See notes below.
Y = rank(Y) Column rank. Smallest Y value gets rank of 1.
Y = zscore(Y) Number of SDs from the column mean.
Y = sin(Y) Y is in radians.
Y = cos(Y) “
Y = tan(Y) “
Y = arcsin(Y) Result is in radians.
Y = ABS(Y) Absolute value.
Y = Y + Random Gaussian. Mean=0. SD=K (you enter).
Y=X/Y
Y=Y/X
Y=Y–X
Y=Y+X
Y=Y*X
Y=X–Y
Y=K-Y Enter K in the box provided.
Y = K/Y “
Y = log2(Y) Logarithm base 2
Y = 2^Y 2.0 to the power of Y
Many of the functions include the variable “K”. Enter a value for K on the
dialog. When transforming Y values, you can enter one value of K for all
data sets or a separate value of K for each data set. To enter different K
values for each data set, choose a data set, enter K, choose another data
set, enter its K, and so on. Rather than entering the value of K, you can
hook an analysis or info constant 147 .
Notes about the probit() function. Prism uses an older definition of probit
that adds 5 to all results, to avoid use of negative numbers. No, it doesn't
really make sense, but this was once a common standard. So
probit(0.025) = 3.04 (which is -1.96 + 5.00), probit (.975) = 6.96 (1.96
+ 5.00), and probit(0.5) = 5.00 (0.00 + 5.00).
Interchanging X and Y
When you choose a standard function, you can choose to interchange X
and Y values and also choose transforms of X or Y or both.
· Prism can interchange data on tables with more than one data set (more
than one Y column), even though the results sheet has only a single X
column. It does this by creating additional rows. The results will be
staggered down the page with only one data set in any particular row.
If you are transforming X values, you may use Y in the function. If the
data table contains several data sets (so has several Y values for a
single X value), Prism will stagger the results down the page, repeating
X values as needed. The results for column A will appear on top of the
results table. Below that Prism will place the results for column B. For
these rows, column A will be empty.
Usually, the same function is applied to all data sets. But you can write
a transform to apply different transforms to different data sets. To do
this, precede the lines that apply to only some data sets with column
designators. For example, put <B> in front of a line in your transform
that only applies to data set B. Put <~A> in front of a line that applies
to all data sets except data set A.
Prism allows you to introduce some branching logic through use of the
IF function. The syntax is:
IF (conditional expression, value if true, value if false)
You can precede a conditional expression with NOT, and can connect
two conditional expressions with AND or OR. Examples of conditional
expressions:
MAX>100
Ymax=Constraint
(A<B or A<C)
NOT(A<B AND A<C)
FRACTION<>1.0
X<=A and X>=B
Note: “<>” means not equal to, “<=” means less than or equal to, and
“>=” means greater than or equal to. Here is an example:
Y= IF (Y<Y0, Y, Y*Y)
Y = IF (Y<0, Y/0, Y)
Many kinds of data combine a measurement (signal) you care about with
a baseline or background (noise) you don't care about. You can analyze
these data using two approaches. One approach is to perform analysis on
the total signal. The other approach is to subtract or divide by a baseline
or nonspecific value and then analyze and graph the results.
To access this analysis, start from the data table, click Analyze, then
chose the analysis Remove Baseline and Column Math.
Definition of baseline
· A value (or mean of replicate values) in the baseline column of the same
row of the table as the value from which the baseline will be subtracted
(or divided) . So the baseline value for a value in row 13 of column B
would be the value in row 13 of column A, or the mean of replicate
values in row 13 of column A. Choose between a single column being
the baseline values for all other values, or of having every other column
be baseline vs. total.
If you check the option to assume a linear baseline, Prism performs
linear regression using the X values as X and the background
(nonspecific) values as Y. It computes a predicted Y value (from the
line) at each value of X. Finally, it subtracts (or divides) this predicted
baseline Y value. This method is appropriate when you know that the
nonspecific or background values must be linear with the X values (for
example nonspecific binding is often linear with ligand concentration),
and is particularly useful when you have not collected baseline or
nonspecific measurements at every value of X (Prism will fill in the
missing nonspecific values from linear regression). When Prism fits
linear regression, it does not assume that the line goes through the
origin and does not display the regression results.
· The mean of all the values in (say) column A. You need to approach the
calculation differently if you wish to define the mean (or median) of one
column to be the baseline to subtract from (or divide into) all the other
values,. You cannot accomplish this goal using Prism's Remove Baseline
analysis. Instead, do two analyses of your data table. First do a Column
Statistics analysis to compute the column means. Then do a Transform
analysis, and choose the transform Y = Y-K (or Y = Y/K). Hook 267 the
value of K to the mean of the column you wish to use as the baseline
(usually the first column) as computed by the column statistics
analysis.
Calculation
The choices are self-explanatory. You can ask Prism to subtract the
baseline, divide by the baseline, compute the fractional or percentage
difference, etc. Note that you can also use this analysis to add or
multiply two columns.
12.4.5 Normalize
Notes:
· If you have entered replicate values, zero and one hundred percent are
defined by the mean of the replicates. It is not possible to normalize
each sub column separately.
· The X values are copied to the results table. They are not normalized.
· If you normalize to the smallest and largest value in the data set, you
can remove those values (which would become 0.000 and 1.000) from
the results.
· If your data are entered into subcolumns, you can normalize the means
or separately normalize each subcolumn (new to Prism 7).
Each row of Y values becomes one column (data set) in the results table.
The first row becomes the first data set, the second row becomes the
second data set, etc. You may not transpose a data table with more than
104 rows, because Prism cannot create a table with more than 104
columns.
The column and row titles in the results table are determined by your
choices in the dialog.
Don't forget to check the option box if you want to create a new graph
from the transposed data.
This analysis takes a large table, and creates as output a smaller table,
which can be easier to work with. But note that your original data are still
part of the file, which means the project file gets larger not smaller. But it
can be useful to have a smaller data set for quicker analyses and
graphing.
To prune, click Analyze and choose Built-in analyses. Then choose Prune
from the list of data manipulations to bring up this dialog.
Note that Prism can also decimate while importing data 241 – import a row,
skip K rows (you enter K), then import another row.
The fraction of total analysis divides each value by its column or row total,
or by the grand total.
This analysis is most often used for parts-of-whole data or for contingency
tables, but it can be used for column data and for XY or Grouped data
tables, so long as they have no subcolumns.
The calculation of confidence intervals only makes sense when each value
entered into the table is an integer that represents the actual number of
objects or events counted. In this situation, computing a confidence
interval based on the binomial distribution makes sense. Don't check the
option to compute confidence intervals if the data are measured values, or
percentages, or normalized.
With the first three choices, the results are presented on a results table
with green grid lines, so those results can be plotted. With the last choice
(presenting fraction of all three totals), the results are on a table with red
grid lines so cannot be plotted. With this last option, the option to report
confidence intervals is not available.
13 Graphs
When you create a new data table, Prism automatically creates a linked
graph.
You'll create your first data table from the Welcome dialog, and a graph
will be automatically created. When you first go the the graph, a dialog
will pop up 313 where you can choose exactly the kind of graph you want.
To make additional data tables with linked graphs, click on the New
button in the Sheet section of the Prism toolbar and choose New Data
Table With Graph. Choose the kind of data table on the New Data Table
dialog.
There are two ways to create an additional graph from an existing table.
· Duplicate the graph using New...Duplicate Sheet. Then edit one of the
copies.
1. From the data table, click the New button and choose New Graph
of Existing Data.
3. Check "Plot selected data sets only" if you don't want to plot all
the data on the graph. Then click Select and choose the data sets
to plot. You can only choose data set columns. You cannot choose
a range of rows. All rows are always plotted.
5. Usually Prism creates one graph from the entire data set. If you
want one graph for each data set, check that option.
6. Choose the kind of the graph you want. The drop down lets you
choose between XY, Column, Grouped, etc. graphs, but you'll
almost always want the kind of graph that matches the data
table.
To bring back this dialog, drop the Change menu and choose Graph Type,
or click the button shown below.
· You want two graphs of the same data. Maybe one in color and one in
black and white. Or one plotting raw replicates and another with mean
and error bar.
· You are experimenting with different ways to graph your data. You can
keep one graph while editing a copy, so exploring new ideas won't
delete a graph you sort of like.
To duplicate a graph, go to it, click the New button, and choose Duplicate
Current Sheet. Prism will create a duplicate of the graph, and both copies
will plot the same data. Now you can change the properties of the new
graph without affecting the original. In some cases, you may want to later
delete one of the copies 143 .
Duplicate a graph, and plot new data
Prism offers two ways to duplicate a graph with its data table (and
analyses).
· Duplicate family 134 . Go to the graph you want to duplicate, click the New
button, and choose Duplicate Family. The duplicated graph will look
identical to the one you started with, but it will be graphing a duplicated
data table and perhaps duplicated analyses. Edit the data on that table,
and the new graph will update.
· Clone a graph 135 . Click the New button and choose New Data Table (+
Graph). On the New Data Table and Graph dialog, choose the Clone tab,
and choose which graph to clone. Use tabs at the top to clone from a
current project, a recent project or a saved example file 126 .
When you clone a graph, you'll be asked if you want to delete the data (or
part of the data) on the new data table. When you duplicate a family,
you'll need to select and delete the data manually.
One tab on the Welcome dialog that comes up when you start a new
Prism project is portfolio. Choose that tab and view dozens of graphs
made by Prism. Open any of these to explore how they were made. You
can also edit the data and graph to make your own graph using the
portfolio graph as, essentially, a template.
Shortcut menus
Move the mouse over a symbol, title, axis, etc, and right click (control
click with Macs). The shortcut menu will give you many choices to format
the object.
Color schemes
Choose a color scheme 330 to change all colors at once.
Click to open the Change Graph Type dialog. Choose one of the
thumbnails showing the various kinds of graphs, and inspect the preview
below to make sure it is the graph you want.
This dialog is really designed to let you change graph type within a
family. For example, change from column chart with error bars to a
scatter plot. It also lets you change to a graph in a different family, but
this rarely makes sense. If you need to change to a different family of
graphs, there is a good chance you also need to change the format of
the data table 189 .
You can also change graph type within the Format Graph 368 dialog. Click
Global (upper right corner of dialog), then change all data sets, so the
changes you make apply to all the data sets and not just one.
Note that there are three ways to create bar graphs in Prism, as a
Column, Grouped, or XY graph. Don't think "make a bar graph". Instead
think about how the data are organized.
Once you understand these key points, it will be easy to control which
data sets are graphed.
· A graph can include many data sets from many tables. By default, Prism
makes one graph per table. But you can plot the same data table on
multiple graphs, and a single graph can have data from multiple tables.
The second tab of the Format Graph dialog lets you add, remove and
replace data sets on the graph (and also change the front-to-back or left-
to right order of data sets).
You don't need to use the dialog to add or remove data sets.
Remove or replace a data set by right clicking on any point or bar, and
choosing Remove Data Set or Replace Data Set.
Overview
Double click anywhere on a graph (except the axis) and the Format Graph
dialog will appear. Or click the Format Graph button.
The choices on this dialog 339 depend on whether you are editing a XY 351 ,
Column 368 , Grouped 387 , Contingency 387 , Survival 415 or Parts-of-whole 422
graph.
The Format Graph dialog gives you a huge number of options to change
the appearance of your graph. Click “Apply” to see the effect of the
changes you made without closing the dialog. This saves you time over
clicking OK to see the changes, then bringing back the dialog to make
more changes. Once you click Apply, the changes are made even if you
then cancel the dialog. To reverse those changes, use the Edit..Undo
command.
Overview
Double click on an axis, or drop the Change menu and chose the X-axis or
Y-axis command, and the Format Graph dialog will appear. Or click the
Format Axis button.
· X axis 426
The Format Axes dialog gives you a many options to change the
appearance of your graph. Click “Apply” to see the effect of the changes
you made without closing the dialog. This saves you time over clicking OK
to see the changes, then bringing back the dialog to make more changes.
The Prism Magic tool makes one (or more) graph look like another. It is a
huge time saver.
Start from a graph, or several graphs selected in the Prism Gallery 144 .
Click on the Prism Magic tool in the Change section of the Prism toolbar.
Select the graph whose appearance you'd like to match. You can choose
from any graph in the current project or any open project, any graph
saved as an example 135 , or any graph recently (two months) printed,
exported, copied to clipboard, or used as a Magic example (so long as it
had been saved, so the file has a name).
Click Next, and then select the properties of the example graph you would
like to apply to your current graph or graphs.
If you check the option "Range and ticks of axis", what Prism does
depends on whether you have checked the option to use an automatic
range for the axes. If that is checked (on the example graph, then Prism
will check that option on the graphs you are changing. Depending on the
range of the axes, they may or may not end up with the same range. If
that option is not checked on the example graph, then all of the graphs
you are changing will be changed to have the range and interval of the
example graphs.
The Rotate/Flip button in the Change section of the Prism toolbar drops a menu with
three choices. You can reverse the order of data sets on a graph, flip the page settings
between landscape or portrait orientation, or rotate column graphs from vertical to
horizontal
For XY and Survival graphs, this reverses the front to back order of data
sets. This only matters when data points overlap. You can fine tune that
order 356 in the Format Graph dialog.The two graphs in this example are
identical except that the plotting order was reversed.
Right-mouse click on the point and choose Format this Point. You can
change the symbol shape, size, color and the format, color and direction
of the error bars.
For example, on the graph below, each data point is formatted to show
different symbols and error bars directions for alternating data points.
Open the Format Graph dialog by double clicking on the graph. Go to the
Graph Settings tab, and check the option to "Remove all individual
formatting."
Click on the axis to select it. Knobs appear on both ends. Move the mouse
over the right knob of the X-axis or the top knob of the Y-axis. Drag the
axis to make it longer or shorter.
Double-click on the origin to bring up the Frame and Origin tab of the
Format Axes dialog. Select a custom shape, and enter both width and
height. Or select a standard (tall, square, or wide) shape, enter either
height or width, and Prism will automatically change the other value. On a
wide or tall graph, the long axis is 1.5 times the length of the short axis.
Pull down the Edit/Preferences menu and from the Graph tab choose Wide, Tall, or
Square. On a wide or tall graph, the long axis is 1.5 times the length of the short axis.
Click the Resize Graph button in the Change section of the Prism toolbar.
This button enlarges or shrinks both the graph and also the graph's
symbols and fonts. In contrast, resizing by stretching an axis doesn't
change symbol or font size.
If you click More Choices, a dialog lets you pick the exact size you want
and lets you resize the graph without changing font size. This can be
useful when the journal specifies both the exact graph size it wants and
the point size of fonts.
Point of confusions
magnification you see on screen, and do not change its size when printed
or exported.
Changing the length of the axes vs. changing the size of the graph
The first tab of the Format Axes dialog lets you choose exactly how long
the X and Y axes will be. The graph itself will be larger, depending on
what you choose for numbering and titles. The graph object will also be
larger if you place tables, text, drawings or images on the page.
Click the Resize Graph button in the Change section of the Prism toolbar
to change the size of the entire graph object (including everything placed
on the graph page). You can tell Prism to make the graph have a specified
width. Or you can tell Prism how long you want the X axis to be.
You can also resize a graph when exporting 518 . Specify the width of the
graph (with all the other objects on the page) in the export dialog. Prism
won't change the graph in Prism, but will create the exported graph file to
match your choices.
13.2.10 Legends
When are legends created automatically?
When your graph includes more than one data set. If it plots only one
data set, no legend is created automatically. The legend is a symbol or
bar followed by the data set name. (Legends for heat maps are
different 414 .)
When you have not entered a column title. Enter a column title on the
data table, and the graph legend will update.
You can edit the column titles of the data table, or you can edit the
legend directly on the graph. If you edit on the graph itself, you will
lose the linkage between the column title and the legend.
The two are hot linked so long as you only edit the column titles. When
you edit the legend itself, the link is broken. Restore it by double-
clicking on the graph to bring up Format Graph dialog, go to the
appropriate data set, and then check the option "Revert legend to
column title".
You can also delete legends by unchecking the "Show legend" option at
the bottom of the first tab of the Format Graph dialog.
Moving legends
Select one or more legend, and drag it around. You can also use the
commands on the Arrange menu to align and equally space legends.
Master legends
Learn how to make one legend be a master legend 488 for several graphs
on a layout.
When you plot several kinds of data on one graph, it is often useful to plot
some data using a Y-axis on the left side of the graph and other data
using a Y-axis on the right side of the graph. Note that you never can
have more than one X-axis.
Select the tab for "Right Y axis" and choose how you want the axis to be
formatted.
All data sets are initially graphed using the left Y-axis. Depending on the
scale of the axis, the data points may or may not all be visible.
If you assign a data set to the right Y-axis, and there is no Y axis. Prism
will automatically create the axis.
Turn off the left Y axis, and create a right Y axis to which you assign data
set.
Limitations
Prism always draws the right X axis, if there is one, the same length as
the left axis and starting and stopping at the same position on the page.
If you want one of the axes to be shorter, the only way to get that effect
is to cover up part of one of the axes with a drawn rectangle that you
then make white (or whatever your page background color is).
You can change the font and point size of all the text that appears on
Prism graphs:
· Set the default for fonts used in new graphs on the New Graphs tab of
Preferences 153 .
· The font of the graph title and axis titles can be changed by selecting
the title, then changing the font using the toolbar or Change..Selected
text. You can also change these fonts on the last tab of the Format Axes
dialog 316 .
13.2.14 Colors
.
13.2.14.1Color schemes
Color schemes change all of a graph's colors at once. They are very
useful, but underused.
In the color schemes dialog, you can choose a color scheme by name
using the drop down menu, or use the arrow buttons to browse the
schemes.
You are not limited to the color schemes we provide. To create a new
color scheme, start from a graph with the colors you like. Click the Color
Scheme toolbar button and choose Define Color Scheme. The new color
scheme will default to the colors of the current graph, so this makes it
easy to transfer a color scheme from one computer to another: Apply the
color scheme to a graph on one computer, open the file on another, and
create a new scheme based on that graph.
After you click the "Define Color Scheme" button shown in the screenshot
above, you can also delete any user-defined color scheme you no longer
use.
· Change the color of a data set in the Format Graph dialog. Double-click
on the graph to open it.
· To change the color of just one data point, right-click on that point or
bar and choose Format this data point 320
· Change the color of axes in the first tab of the Format Axes dialog.
Double click on an axis to open it.
When you choose any color in Prism mac, you use the standard OSX color
picker.
· The toolbar at the top lets you choose one of five different approaches
to choosing a color. The example above shows choosing from a list of
colors, but the buttons above let you also use a color wheel, color
sliders (to enter RGB or CMYK), crayons, or choosing from a color
palette.
· The magnifying glass icon gives you a sixth way to choose colors, by
matching a color elsewhere on your screen. Click on the icon, then click
on any color showing on your screen (ok if it is a different application).
· The custom colors at the very bottom of the dialog offer a seventh way
to choose a color. Simply click on the desired color. See below for
adding colors to this palette.
· Once you choose a color in the color picker, you'll also see it in the
Prism dialog. You don't need to click OK or close the Color Picker dialog.
· To add new colors to the custom color palette at the bottom of the
dialog, drag a color down from the large swatch on top of the dialog
(white in the example) and drop onto one of the custom color boxes. To
choose a custom color, click on it.
Within Prism, there are color drop downs on many dialogs (Format Graph,
Format Axis...) and on some right-click shortcut menus. Choose a color
shown and you are done. Or click the bottom choice, More Colors. for
more choices. This brings up the custom color dialog, which is part of
Windows and was not created by GraphPad. This offers three ways to
choose a color:
· Basic colors.
· Use the large color picker on the upper right of the dialog. To use this,
you need to make two choices. First, choose a color in the large box.
Then use the vertical slider to choose the shade of the color, basically
how much black is added. The slider often defaultsall the way down to
maximum blackness. In this position, all the colors look black! You'll
need to move that slider up.
Custom colors
There is room at the bottom left of the dialog for 16 custom colors you
create. To define and save a custom color, follow these steps:
1. Make sure that "Use these custom colors in new projects" is checked.
2. Check on one of the 16 custom color boxes in the bottom left of the
dialog, an empty one if there is one empty.
3. Choose a color by clicking on the large box of colors at the upper right,
or by entering RGB (or HSL) values. (See the note below about the
intensity slider.)
4. Click the button Add to Custom Colors. The color will appear in one of
the 16 custom color boxes at the lower left, and wil stay there.
Notes:
· The checkbox "Use these custom colors in new projects" only applies to
custom colors you choose after you checked that option. It won't apply
to colors you chose before checking the option.
· The colors you assign to the Custom Color slots at the bottom of the
dialog cannot be transparent. The transparency of all custom colors is
set to 0% (because this standard Windows dialog cannot store
transparency).
· If you want to create a custom color by entering RGB (or HSL) values,
you need to add a first step: click on a basic color. You'll see the RGB
values of that color. Change them to enter your RGB values. Finally click
"Add to Custom Colors". If you don't do that first step, you won't be
successful at creating a new custom color.
Transparency
13.2.14.4Semitransparent colors
The main reason is shown below. When points on a graph overlap, you
usually won't see some of them. If you choose a semitransparent color,
the overlapping points will get darker.
Transparent colors can also be useful when shading the area under a
curve (or between error bands) so you don't hide other objects on the
page.
Terminology
· Prism Mac uses the Mac term opacity, which is the complement of
transparency. At 100% opacity, the color is completely opaque. No
matter how light the color, objects that are underneath this object won't
show through. At 0% opacity, the object is completely clear so it doesn't
matter what color you picked. With values in between, you'll see the
color you chose, but objects underneath will show through.
Prism Windows lets you choose a few semitransparent colors in the color
drop down menu.
Or press "More colors & transparency" to bring up the full Choose Color
dialog where you can choose any color and apply any degree of
transparency.
Prism Mac doesn't have the color drop down menu, so you choose opacity
at the bottom of the color picker.
You can make any color with a light shade, with lots of white mixed in.
But a light color is not semitransparent. Light colors are opaque. To get a
color that looks light but is semitransparent, start with a darker color and
choose transparency.
When you export your graph or layout, the transparent (opaque) colors
may or may not look transparent (opaque) in the exported image. It
depends on what format you export in and on whether you use Prism
Windows or Mac. In the table below, "Yes" means that semitransparent
colors will be encoded in the exported file, and No means they will not be.
Note that the EPS, JPG and BMP formats simply cannot encode
transparency.
plotted.
XY graphs 349
Parts of whole and heat maps are distinct, but for XY, Survival, Column,
Grouped and Contingency table graphs, the first tab of the Format Graph
dialog controls the appearance of symbols, bars, lines, etc.
· All data sets. Drop the Global menu at the top right of the dialog,
choose to change all data sets, and then make the changes. Any
changes you make after choosing the change all data sets will apply to
all the data sets. Changes you made before making that choice only
apply to the data set selected then. Shortcut: Hold the Control key
down while you make changes. Any changes made while holding Control
down will apply to all data sets on the graph. When you are in global
mode (change all data sets), the cursor changes to include a globe:
· Several data sets. Drop the Global menu at the top right of the dialog.
Choose some data sets, and then change the settings for those data
sets. For example, you could select data sets A, C, E, and G and change
all of those data sets to use a dotted line. Then choose B, D, F and H,
and make those lines solid. Any changes you make after choosing a set
of data sets to change will apply to those data sets.
13.3.1.3 Style
The Style selections at the top of the Format Graph dialog is where you
make the most important choice about your graph. Graph points or bars?
Error bars? If so, showing what error value (SD, SEM, Range, ...)?
The choices available here depend on what kind of graph you have chosen
(XY, Column, Grouped, etc.) and how you entered your data.
Notes:
· You will only have choices of border color and thickness when you
choose an open symbol (as shown above).
· The standard set of symbols will probably have what you want:
· If you click "More", you can choose any character from any font to use
instead of a symbol.
Formatting bars works a bit different than it did in prior Prism versions.
By default, the bars are solidly filled with a color you choose, and have a
border whose thickness and color you choose.
If you want a fill pattern, check the option box, then choose the design
(the fill pattern) and color.
If you make a Column or Grouped bar graph with some bars having
positive values and others having negative values, Prism will by default
make a graph like this:
· The X axis in the graph above is at the bottom of the graph. Set the
location of the X axis (and move it to Y=0) if you want in the first tab of
the Format Axis dialog. Bring up this dialog by double clicking on any
axis.
· The bar baseline is set in the third tab of the Format Graph dialog. In
the graph above, the setting is at its default value of "automatic" which
means Prism draws the baseline at Y=0. Bring up this dialog by double
clicking on the main part of the graph.
The default graph, as above, is usually what you'll want. But if you want
something different, knowing how the X axis and the bar baselines are set
will let you make the kind of graph you want.
Prism lets you create box-and-whisker plots from stacks of values entered
into a Column table, or side-by-side replicates entered into an XY or
Grouped table.
The box
The box always extends from the 25th to 75th percentiles. These limits
are sometimes called the hinges of the plot. Surprisingly, there are
multiple ways to compute these percentile values. Prism uses a standard
method, but a different method than Excel uses.
The line in the middle of the box is plotted at the median. You can not
choose a different value, but Prism also lets you put a "+" at the mean.
The location of the whiskers
§ 5 and 95 percentiles
§ 1 and 99 percentiles
§ Min to max, show all points. This method plots whiskers down to the
minimum and up to the maximum value, but also plots each
individual value as a point superimposed on the graph.
Other options
§ Choose the border color and thickness, and fill the box with color.
§ Choose a fill pattern for the box, and choose the design (pattern)
and color.
2. Add the 75th percentile plus 1.5 times IQR. If this value is greater
than (or equal to) the largest value in the data set, draw the upper
whisker to the largest value. Otherwise stop the upper whisker at the
the largest value less than the sum of the 75th percentile plus 1.5IQR,
and plot any values that are greater than this as individual points.
3. Calculate the 25th percentile minus 1.5IQR. If this value is less than
the smallest value in the data set, draw the lower whisker to the
smallest value. Otherwise stop the lower whisker at the lowest value
greater than the 25th percentile minues 1.5IQR, and plot any values
that are greater than this as individual points.
Notes on Tukey boxplots
§ Why 1.5IQR? There is no statistical rationale; it is simply how Tukey
decided to do it, and he invented the idea of box-and-whisker plots.
§ When the Tukey method is used to create the whiskers, the ends of
the whiskers are sometimes called the inner fences.
§ If you only enter three values per group (n=3), Prism will plot the
median and range. It will not plot the percentiles and will ignore your
choice for how to plot the whiskers.
maximum. These are the same values plotted in a box-and-whiskers plot when
the whiskers extend to the minimum and maximum.
The term outlier
The term outlier has many definitions. The individual points that are plotted
beyond the whiskers of a box-and-whiskers plot are sometimes called outliers,
but this definition does not match the definition used by the Grubbs' or other
outlier tests.
You need to be just a tiny bit sneaky if you want to enter the minimum,
25th percentile, median, 75th percentile, and maximum values directly.
Prism won't know that you entered those values, but will make the graph
as if you entered raw data. If you just enter those five values, Prism will
correctly compute the minimum, maximum, and median, but the 25th
and 75th percentile values Prism computes will be incorrect.
Here is a trick: Enter the median value three times, rather than once, and
also enter the minimum, maximum, and 25th and 7th percentile values.
For each group, enter seven values on separate rows of one column (the
order doesn't matter. )
§ Minimum
§ 25th percentile
§ Median
§ Median
§ Median
§ 75th percentile
§ Maximum
You can't "tell" Prism that these values are not the raw data, so it
"thinks" they are raw data and so computes the minimum, maximum,
median and quartiles from those seven values, and plots those computed
values, which exactly match the values you entered.
You'll be able to choose error bars for XY, Column and Grouped graphs.
Choose the color of the error bars, the direction (up, down or both), the
style (capped or not) and thickness.
· Prism can plot several kinds of error bars: SD, SEM, interquartile range,
etc. Choose this at the top of the Format Graph dialog in the style
section.
· In XY graphs, the Style drop down has two choices that look like
horizontal lines (solid or dotted). These create error envelopes, rather
than error bars. You can use the Area Fill settings at the bottom of the
Format Graph dialog to choose to add color within the error envelope.
13.3.2 XY graphs
Format graph: XY
Prism lets you customize every feature of a graph. Change the shape, color, and size
of graph symbols; change the fill pattern and color of graph columns; change the
order that data sets are plotted.
Double-click on any symbol (or the background of the graph) to bring up the Format
Graph dialog.
See also
Use this dialog to change the look of all the symbols in a data set. If you
want to change the look of a single data point 320 , use the right-click
menu. You can also change all the symbols in selected data sets 320 .
Select a data set from the list, and choose the symbols, lines, and error
bars for that data set. Then select another data set and choose settings
for that data set. Click OK to redraw the graph with all the changes. Note
that a fitted curve is treated as a separate data set.
To change all data sets on the graph for example to make all symbols larger:
Click the Global button at the top right of the Format Axis dialog. Then
choose to change all data sets, or a subset of data sets. Changes you
make after that will then apply to all selected data sets.
There is a shortcut for changing all -- Hold down the control key. (Note
the 'globe' cursor since you are making global changes.)
Style
Choose how to plot error bars. Select SD, SEM, 95% CI, or range. If you
created a data table with no Y subcolumns, you won't have any choices
here. But, if you entered data into subcolumns, the choices here
determine how the replicates or error bars will appear. The Appearance
drop-down on the left determines the basic look, with further choices in
the Plot drop-down to the right (the choices will depend on what you
select in Appearance).
Since it rarely makes sense to choose different error bar (or replicate)
choices for different data sets, you will probably want to click Global and
then All before changing error bar format.
If you have four or more subcolumns for replicates, Prism will also give
you the opportunity (at the bottom of the appearance drop down menu)
to plot box-and-whiskers at each X value.
Show symbols
Choose symbol type, size, and color. If you pick an open symbol, also
select a border thickness. Check the Clear option if you want point-to-
point lines and overlapping points to show through the open symbol. Use
the Shape drop-down menu to select one of Prism's symbols. Use the
More option on this drop-down to use any character from any font on your
computer as a symbol.
Choose the format and color of error bars. Choose dotted or solid lines to
show an error envelope as in the graph on the right below. Use the Show
Area Fill button to add fill between the limits of the error envelope. Learn
more about error bars on XY graphs 213 .
Connecting line/curve
Choose the line's color, style (straight vs. staircase), thickness, and
pattern (solid, dotted, dashed, etc.). This section includes choices to have
your connecting start either at the first data point or the graph origin,
whether to leave gaps in the connecting line next to the symbols, and
whether to plot lines connecting each replicate for repeated measures in
corresponding data sets as in the graph on the bottom row below.
If you choose arrows as connecting lines, it looks best to not also show symbols.
Don't confuse the connecting line – which always goes from point to point – with a
best-fit curve or line created by regression. If you have fit a curve through the data,
Prism sees the curve as a separate data set. Select the data set for the values, and
choose symbols but no connecting line. Then choose the data set for the curve, and
choose a connecting line (curve) with no symbols.
Area fill
Fill below (or above) the point-to-point line that connects your points with a fill. Specify
fill color and pattern, and background color. You may need to adjust the front-to-
back order 356 of the data sets to get the effect you want.
Specify which axis to use for each data set. If you choose to plot on the right Y-axis,
Prism will create that axis if it doesn’t already exist.
Legend
Select Show legend to create a legend for this data set. Select Label each point with its
row title to link the legend to the column title entered on the data table. Uncheck this
box if you don’t want the legend to change when you edit the column title. You can
also edit the labels in the legend. Double-click on the legend text and type whatever
you like. Use the checkbox Revert Legend to Column Titles in the Legend to change the
labels back to match the column titles. (This checkbox is only available if you have
edited a legend.)
move each label by selecting it and dragging. Even if you move a label,
however, it remembers its position relative to the data point -- so when
you change your data, or the size and shape of your graph, the labels will
adjust accordingly. Click on any label on the graph and use the Text
section of the Prism toolbar to change the size, font, and color of all the
labels.
To fine tune the front-to-back order, go to the Data Sets on Graph tab of
the Format Graph dialog.
Changing the order of the data sets on this dialog changes the font-to-
back order of the data sets on the graph. It does not change the order
of the legends 325 .
Front to back order within a data set
Within a data set, the front-to-back order of plotting is determined by the
order of rows in the data table. The first row is furthest back and the data
entered in the last (bottom) row is plotting in front. If you need to control
which point is on top, change the order of the rows.
See also:
Click on the Format graph button on the Prism toolbar or double-click on any data point
to open the Format Graph dialog. Check the "Label each point with its row title" box to
include row titles 198 on graph near each data point.
You can move each label by selecting it and dragging. Even if you move a
label, however, Prism remembers its position relative to the data point --
so when you change your data, or the size and shape of your graph, the
labels will adjust accordingly.
Click on any label on the graph and use the Text section of the Prism
toolbar to change the size, font, and color of all the labels. You cannot
adjust the font individually. All the row labels always share the same font,
size and color.
Double-click on any data point to open the Format Graph dialog. Use the
bottom of the Data sets on Graph tab to nudge a data set.
The values you enter are the actual increments of X and Y. Each data set
is incremented as much as you specify. The nudging is not cumulative
The graph below is an alternative way to plot the same data. The second
data set is nudged by 25 units in the Y direction (and not at all in the X
direction).
The third data set is nudged 50 units in the Y direction (with zero in the X
direction).
All three tracings still have about the same baseline Y coordinates. But
nudging the data in the Y direction separated them. Because the Y values
were nudged, it would be misleading to plot the data on a full Y axis as in
the graph above. So instead, the Y axis is hidden (Frame and Origin 471 tab
of the Format Axes dialog) and replaced with a scale bar.
Waterfall graphs
Nudging can also help you create a waterfall graph. All the data sets in
the following graph cover the same range of wavelengths, and all have a
baseline of zero optical density. Each data set was nudged in both X and Y
directions, each data set nudged a bit more than the previous data set.
This creates a pseudo 3D look that is sometimes called a “waterfall”.
Leave a gap when you have missing data points, or connect the existing
data points on each side.
If you have changed the format of individual data points (by right clicking
on them 320 ) , check an option here to reset everything back to the format
set in the dialog.
The regression guide explains how to fit a line or curve to your data. But
sometimes you just want to add a line of identity -- a line where every
point has equal X and Y coordinates. Prism offers three ways to do this:
· Enter a data set on your data table with only two points defining the two
ends of the line you want to draw. In both cases X and Y will have the
same value. You can put these points on the same data table as your
data. Just put the Y values into a new Y column. Place the X values
below the X for your main data set, and make sure the Y values are in
the same row (many rows above that will be blank). When you look at
the graph, that new data set will first show as only two data points.
Double click on one to bring up the Format Graph dialog. Uncheck the
option to show symbols, and check the option to show a connecting
line.
· Use the analysis to plot a function. From the graph, click Analyze, and at
the bottom of the analyze dialog go to the Generate a Curve folder and
choose the Plot a function analysis. Choose a straight line (from the
Lines folder of equations), enter the first and last X values you want to
plot the line between, and then go to the third tab of that dialog and
enter values for the Y intercept (probably 0.0) and slope (probably
1.0).
and easy way to produce the line for someone who doesn't want to think
about math.
Check on option in the Format Graph dialog to connect the XY points with
point to point lines.
Notes:
· You can choose to have the line begin at the origin by checking an
option. Otherwise, the line starts with the first point.
· The lines are drawn point to point in the order the data are in the data
table, which may not be in the order of the X values. But note that there
is a Sort-by-X-Value command on the Edit menu you see when looking
at an XY data table.
· The graphs below show four ways to draw these lines. The bottom
choice is one you probably want to avoid. That graph uses open symbols
with clear (transparent) fill so the point-to-point lines will be visible
going through (or under) the points. You want to use white (or whatever
the background color is) fill color which covers up the point-to-point
lines.
2. At the top of the Format Graph dialog, choose to plot each replicate.
Use this dialog to change the look of all the columns in a data set. If you
want to change the look of a single column 320 , use the right-click menu.
You can also change all the symbols in selected data sets 320 .
Select a data set from the list, and choose the bar format and fill, lines,
and error bars for that data set. Then select another data set and choose
settings for that data set. Click OK to redraw the graph with all the
changes. Note that a fitted curve is treated as a separate data set.
To change all or some data sets on the graph (for example to make all
symbols larger):
Drop the list of data set names and choose --Change ALL data sets--. Or
click the Global button just to the right. Changes you make after that will
then apply to all data sets, or to data sets you select from the Global
button.
Style
Choose how you want to display your data for each data set. Choose from
a simple bar (with or without error bars), a scatter plot, an aligned dot
plot, box and whiskers, before and after plot, or floating bars. Also choose
how to plot error bars. Select SD, SEM, 95% CI, or range.
Note that your choice applies to only one data set column unless you used
the global button to choose all or a subset of columns.
It can also be useful to use different styles for each bar of a graph, as
below:
Symbols
Choice here apply to symbols used for column scatter and before-after
graphs, as well as the symbols shown (if any) on box-and-whisker plots.
If you check the option to plot a symbols on a bar graph, every value will
be plotted. If you check the option top plot symbols on a box-and-
Error bars
Choose the format and color of error bars. Learn more about error bars on
Column graphs. 215
Lines
Choose the color and thickness of lines for means on scatter plots and the
connecting lines on before and after plot.
Legend
Select Show legend to create a legend for this data set. You can edit the
labels in the legend. Double-click on the legend text and type whatever
you like. Use the Revert Legend to Column Titles checkbox in the Legend
section of the Format Graph dialog to change the labels back to match the
column titles. (This checkbox is only available if you have edited a
legend.)
On before and after plots and scatter plots, you can choose to have X
labels show on the graph to identify individual data points. Check this box
to include X titles on graph near each data point. Select each label on the
graph and drag it to move it where you want.
To change the space between two adjacent bars, go to the Data Sets on
Graph tab. Select a data set and enter (at the bottom of the dialog) the
percentage of normal spacing you want between that data set and the
prior one. If you enter 200, that bar will be twice as far from its left
neighbor as the others. If you enter 50, it will be half as far. If you enter
0, this bar will touch the neighbor on the left.
When you enter a value here as the space between a selected data set
and the prior one, you enter the percentage of the gap set in the Graph
Settings tab. If you enter 200%, you are setting the gap between the
selected data set and the prior one to be twice as large as the usual
gap.
When you enter a value on the graph settings tab as space between
columns, enter the percentage of the width of the bar (or plotting area for
the set of scatter symbols). A value of 50%, means the gaps between
bars are half as wide as the width of the bars.
To add a line between two adjacent bars, go to the Data Sets on Graph
tab. Select a data set and enter (at the bottom of the dialog) the
specifications of the line you desire.
Empty columns
The two graphs in this example are identical except that the plotting order
was reversed.
To reverse the plotting order of the data sets on a graph, click the Rotate/Flip butto
To fine tune the left-to-right order, go to the Data Sets on Graph tab of
the Format Graph dialog.
Changing the order of the data sets on this dialog changes the left-to-
right order of the data sets on the graph. It does not change the order
of the legends 325 .
Baseline
Fill in this box to change the baseline for bars on a graph. The default is
for the bars to begin at the X axis.
Check the box for "Hide Baseline" if you want your bars to float as in the
example below.
Dimensions
Prism automatically adjusts the width of columns based on the size of the
graph, the number of data sets, and the gap between bars which you set
here. To make the columns wider, make the spacing smaller.
Discontinuous axis
When you have a gap in a Y axis, choose whether or not you also want a
gap in any columns that span the gap.
If you have changed the format of individual bars (by right clicking on
them 320 ) , check an option here to reset everything back to the format set
in the dialog.
This graph below is a Forest plot, also known as an odds ratio plot or a
meta-analysis plot. It graphs odds ratios (with 95% confidence intervals)
from several studies.
1. Enter the data into a Column table. Keep the default choice to enter the
"replicates" into columns. (In this case, they are not replicates, but
Prism thinks they are).
2. Enter the results for each study into a separate data set column. Enter
in each column the Odds ratio itself, as well as the high and low
confidence limits (computed elsewhere). The order of those three
values doesn't matter. Label the studies by entering column titles. You
may use Greek symbols, super and subscript, etc.
2. Go to the graph.
3. When you first go to look at the graph, Prism will prompt for the kind of
graph you want. Choose "Column Mean with error bars".
4. Note that Prism's doesn't (can't) know that each column contains a
computed odds ratio along with the low and high confidence limits. It
just sees three values. So choose to plot the median and range. The
median and range of the three values you entered are the odds ratio
and its confidence limits computed elsewhere. If you choose to plot a
confidence interval instead, Prism will compute a confidence interval
from the three values you enter and the results will be nonsense.
5. The default graph is vertical. To make the graph horizontal (as below),
double click on the graph to bring up Format Graph, and go to the third
tab.
When you make a column scatter graph with Prism, all symbols in each
column have the same shape and color. This page explains how to make
some points have a different color.
Choose selected points in the table
1. Select a range of data, perhaps one row or one subcolumn.
Notes:
· If you have more than one graph from the same data, Prism will ask
whether you wish to change one or all graphs.
· If you have an XY table and the X column is entirely blank, then the
Format Points command will be gray. The workaround is to change your
table to a Column table.
2. Right click on a point, choose Format This Point, and choose a different
color.
2. Enter data on a Grouped data table, rather than a Column table. Put
the values you want to have different colors into different data set
columns. Put the values you want to appear in different columns of the
graph into different rows.
Example:
Use this dialog to change the look of all the columns in a data set. If you
want to change the look of a single column 320 , use the right-click menu.
You can also change all the symbols in selected data sets 320 .
Select a data set from the list, and choose the symbols, lines, and error
bars for that data set. Then select another data set and choose settings
for that data set. Click OK to redraw the graph with all the changes. Note
that a fitted curve is treated as a separate data set.
To change all data sets on the graph (for example to make all symbols
larger):
Click the Global button and choose, Change all data sets. Changes you
make after that will then apply to all data sets.
Style
Choose how to display a data set on the graph. Choose between a scatter
dot plot, an aligned dot plot, a before-and-after plot, a column or bar, a
box-and whiskers plot, or a floating bar showing the minimum and
maximum.
Choose how to plot error bars. Select SD, SEM, 95% CI, or range. Your
style choice will determine which options below you can use. Since it
rarely makes sense to choose different error bar (or replicate) choices for
different data sets, you will probably want to click "All" before changing
error bar format.
Format Symbols or bars
Use the Data Sets on Graph tab 393 to adjust the spacing between bars,
and whether you want the bars stacked, interleave, grouped, or
superimposed.
Prism labels bars with the row titles from the data table. You may not edit
the bar titles directly, but if you edit the row titles, the graph will update.
Use the Format Axes 428 dialog, not the Format Graph dialog, to turn off
the labels or to change to vertical or angled labels.
If your data table has only one row, Prism will place the column titles,
rather than row titles, under each bar.
Format error bars
Choose the format and color of error bars. Learn more about error bars on
Grouped graphs 216 .
Lines
Specify which axis to use for each data set. If you choose to plot on the
right Y-axis, Prism will create that axis if it doesn’t already exist.
Legend
Select Show legend to create a legend for this data set. Prism
automatically uses the column titles in your data table as legend labels.
You can also edit the labels in the legend. Double-click on the legend text
and type whatever you like. Use the "Revert Legend to Column Titles"
checkbox in the Legend section of the Format Graph dialog to change the
labels back to match the column titles. (This checkbox is only available if
you have edited a legend.)
If you entered XY data, with values in the X column, you can make a
Grouped graph. But note:
The whole point of a grouped graph, is that the X axis is spaced by row
number. The X value does not determine the position of the points on the
graph. This can be useful in some situations, but very misleading in
others. As shown below, the X values may not even be plotted in order of
their value. On a grouped graph, position is determined by row number
only.
The graph above has the X axis labeled with the X values. On the Format
Axes dialog, you can choose to use Row titles (if you entered any), row
numbers or column titles instead.
You can change the overall grouping scheme of a graph or make hybrid
graphs types using the order tab of the Format Graph dialog. For each
data set, choose whether it should be interleaved with, stacked on top of,
placed apart from the (starting a new group of bars, hence the name
“grouped”) or superimposed on the preceding data set.
Check the box to Separate a data set from another with a line.
Note the difference between Stacked and Superimposed below. When the
bars are superimposed, each extends from the X axis to the value. When
bars are stacked, each bar starts at the top of the one below.
Notes:
· If bars are stacked, it can help to use semitransparent colors. That way
you can see the smaller bar that is hidden behind a taller one.
· Use the middle tab of Format Graph to change the order of data sets,
which controls which data set is on top.
· With stacked bars, each row in the data table becomes one stack of
bars. If you want to leave a gap between stacks (to create a natural
grouping of sets of stacked bars), leave a blank row in your data table.
The two graphs in this example are identical except that the plotting
order was reversed.
Click the Rotate/Flip button in the Change section of the Prism toolbar to
reverse the order of data sets on a graph.
To fine tune the left-to-right order, go to the Data Sets on Graph tab of
the Format Graph dialog.
The two graphs in this example are identical except that the plotting order
was reversed.
To reverse the plotting order of the data sets on a graph, click the Rotate/Flip butto
To fine tune the left-to-right order, go to the Data Sets on Graph tab of
the Format Graph dialog.
Changing the order of the data sets on this dialog changes the font-to-
back (if you plot symbols) or left-to-right (if you plot bars) order of the
data sets on the graph. It does not change the order of the legends 325
.
The previous section explained how to adjust the order of the data sets.
The order of the individual columns within a particular data set is
determined by the order they are entered in the data table.
To reverse the row order on a data table, go to the data table. From the
Change menu, click the Sort button, and choose "Reverse row order".
See another example of changing plotting order for Grouped graphs. 200
Use the Graph Settings tab of the Format Graph dialog to change the
direction of bars on a graph from vertical to horizontal, to set the baseline
for graph columns, and to adjust the space between columns. Double-
click on any bar to open the dialog.
Direction
Baseline
Fill in this box to change the baseline for bars on a graph. The default is
for the bars to begin at the X axis. Check the box for "Hide Baseline" if
you want your bars to float as in the example below.
Spacing
Use these settings to set adjust the spacing between columns on a graph,
between groups of columns, and before the first column and after the last
one. If you make the spacing smaller, the columns will become wider.
Choose how much if any space to leave for the bar where no value was
entered (the cell is blank). In the example below, no data was entered
for black dogs. You can choose to leave no space, normal 100% space, or
something in between.
Discontinuous axis
When you have a gap in a Y axis, choose whether or not you also want a
gap in any columns that span the gap. Not leaving a gap may make your
graph misleading.
You can change how the bars are labeled. Double-click on the axis to
bring up Format Axis, then choose to use Row numbers, Row titles, or
Column titles. If you choose Column titles, you'll want to use the text tool
to add text to the graph to explain the grouping.
To plot the grand mean or grand median, plot grouped data (with
individual replicates) on a Grouped graph, choose "Plot individual values"
and then choose to plot a superimposed scatter plot. The last two choices
on the Plot drop down menu will be to plot the Grand mean or Grand
median. It is not possible to plot the grand mean or median with error
bars, and it isn't totally clear even how those error bars would be
defined.
A three way graphs is a Grouped graph set up to show data entered into
Prism for analysis by three-way ANOVA. The entered data look like this.
Note the two line column titles. These are essential both to properly label
the data, but also to properly label the graph.
From the Change Graph Type dialog, choose the Three-way tab and one
of the thumbnails.
Note that the first line of the column titles in the data table become labels
for the two halves of the graph, and the second line of the column title
becomes the graph legend. The row titles become labels under the bars.
New to Prism 7, Heat maps are a standard way to plot grouped data. The
basic idea of a heat map is that the graph is divided into rectangles or
squares, each representing one cell on the data table, one row and one
data set. The rectangle or square is color coded according to the value of
that cell in the table. Prism offers lots of options to make Heat maps
useful.
Enter data on a Grouped table. Most often you'll want to format this with
no subcolumns. In this case, each number you enter maps to one
rectangle on the heat map. Unless you choose to reverse or transpose
axes, the arrangement of the colors on the heat map will correspond to
the numbers in the table. If your table as three rows and four columns,
the heat map will also have three rows and four columns.
If you enter replicate values in side by side subcolumns, you can later
choose if you want the heat map to be based on the mean, median or
geometric mean of the replicates. You can even choose to make a heat
make of variation, and use the SD, %CV or SEM as the basis of the
heatmap.
Once you see the graph, double click on it to bring up the Format Graph
dialog with many more choices.
Color mapping 409
Gaps 413
Legends 414
Mapping replicates
If you entered replicate values (side by side), you can later choose if you
want the heat map to be based on the mean, median or geometric mean.
You can even choose to make a heat make of variation, and base the heat
map on the SD, %CV or SEM of replicate values.
· Single gradient. Choose the range of values to convert to colors and the
range of colors.
· Grayscale. You can choose whether black or white is for the lower value,
and what the range of values is.
· Rainbow. You choose the range of data values. Prism chooses the range
of colors by fixing the saturation and brightness to 100% and varying
the hue from 0° to 300°.
· Double gradient. You choose the baseline value (almost always 0.0) and
color (usually white or black). Then separately enter the maximum
value (almost always positive) and color, and also enter the minimum
value (almost always negative) and color.
· Categorical. Enter a value or range of values, and what color you want it
to show as. Repeat as needed. Ranges cannot overlap. You can't have
one range "1-2" and another "2-3" because that would leave it
ambiguous how to encode 2.0000. Instead do something like this -- one
rage for "1.0-1.99" and another for "2.0-2.99".
Choose a color to use if the value is too large or too small to be part of
the map you defined, or if the value is blank or excluded. Of course, these
choices are irrelevant if no values are off scale, blank or excluded. To
make blank and excluded values more apparent, check an option to place
a black X through those cells. Since the X is always black, use a light
color to encode these cells.
You can set the size of the entire heat map, in which case the size and
shape of each cell will be determined by that size and how many rows and
columns there are. Or set the size and shape of each cell, and the overall
size is determined from that.
Order
Choose whether to show an overall title over the graph, whether to write
the numerical value (or the mean, median, etc) of each cell in that cell,
and how to label the two axes.
13.3.5.5 Gaps
· You can put gaps next to and under all the cells just because you think
the heat map is more attractive that way. The gaps server as borders.
· Use gaps to group sets of rows and columns to give some structure to
the heatmap. Maybe the left half is for one cell line, and the right half
for another. Put a gap in between to distinguish them.
Use this dialog to change the look of all the points in a data set. If you
want to change the look of a single data point 320 , use the right-click
menu. You can also change all the symbols in selected data sets 320 .
Select a data set from the list, and choose the symbols, lines, and error
bars for that data set. Then select another data set and choose settings
for that data set. Click OK to redraw the graph with all the changes. Note
that a fitted curve is treated as a separate data set.
To change all data sets on the graph (for example to make all symbols
larger):
Drop the list of data set names and choose --Change ALL data sets--. Or
click the ALL button just to the right. Changes you make after that will
then apply to all data sets.
Show Symbols
Choose symbol type, size, and color. If you pick an open symbol, also
select a border thickness. Check the Clear option if you want point-to-
point lines and overlapping points to show through the open symbol. Use
Shape the drop-down menu to select one of Prism's symbols. Use the
More option on this drop-down to use any character from any font on your
computer as a symbol.
In survival plots, you can choose to show symbols only for censored data
points, or for all data points. Make this selection on the Graph Settings
tab of the Format graph dialog.
Error bars
Choose how you want error bars to be formatted. Error bars can go above
or below a data point or in both directions. Learn more about error bars
on survival graphs 218 .
Depending on the choices you made when you created a survival plot, the
option for error bars may or may not be available. You can change this
setting in the Graph Settings tab of the Format Graph dialog.
Connecting lines
Choose the format of lines connecting data points for a survival plot.
Consider using arrows.
Show Area fill
Choose to show area fills below or above the plot lines. Click the All
button to turn on this feature for all the data sets on a graph.
Specify which axis to use for each data set. If you choose to plot on the
right Y-axis, Prism will create that axis if it doesn’t already exist.
Legend
Change the order to determine plot lines are in front. Plotting area
matters the most when you use area fills. The graphs below are the
same, but with the plotting order reversed. Note that on the graph on the
left the area fill from the one data set hides the other.
Use this dialog to change the look of a Parts of whole graph If you want to
only change the look of a single segment, it is simpler to use the right-
click menu on the graph. You can also change all the symbols in selected
data sets 320 .
Click on the thumbnail image in the Graph type section of the dialog. You
can change from a pie chart to a donut, slice or dot plot. Prism will
preview how your graph will look after the change.
Select a data set from the Colors and legend list, and choose the color of
the graph segment, and what information you want the legend to show.
Click the Global button to select all data sets on the graph. Or, select the
data sets from the drop-down menu that you want to change.
The graph is generated from values entered into a single column in the
data table. If you enter data into a row, you'll only get one value on the
graph.
Color and Legends
You can set the color of each graph segment. Use the roll-down to choose
if you want a particular segment of a pie or donut graph exploded from
the main graph and how far away you want it to be. Or choose to hide a
segment.
Choose the information you want the legend to show and how you want it
formatted.
If one of the rows of data is zero, then Prism will not create a legend for
it. If you want a legend, even though the value is zero, enter some tiny
nonzero value like 0.000001. The slice will be invisible on the graph, but
Prism will create a legend for it.
Borders
Choose the color and thickness of the segment borders. You can have
different settings for the inner and outer borders. These setting are for all
the segments of the graph.
Titles
Click to show or hide the graph title, or to show or hide the total of all
values, The graph shows each category as the fraction of the total, so it
can be very useful to show the total near the graph. With donut graphs,
it is common to put the total in the center of the graph as shown below.
Optionally add a prefix or suffix, for example: 1,000 Applicants, or Total:
1,000.
Order
Reverse the order of the data sets to change the orientation of the graph.
Size
Use the Size portion of the dialog to change the size of the whole of a
donut graph.
A parts-of-whole graph can only plot one data set. To plot a different data
set, double click on the graph to bring up the Format Graph dialog, go to
the second tab, and choose the table and data set to plot.
Note that the values that comprise a parts-of-whole graph are always
entered into one data set column. If your values are in a row, you'll need
to select them, copy to the clipboard, go to an empty row and paste.
You can only set the range on numerical axes, not the categorical axis
of Column and Grouped graphs.
Double-click on an axis to open the Frame and Origin tab of the Format
Axes dialog. Or click on the Format Axis tool in the Prism toolbar.
XY graphs
Normally, of course, the X axis of an XY graph plots the X values you
enter into the X column. The X axis, in this case, is numerical. If the X
values are not evenly spaced, then the points won't be equally spaced.
If you enter row titles for each row, then you can choose to label the X
axis with row titles rather than X values. Choose in the Number format
drop down in the Regularly spaced ticks part of the Format Axes dialog.
When you choose to label the X axis with row titles, the spacing of the
points may change. Essentially, Prism uses the row number as the X
value, so the points will be equally spaced along the X axis, even if the X
values are not evenly spaced. By default, each row label will appear. If
you have many rows, this will be cluttered. You can ask Prism to plot,
say, every fifth row label by changing the major tick interval to 5, starting
at 0. Keep in mind that Prism uses the row number (first row is 1) rather
than the values entered into the X column, to position the data and
determine which labels to show. Prism also uses this coordinate system
(row numbers) when positioning additional ticks and grid lines. If you ask
for a grid line at X=5, Prism will draw it to correspond with the point in
the fifth row, whatever its X value.
will ignore the X values entirely, and use the row number (top row is 1) as the X
values when plotting values.
Prism lets you show both major and minor ticks on an axis. By default,
Prism automatically sets the minimum and maximum range for an axis as
well as the major tick interval.
Double click on an axis to open the Format Axes dialog. Uncheck the Auto
box in the Range section and set the major tick interval in the Regularly
Spaced Ticks section.
To hide ticks:
Double-click on an axis to open the Format Axes dialog. In the All Ticks
section, choose None for Direction in the All Ticks section.
Double-click on an axis to open the Format Axes dialog. Set the direction
and length of ticks in the All Ticks section.
Double-click on an axis to open the Frame and Origin tab of the Format
Axes dialog. Or click on the Format Axis tool in the Prism toolbar.
Choose to hide one or both axes in the Frame & Origin section. You can
also choose to Show Scale Bars.
Scale bars
To show or hide the scale bar:
If you choose to hide either axis, a new option will appear in the Frames and axes
section of the Format Axes dialog: Show Scale Bar. Check this option to see scale
bars; uncheck to hide scale bars.The length of the scale bar equals one axis
interval, set in the Regularly Spaced Ticks section of the Format Axes dialog.
To delete a scale bar:
Click on a scale bar to select it, and then press DEL to delete it.
To change the color or thickness of a scale bar:
Double-click the scale bar to bring up the Format Object dialog. Choose color and
thickness in this dialog.
To remove the caps on the end of the scale bar:
By default the scale bars are capped. These are actually "arrowhead" styles. To
change, double-click the scale bar to bring up the Format Object dialog. Then
change the arrowhead settings.
To move a scale bar:
Drag with the mouse. Or click to select, and then nudge by clicking the arrow keys.
To label a scale bar:
Scale bars don't include any numbering. To create a label for a scale bar, click the
text tool ("T" button), click near the scale bar, and type.
To change the length of the scale bar:
You can’t change the length of a scale bar by stretching it. Instead, bring up the
Format Axes dialog and change the tick interval in the Regularly Spaced Ticks
section.
Set the range of each segment and its length as a percent of the total
length of the axis.
Finally, choose the spacing and properties of minor and major ticks
separately for each axis segment.
Limitations:
· You cannot make one segment have a logarithmic or probability scale 455
· You cannot make one segment go in reverse while the other segment(s)
go in the usual direction.
To move the gap along the axis, or make the graph wider or narrower,
click on ab axis segment to select it and then:
· Drag the knob just to the left of (or below) the gap to change the
position of the gap along the axis (make one segment larger, and
another smaller).
· Drag the knob to the right of (or above) the gap to widen or narrow the
gap itself
Example 1
Because of the change in scale, the linear regression line appears to have
two slopes. In fact the mathematical slope is identical everywhere, and
the line was created by a single analysis of all the data. The slope just
appears steeper in the right half of the graph because the time scale is
more compressed.
Example 2
To place "0" on the left segment, create an additional tick label at X=-9
and enter the text "0".
Example 3
The line was drawn by linear regression, which automatically skips over
the axis gap.
Click on the Details button of the Additional Ticks and Grid Lines
section of the Format Axes dialog. This appears as a gear on Prism
Window, and as "..." on Prism Windows. This lets you format how the tick
or grid line will look and add text to the label.
Choose whether you want each tick in line with or across from the regular
ticks.
You also can Hook 267 the text of the label to an analysis or Info Sheet
constant 147 .
Check the option to fill between two ticks, and choose the fill color and
pattern.
Prism can only fill the space between an Additional Grid line and one of its
neighbors. To create fills on both sides of a grid line, place two additional
grid lines at the same position and fill backward from one and forward
from the other.
Double-click on the axis to bring up the Format Axes dialog. Locate the
tick you want to change in the Additional Ticks and Grid Lines section.
Click the Details button to open the Format Additional Ticks and
Grid lines section and click on the Delete Tick button.
In this example, the dashed line for Kd and BMax were hooked to the Kd
and BMax values in fit by nonlinear regression. If the raw data are
subsequently changed, both the results information and this grid lines will
update.
By default, Prism puts generic titles on top of the graph and on the X-
and Y-axes.
To edit a title:
Click in the title to select it. If you are editing a Y-axis title, it will be
temporarily displayed horizontally for editing. Edit (or replace) the text.
To enter more than one line, press Enter between lines. When you are
done, click in the background of the graph. While editing the title, you can
use controls in the text toolbar.
To delete a title:
Drag over the title to select it. If you are editing a Y-axis vertical title, it
will be temporarily displayed horizontally for editing. Then press the DEL
key.
Double-click an axis to open the Format Axes dialog. Check "Show title" in
the Titles and Fonts tab of Format Axes dialog.
Select the title, or a portion of it, and use Font tools on the toolbar. Or
double-click on the axis to open the Format Axes dialog, go to the
Titles&Fonts tab, and click the Font button.
Select the title, Then press the or buttons on the toolbar to change
the point size.
To justify a title:
Usually a title is centered over the graph or axis. To justify the title with
the right or left edge of the axis, select the title and click on the text
justify button on the Text toolbar.
To move a title:
Move the mouse near the title until the cursor turns into a two-headed
arrow . Then hold down the primary mouse button and drag the title
into position. Alternatively, go to the Titles and Fonts tab of the the
Format Axes dialog, where you can position the title by entering the
desired distance between the axis and the title.
A graph title is always centered over the graph, and an axis title is always
centered on the axis. If you want noncentered titles, delete the title and
then create a separate text object which can be positioned anywhere.
To change the default font and size used for new graphs:
Pull down the Edit menu (the Prism menu in Mac OS X), and choose
Preferences, then go to the New Graphs tab.
From XKCD.
New to Prism 7 is the ability to enter elapsed times instead of X values 191 .
This page explains how to make graphs when you entered data this way.
Another page 450 explains how prior versions of Prism dealt with dates and
times.
The Format Axis dialog offers two new choices when you entered elapsed
times instead of X values. '
If you choose an Elapsed time scale, you can choose to display just hours,
hours:minutes, hours:minutes:seconds or
hours:minutes:seconds.milliseconds. To avoid overlap, you'll probably
want the dates to be displayed vertically or at an angle (not horizontal).
If you choose a linear (numerical) scale, Prism tells you what the units
will be. These units are set on the data table.
To change those units, go back to the data table, drop the Change menu,
choose Format Data Table 189 , then uncheck "Choose automatically" and
select the units you want to use.
What values are used when you fit a curve or line to the data?
Note that the choice of months is not offered, because months have
different lengths. Years also have different lengths, so if you choose
years, Prism does the conversion based on 365.25 days per year.
New to Prism 7 is the ability to enter dates instead of X values 191 . This
page explains how to make graphs when you entered data this way.
Another page 450 explains how prior versions of Prism dealt with dates.
The Format Axis dialog offers two new choices when you entered dates
instead of X values.
Date scale
If you choose a Dates scale, you can choose many ways to display the
dates. To avoid overlap, you'll probably want the dates to be displayed
vertically or at an angle (not horizontal).
When you choose a Dates axis scale, Prism is smart about how far to
space the dates. Jan 3 is much closer to Jan 1 than is Feb 3...
To change those units, go back to the data table, drop the Change menu
and choose Format Data Table.
What values are used when you fit a curve or line to the data?
When you analyze XY data where X values are entered as dates, Prism
converts the dates to numbers using a starting date and a date unit
(days, weeks, etc.). To change this, go back to the data table, drop the
change menu, and choose Format Data Table.
Prism does not understand the concept of dates or times, so you cannot
enter times or dates as X values. But you can get Prism to plot dates or
times as row titles.
Note that Prism does not try to interpret the meaning of row titles. To
Prism, you just entered labels. The data will always appear equally
spaced on the graph, even if the dates are not equally spaced
chronologically. Most of the time, your dates will be equally spaced so this
approach works well. But sometimes, your dates are not equally spaced
on the calendar and the graph is misleading:
You'll need to use a different approach to plot a series of dates that are
not equally spaced. Enter the data onto an XY table. Enter the dates as
row titles. Into the X column, enter the number of days since some
starting point.
To make this graph, you'll need to make some changes on the Format
Axis dialog. Leave the Number Format set to Decimal. Don't choose Row
Titles, or you'll end up with an equally spaced graph. Use the Additional
Ticks section to specify each label and its location (X coordinate). Check
the option to plot only these additional ticks (instead of regular ticks).
If your data span a wide range of values, it can be easier to visualize the results
when you make one of the axes logarithmic. Distinguish between transforming
your data to logarithms (which creates a new results table) from choosing a
logarithmic axis (which only stretches the axis, without altering your data).
Advice: Bar graphs with log axes rarely make sense 461
The graph on the left has a linear (ordinary) axis. The difference between
every pair of ticks is consistent (2000 in this example).
The graph on the right has a logarithmic axis. The difference between
every pair of ticks is not consistent. From the bottom tick (0.1) to the
next tick is a difference of 0.9. From the top tick (100,000) down to the
next highest tick (10,000) is a difference of 90,000). What is consistent is
the ratio. Each axis tick represents a value ten fold higher than the
previous tick.
The red dots plot a data set with equally spaced values. Each dot
represents a value with a Y value 500 higher than the dot below. The dots
are equally spaced on the graph on the left, but far from equally spaced
on the graph on the right. To prevent overlap, the points are jittered to
the right and left so they don't overlap. The horizontal position of the red
dots has no other meaning.
The blue dots represent a data set where each value represents a Y value
1.5 times higher than the one below. On the graph on the left, the lower
values are almost superimposed, making it very hard to see the
distribution of values (even with horizontal jittering). On the graph on the
right with a logarithmic axis, the points appear equally spaced.
Why “logarithmic”?
In the example above, the ticks at 1, 10, 100, 1000 are equally spaced on
the graph. The logarithms of 1, 10, 100 and 1000 are 0, 1, 2, 3, which are
equally spaced values. Since values that are equally spaced on the graph
have logarithms that are equally spaced numerically, this kind of axis is
called a “logarithmic axis”.
What value is halfway between the tick for 10 and the one for 100 on a
logarithmic axis? Your first guess might be the average of those two
values, 55. But that is wrong. Values are not equally spaced on a
logarithmic axis. The logarithm of 10 is 1.0, and the logarithm of 100 is
2.0, so the logarithm of the midpoint is 1.5. What value has a logarithm
of 1.5? The answer is 101.5, which is 31.62. So the value half way
between 10 and 100 on a logarithmic axis is 31.62. Similarly, the value
halfway between 100 and 1000 on a logarithmic axis is 316.2.
Lingo
The term semilog is used to refer to a graph where one axis is logarithmic
and the other isn’t. When both axes are logarithmic, the graph is called a
log-log plot.
This page shows how to stretch the axis to a logarithmic scale (which
doesn't alter your data). Distinguish this from transforming your data to
logarithms, which you can plot 463 .
Double-click on the axis to open the Format Axes dialog. In the Scale
section, set the axis to Log10 or Log 2.
This graph shows log spaced minor ticks. The minor ticks between the
major ticks at 10 and 100 show the position of 20, 30, 40, 50, 60, 70, 80
and 90. Since the minor ticks are not equally spaced on the graph, we
call these log spaced minor ticks. Note that there are eight ticks, so nine
intervals (many people mistakenly think there ought to be ten).
Double-click on the axis to open the Format Axes dialog and set the
number format in the Numbering or Labeling section. Prism offers three
numbering format choices for log axes.
Double-click on the axis to open the Format Axes dialog and set the
number format in the Numbering or Labeling section. Choose Equal or
Log spacing, and the number of intervals you want between major ticks,
not the number of ticks.
The logarithms of negative numbers and zero are simply not defined
Let’s start with the fundamental definition of a logarithm. If 10L = Z, then
L is the logarithm (base 10) of Z. If L is a negative value, then Z is a
positive fraction less than 1.0. If L is zero, then Z equals 1.0. If L is
greater than 0, then Z is greater than 1.0. Note that there no value of L
will result in a value of Z that is zero or negative. Logarithms are simply
not defined for zero or negative numbers.
Bottom line: A logarithmic axis can only plot positive values. There simply
is no way to put negative values or zero on a logarithmic axis.
Choosing a log scale changes the way the graph looks, but does not
change the data values. Linear or nonlinear regression fit an equation to
your data. Your choice of axis scaling does not affect any calculations.
The three graphs shown below represent the same data by linear
regression.
The graph below has linear X and Y axes, so the linear regression line
appears straight.
The graph below has a logarithmic X axis. The curve you see is the
results of linear regression. It is the best fit to the equation Y=intercept
+ slope*X. The X values used by the linear regression are the X values
you entered. Graphing this linear regression line on a log axis results in
a curve.
The graph below has a logarithmic Y axis. Again, the best-fit linear
regression line appears curved when plotted on a logarithmic axis.
The two graphs below show the same data. X is dose, and Y is response.
The doses were chosen so each dose is twice the previous dose. When
plotted with a linear axis (left) many of the values are superimposed and
it is hard to see what’s going on. With a logarithmic axis (right), the
values are equally spaced horizontally, making the graph easier to
understand.
The graph above was created with the outcome (the odds ratio) plotted
horizontally. Since these values are something that was determined (not
something set by the experimenter), the horizontal axis is the Y axis even
though it is horizontal.
A logarithmic axis linearizes compound interest and exponential growth
The graph on the right is identical to the one on the left, except that the Y
axis has a logarithmic scale. On this scale, exponential growth appears as
a straight line. Exponential growth has a constant doubling time. For this
example, the Y value (cell count, value of investment) doubles for every
time increment of 7.2657.
13.4.3.6 Advice: Bar graphs with log axes rarely make sense
Prism can create bar graphs with a logarithmic Y axis, but think twice
before deciding to make this kind of graph. The point of a bar graph is to
let people compare heights or areas of the bars. In most cases, you
should try to avoid using a log axis when creating a bar graph.
The treated bar in the graph below (with a linear axis) is four times as
tall (and has four times the area) as the control bar. This tells you that
the treatment increased the response fourfold. This is very
straightforward since the Y axis is linear.
Beware of bar graphs with a logarithmic axis. Since zero can’t be shown
on a log axis, you need to decide where to begin the axis. A log scale
has no logical starting place, so the choice is arbitrary. The graphs below
show the same data as the graph above, but with the Y axis beginning at
0.001, 10 or 100.
The relative height of the bars appears to be almost the same in the
graph on the right and very different on the graph on the left. The
relative heights (or areas) of the bars on a log scale is not informative.
The only case that I know of where it makes sense to plot bars on a
logarithmic axis is when the Y axis is plotting the fold change in
something. Then the baseline of 1.0 is logical (no change), as is a
logarithmic scale.
Click the Analyze button, choose built-in analyses, and then select
Transforms from the list of data manipulations. Choose X = log(X).
Also check the box at the bottom of the dialog to Create a New Graph of
the results. Prism will create a results table of the transformed data, and
a new graph. To perform nonlinear regression on the transformed data,
click Analyze from either the results table or the new graph.
In the graph below, An alternative way to indicate that the data are
logarithms is to choose the antilog numbering format (this choice is only
available when the interval is 1, the minimum is –9 or greater, and the
maximum is 9 or less).
You can take 10 to a fractional power. Ten to the 1/2 power equals the
square root of 10, which is 3.163. So the logarithm of 3.163 is 0.5.
You can take the logarithm of any positive number. The logarithm of
values between zero and one are negative; the logarithms of values
greater than one are positive. The logarithms of zero and all negative
numbers are undefined; there is no power of 10 that gives a negative
number or zero.
You can compute logarithms for any power. Mathematicians prefer natural
logarithms, using base e (2.7183…). Despite that name, use of
logarithms with base e does not seem natural to most scientists who tend
to be far more comfortable with base 10 logarithms.
log(An) = n.log(A)
Antilogarithms
The term logistic sounds like it is related to logarithms. In fact, the term
logistic has three distinct meanings, which don't all involve logarithms.
Within Prism transforms and equations, the log() function computes the
common (base 10) logarithm.
The frame determines the basic look of a graph, and its position on the page.
If you plan to copy and paste the graph to another program, or place the
graph on a Prism layout, you won't really care where it is on the page.
But if you plan to print that page, you might care to position a graph
precisely.
To see the position of a graph on the page: Drop the View menu,
choose Zoom, and choose To Fit Page. This lets you see the entire page
at once, so you can see the position of the graph relative to the entire
page.
To center the graph on the page: Click on the Center on page button
in the Arrange section of the Prism toolbar.
To drag a graph on the page: Grab either axis and drag the graph.
Double-click on the graph origin to open the Frame and Origin tab of the Format Axes
dialog.
Use this dialog to set the origin of your graph and to format the color
and shape of the axes or surrounding frame.
To change the gap of an offset axis, click and drag the end of the axis.
To draw a box around a graph 495 , including titles and numbering, use the
drawing tool. Or wait until you assign a graph to a layout, and choose
the border on the second tab of that dialog 483 .
Double-click on the graph origin to open the Frame and Origin tab of the
Format Axes dialog. Or click on the Format Axis tool in the Prism toolbar.
Note that grid lines from the Y axis always come from the left Y axis.
Even if you turned off that axis, its major tick interval controls the grid
lines. It is not possible to create automatic grid lines from the right Y
axis. But you can place additional ticks or lines 437 one at a time to give
the same effect.
Double-click an axis and use the Format Axis tab to place additional ticks
or lines 437 (formerly called Custom Ticks) on your graph. You can create
as many grid lines anywhere on an axis.
The position where the X and Y axes intersect is called the "origin." To
change the position of a graph's origin, double-click on the origin of a
graph to open the Frame and Origin tab of the Format Axes dialog). Or
click on the Format Axis tool in the Prism toolbar.
Here is a list of shortcuts to change the axis directly, without using the
Format Axes dialog:
· To change the thickness or color of one axis: Click once to select that
axis. Then click Change and choose Selected Object(s). If you change
the X axis, the top part of the axis frame will also change accordingly.
If you change the left Y axis, the right part of the axis frame will also
change accordingly (unless you created a right Y axis).
· To move the graph and axis titles closer to, or further from, the
graph: Position the mouse near the title until the cursor becomes a
two-headed arrow. Then drag.
· To move the axis numbering (or labeling) closer to, or further from,
the axis: Move the mouse over the numbering until the cursor
becomes a two-headed arrow. Then drag. Or click once on the
numbering to select. Then nudge one pixel at a time by clicking the
arrow keys.
· To format the size, font, color, etc. of the numbering of an axis, click
to select the axis and then use the controls in the Text toolbar.
14 Page layouts
Click the New button on the toolbar, or drop the Insert menu, and choose
New Layout
2. Choose the number of placeholders you want on the page, and their
arrangement.
Choose a graph
Note that you can not only choose a graph from the current project, but a
graph from any project. It is especially easy to choose a graph from
another open project. If you choose to Pick a graph from another project
that is not open, after you choose that project file, Prism will open it.
By default, Prism the graph on the layout automatically updates when the
graph itself updates. You can choose to place the graph on the layout as
an unlinked picture with no link to the original data.You might prefer this
option if you are sure the graph is 100% done, and you don’t want to
allow any possibility of it being changed. It can also be useful if you want
to keep "snapshots" of a graph as a project continues and additional data
are collected and graphed.
· Merge a copy of the graph, along with its data tables and analyses, into
the project you are working on now. The advantage of this approach is
that everything will be in one project. The disadvantage (which some
consider to be an advantage) is that the data will now exist in two files.
Another advantage is that your file is self-contained. Even if you move it
to a different computer, everything will work.
· Live link. The data and graph are not in the project with the layout. To
edit a linked graph, go to the original project. Then resave that file and
the changes will appear in the layout. Prism can maintain the link only if
the name and location of the linked file doesn't change. If you change
computers, you'll almost certainly break the link. The advantage of
linking (which some will consider to be a disadvantage) is that the data
and graph only exist in one place.
· Unlinked picture. As the name suggests, this is just a picture that is not
linked to any data or any editable graph.
Cropping
The first tab of the Assign Graph graph dialog lets you crop a graph to get
rid of white space or portions of the graph you don't want to see on the
layout. If you enter a negative value for any of the cropping values, you'll
make the graph placeholder larger, adding white space in that direction.
The second tab (Size and Location) lets you add a border around the
graph and change its size and position.
When you first create a layout page, Prism asks you to choose an
arrangement of placeholders. But you are not stuck with this
arrangement. You can always go back to the Format Layout dialog by
clicking its toolbar button, or dropping Change and choosing Arrangement
of graphs.
If you assign the wrong graph to the layout, it is easy to replace. Just
double-click on that graph (in the layout) and choose its replacement. Or
drag the name of the replacement graph from the Navigator and drop on
to the graph you want to replace.
It is easy to add a new graph to the layout. First resize and move graphs
to make space for a new one. Then do any of the following (the first two
choices are the simplest).
· Drag a graph from the Navigator 100 and drop onto a blank spot on the
layout page. Don't drop too close to an existing graph or you might end
up replacing that graph rather than adding a new one.
· Go to the Graphs section of this project or any other, and copy a graph
to the clipboard. Then go to the layout, and paste.
· From the layout, drop the Insert menu and choose Prism Graph. Then
choose the graph you want to add to the layout.
· From the layout, bring up the Format Layout dialog (by clicking its
toolbar button, or dropping Change and choosing Arrangement of
graphs) and choosing to add another graph to the layout.
The scaling factor is defined as the ratio of the size of the graph on the
layout compared to its size on the graph.
If the graph and layout are zoomed 103 differently, it will be hard to
understand the scaling factor. Set the zoom to be the same for the graph
and layout (using controls in the lower right of the Prism Window).
Changing the scaling factor of one graph
If you want to set the scaling factor to a particular value, double click on
the graph (on the layout) to bring up the Place Graph on Layout dialog.
Go to the second (Size and Location) tab, and set the Scaling Factor.
Note that this tab also lets you specify a border to go around the graph.
1. Select the graphs to equalize (not necessary if you want to equalize all
graphs on the layout).
2. Choose Equalize Scaling Factor from the Arrange menu, or press the
button in the Change section of the toolbar.
3. Choose to reduce the scaling factor of graphs that are larger than the
others, or to increase the scaling factor of those that are too small. If
graphs were originally made the same size, with same size fonts, this
command will get things the same again on layout. If graphs started off
with different sizes, however, this command won't be very helpful.
1. Choose Resize Layout from the Change menu, or click the Resize
button.
2. You'll see a menu with choices: Smaller, Larger, To Fill Page, or More
Choices (which brings up the dialog shown below).
1. Enter data on an XY table. Enter the depth values into the X column.
3. Use the Format Axis dialog to change the direction of the numbering on
both axes. Set the X axis numbering to: Below, vertical. Set the Y axis
numbering to: Left vertical.
5. Create an independent text object with the depth title. Select it, and
rotate 180 degrees by clicking the rotate button in the lower right of
the Text area in the toolbar several times.Your graph will now look like
this:
The result, on the layout, is a depth chart, with the X variable on the
vertical axis and the Y variable (and its error bars) plotted on the
horizontal axis.
If you place several graphs on a layout, you can create a master legend
that applies to all the graphs as shown below.
It took four graphs to make the figure above. The first three graphs show
data, and the fourth (not shown) was used only to create the master
legend. Follow these steps:
3. Select and copy the legends from the extra graph and paste onto the
layout (Paste Drawing). That is a simple copy and paste operation and
does not use a placeholder.
Notes:
· The master legend is linked to the graph from which it was copied. If
you change the symbols in that graph, Prism will update the master
legend.
· Master legends on a layout have exactly the same size as the legend on
the graph from which you copied the legends. When you place graphs
on layouts, you can shrink or expand it to fill a placeholder. You can't do
this with master legends. The only way to change the size of the symbol
in the master legend is to change the font size and symbol size on the
graph used to create the master legend. This is done in the graphs
section, not the page layout section.
· You can edit the text of the legend, change the size of the text in the
master legend, and make it bold or italics. These changes will be lost if
you make any changes to the graph from to which the master legend is
linked.
· After copying a legend, you can only paste it into a layout. It is not
possible to paste a legend onto another graph.
If you have a page layout whose arrangement you like, duplicate it (just
the sheet). Then double-click on each graph to bring up a dialog where
you can choose a different graph for that slot.
Prism offers a direct way to clone a graph from the Welcome dialog, but
does not allow cloning of layouts. However, you can accomplish the same
goal by going to a layout, dropping the New button, and choosing to
Duplicate Family. This will duplicate not only the layout, but all graphs on
the layout and all results and data sheets linked to those graphs. Edit all
the data tables, and you are done. The analyses and graphs will
automatically update.
Click one of the text 'T' tool buttons (the one on the right makes a text
box with a border).
Click on the graph or layout at the spot where you want to enter text and
type. If you want to enter several lines, press Enter between lines. When
you are finished, click to select another tool or somewhere else on the
page.
Enter the first line or two using one of the text tool buttons and then click
anywhere else on the page to create the text object. Go back and click on
your text. You'll see knobs on the corners. Drag the knobs to make the
text object the correct width. Don't worry about the height, as it will grow
automatically as you enter text. Now, double-click inside the text object
to bring back the insertion point and type your text. Don't push Enter at
the end of each line, as Prism will automatically wrap the text as you
type.
First be sure you have a text insertion point. Then press the Greek button
and choose a character from the menus provided.
You can add a constant from an Info sheet 147 in a text box. Use the Insert
Info Constant button on the Write section of the Prism toolbar. Once
you've added an Info constant to your graph or layout, double click to
open the text box and add a label. If you edit the Info constant, the linked
text-box on your graph will be updated.
To enter en or em dashses:
Changing text
Standard toolbar buttons let you make text bold, italicized, underlined,
subscripted, or superscripted, and let you change font, color, and size.
Select a block of text or an entire text object and use the right-click
menu. Or pull down the Change menu and choose Selected Text to bring
up the font dialog.
GraphPad does not provide any fonts with Prism, but lets you choose any
TrueType font already installed in your system. If you want a font that is
not in the list, you'll need to find or buy the TrueType font and then install
it into Windows or Mac OS X.
To rotate text:
You can rotate text objects to any angle you want. Click on one of the
text-rotation buttons in the Text section of the Prism toolbar. The buttons
will rotate your text in 15 degree increments.
Click on the Justify button on the Font toolbar. From the drop-down
menu, you can choose, right-justified, left-justified, or centered.
Whichever you choose will be the default the next time you enter text.
Right click on a text block, and you'll see choices to change the border
thickness and color, and the background. Or double-click on the border of
the text object to bring up the Format Text dialog. Then click the Border
and Fill button to bring up a dialog where you can change border and
background.
To move text:
Move a text object by dragging. You can also stretch the object to change
its shape. This will change how text wraps, but will not change the point
size of the lettering.
You can only move axis and graph titles closer to, or further from, the
graph. They always remain centered. To move a title closer to or further
from the graph, point the mouse near the title until the cursor changes to
a two-headed arrow. Then drag the title.
You can also move axis numbering (or labeling for bar or category
graphs) closer to, or further from, the axis. Click on the numbering to
select. Then use the arrow keys to nudge its position. Or set the distance
between numbers and axis within the Format Axis 428 dialog.
When you drag a text object around the page, it will automatically snap
into alignment with other text objects, with the center of bars (etc.) on
graphs, or with the center of a graph. You can keep dragging, so are not
trapped by the snapping. If the snapping gets in the way of moving the
text object where you want it, hold down the ALT key while dragging and
there will be no snapping.
You can use Microsoft Word® to add text to a Prism graph or layout.
Word gives you more formatting options than are possible with Prism text
tools. You can create bulleted lists, add tables -- anything you can do
with Word. It's a bit more cumbersome, however, because you must open
Word in a separate window to edit your text, rather than editing it directly
on your Prism sheet. This is most often helpful when you have large text
boxes that require a lot of formatting.
To add text to a graph or layout using Word, click the Word button in the
Write section of the Prism toolbar. This will open Word so you can create
your text box. To finish, click on Update in Word File menu, or simply
close Word.
To edit a Word text box in Prism, double click on the text in Prism. This
will open Word. To finish, click on Update in Word File menu, or simply
close Word.
1. Click on the drawing tool in the Drawing toolbar. If you want to draw
several objects, hold Shift while selecting the drawing tool. Then you'll
be able to use it repeatedly without selecting it again. Click any other
Toolbar button to unselect the drawing tool. The tool you used most
recently will show as the default drawing button on the main Prism
toolbar.
2. Position the mouse pointer (now a pencil) at one end of the line or
arrow (or one corner of the box or oval).
3. Hold down the primary mouse button and drag to the other end (or to
the diagonally opposite corner). Hold the control key if you want to
draw a perfect circle or square.
Draw a line, then double click to bring up the dialog above. Change the
thickness, color and style to what you want the default to be for future
lines and arrows. Check the option "Set as default" and OK from the
dialog. Now those settings will be used when you draw new lines or
arrows.
cases, you may prefer to paste link to the original file. To do that, copy
the object and then choose Paste Special from the Edit menu. Then
choose Paste Link. To learn more about the pros and cons of linking and
embedding, see Object linking and embedding 233 . You can also use the
Insert Object command to create a new OLE object.
TIFF (*.tif;*.tiff)
Windows Bitmap (*.bmp)
JPEG (*.jpg;*.jp2;*.j2k;*.jpx)
Macintosh Pict (*.pct;*.pict)
Portable Network Graphics (*.png)
To add a border: Click once to select the image or object in Prism. Then
drop the Change menu and choose Selected Object.
We try to keep Windows and Mac completely compatible, but this is not
possible with images placed on a graph or layout.
If you paste (or import) a WMF or EMF or EMF+ image into Prism 7
Windows, save the file, and open in Prism 7 Mac, that image is gone. We
will try to solve this issue in a future release.
When you work in the Graphs 339 section of your project, the graph itself
as well as text, legends, drawings, and imported images are all referred
to as “objects”. When you are working in the Layout section 478 of your
project, each graph with all of its objects is grouped into a single object.
Images imported onto a page layout are also objects. Images imported
onto a graph are objects in the graph section, but are part of the graph
object in the page layout section.
To select one object:
Point to the object and click the left mouse button. To select a rectangle or oval,
click on its edge. Prism displays knobs on the selected object.
To select several objects:
Hold down the Shift key while selecting the objects one at a time. Or… drag an
imaginary rectangle around all the objects. You've selected all objects entirely
contained within the rectangle.
Moving objects
To rotate an object:
Grab the rotation knob and move your mouse.
Using the Align or Group button in the Arrange section of the Prism
toolbar
For graphs, use the drop-down menu under the Arrange button
For layouts, each task has its own button in the Arrange toolbar.
To align objects:
- Select Align Objects from the Arrange menu or toolbar. Choose horizontal
and vertical orientation, or to align by the X or Y axes.
To position an object by entering distances:
Select Position Object from the Arrange menu or toolbar. Enter the distance
between the upper left corner and the top and left edge of the page.
To lock an object so you can’t accidentally move it:
- Select Lock Object from the Arrange menu or toolbar. When you select a
locked object, the knobs will be gray rather than black, and you will not be able to
move the object or change its size. To unlock the object, select it, then select Lock
Object from the Arrange menu again.
To bring an object on top of overlapping objects.
- Select it, drop the Arrange menu or toolbar and choose Bring to Front. Or use
the toolbar button.
To send an object behind overlapping objects.
- Select the object. Drop the Arrange menu or toolbar and choose Send to
Back. Or use the toolbar button.
To space objects equally:
- Select three or more objects. Choose Distribute Objects from the Arrange
menu or toolbar.
To duplicate one or more objects.
Select the object(s). Drop the Arrange menu or toolbar and choose Duplicate
Object.
To center everything on a page.
To precisely align objects, zoom to a larger magnification. Select a zoom level in the
View menu or click the zoom buttons. Zooming changes the magnification on the
screen, but does not change the size of the graph or layout when printed. Hold Ctrl
and press M to zoom larger or J to zoom smaller.
Rulers
Choose Rulers from the View menu. The rulers appear on top and to the left of the
graph and page layout screens. Show the rulers to align objects precisely. Hide the
rulers to leave more space for the graph.
Alignment guides
Alignment guides are temporary lines you can drop from the ruler to check on the
position and alignment of objects. Click the mouse in either ruler to drop a single
guide (down from the top ruler, or over from the left ruler). Click the corner where
the two rulers intersect to drop both alignment guides at once. The guides will
remain visible only as long as you hold down the mouse button.
Grids
Choose Grids from the View menu. Prism will show a grid that extends from the
minor ticks on the rulers, spaced every 0.1 inches, or 1 millimeter. This grid is
provided to help you align objects. It will not print. Select Grids from the View
menu again to turn it off. Note the distinction between showing this grid, and
choosing a frame with a grid (from the Format Axes dialog) to include a grid on a
graph.
2. Click the Export button on the Prism toolbar to bring up the Export
Graph dialog. Alternatively, drop the File menu and choose Export.
3. Choose the export format 509 , and also choose the location and name of
the exported file.
To export several graphs at once, start from the graph gallery, select two
or more graphs, and then choose to export. With most formats, each
graph or layout will be exported to its own file. With PDF format, you can
choose to export each to its own file or all to one file.
File format
This is the most important choice. If you don't want to think about it, we
suggest using PDF if you can, and use TIFF as a second choice.
Read about the pros and cons of the alternative formats 509 .
The TIF, PNG, PDF, WMF (Windows only) and EMF (Windows only)
formats allow you to choose a transparent (clear) background. With Prism
Mac, but not Prism Windows, you can choose a clear background with EPS
export. If you choose to export as a TIF file with a clear background from
Prism Windows, it will be RGB 32 bits (with no choice of color model). The
.JPG, and BMP formats simply have no concept of a clear background, so
those file formats simply don't allow for the possibility of clear
backgrounds.
Resolution
The resolution is entered in dots per inch (dpi). For perspective, most
computer screens have a resolution of about 100 dpi, and most
professional printing presses (used by journals) use a resolution of 1200
dpi. Laser printers range from 300 to 1200 dpi.
The PDF, EPS, WMF (Windows only) and EMF (Windows only) formats
encode vectors and fonts, so have essentially infinite resolution. If you
export in one of these formats, you will not be asked to enter a
resolution. The other formats offer choices of 100, 300, 600, and 1200
dpi.
If you double the dpi, note that you quadruple the number of pixels in the
image, and thus quadruple the file size.
Color model
If you choose the PDF, EPS or TIF format, you'll need to choose between
RGB and CMYK color models 516 . We recommend that you choose RGB
when possible, as there will be fewer surprises. RGB colors will closely
match what you see on the Prism screen (which internally uses RGB).
Select CMYK only when submitting images to a journal that insists on
CMYK.
With some formats, you can also choose between being able to save
thousands (8 bit) or millions (24 bit) of distinct colors. When you choose
thousands of color, some colors may get changed a bit when saved. When
you save RGB, thousands of colors are stored in an indexed format, but
millions of colors are not. Some journals may specify whether they want
colors expressed in 8 or 24 bits, and whether they prefer indexed or
nonindexed colors.
Prism Mac does not offer the choice of converting text to outlines.
Defaults
Check an option at the bottom of the dialog to lock in your settings of
format and options to use as the default for future exports. To change the
default exporting location, use the File Locations tab of the Preferences 153
dialog.
Set the default for how Prism copies graphs and layouts to the clipboard
in the File & Printer tab of the Preferences dialog. Choose the export
format in the Export dialog, and check an option at the bottom of that
dialog to set the default for future exports.
PDF
Prism's PDF files encode vectors and fonts (not bitmaps), so the resulting
file can be stretched to any size with no loss of quality. Export in PDF
format when you can. Unfortunately, few Windows programs import pdf
images, and few journals accept them. In contrast, PDF is the preferred
format for transferring images between mac programs.
Notes:
· If your graphs are in color, you will need to choose between RBG and
CMYK color model. 516 .
· You don't have to have Adobe Acrobat to export PDF files. Prism creates
PDF files on its own.
· When exporting several graphs or layouts, you can choose to export all
of them to one PDF file. Or you can ask Prism to create a separate PDF
file for each sheet.
· Prism Mac and Windows (new to Prism 7) can encode transparent colors
in PDF export.
TIFF
This is a bitmap format, which means that your graph or layout is turned
into a collection of dots. TIFF files tend to be trouble-free when
submitted to journals. You need to choose resolution, color model, and
size.
Choose between two ways to express colors: RGB (designed for computer
monitors or projectors) or CMYK (designed for printing). If you are
submitting to a journal, follow their instructions. If in doubt, choose RGB.
You can also choose Monochrome files, which are pure black and white
with no grays. If you choose RGB, choose between 256 colors (smaller
files) or millions of colors (larger file, but more accurate color rendition).
Typically, graphs and layouts in Prism projects are much larger than
publication size. If you ask for a graph at 1200 dpi at the actual size in
Prism, you will end up with either a huge file or an out-of-memory error
message. Avoid this problem by specifying the publication size. Let’s say
your graph is 6 inches wide, but you will publish it with a width of 2
inches. By checking the option to reduce the graph to a width of 2 inches,
you will create a file one-ninth as large (one-third the width times one-
third the height).
With TIF format, and no other, Prism lets you choose to compress the file.
The advantage of compressing, of course, is that the file is smaller. The
disadvantage is that compressed TIF files might be less compatible with
other programs. With Prism Windows (but not Mac), choose between two
compression methods. PackBits compression create the smallest file size,
but LZW (Lempel-Ziv-Welch) compression might be more compatible.
Prism Mac does not offer you a choice of compression methods. It uses
LZW compression when you chose Grayscale, CMYK or RGB (millions of
colors, 16 bits), and PackBits compression when you choose Monochrome
or RGB (256 colors, 8 bits).
· Don't use WMF when exporting or pasting bar graphs with fill patterns
to PowerPoint. If you use WMF, the patterns on such a graphs do not
scale and appear of higher density.
· If you paste an EMF image into PowerPoint, the colors may change when
you save then reopen the file. Not with WMF or EMF+.
· Some very old programs do not support EMF. For these choose WMF.
· EMF and EMF+ files exported by Prism 7 encode transparent colors, but
WMF files do not encode transparency. If you choose a transparent color
for your graph, it won't look transparent in the exported file if you
choose WMF.
· EMF+ is a newer format than EMF but both kinds of files use the same
.EMF extension.
EPS
Like PDF, EPS files contain vectors and fonts. The resulting file isn't just
an EPS wrapper around a bitmap file. Prism Windows gives you the choice
of converting text to outlines. If you do that, the image will always look
right (no font files required), but you wont be able to edit text in other
programs. If you include text as text, it can be edited elsewhere but font
files are required. Prism Mac always converts text to outlines.
In theory, the EPS format is the best (tied with PDF). It encodes
everything as vectors and fonts, so can be expanded or shrunk without
any loss of resolution. In fact, submitting EPS files to journals is often
problematic. Fonts and EPS files. 517
JPEG
Most programs that export to the .JPG format give you a choice of
compression. If you compress more, the files will be smaller but the
resulting image won’t be as sharp. Prism gives you no choice. It only
exports to the .JPG format with the least possible compression, making
larger files but sharper images. Any compression would reduce the clarity
of your graphs noticeably.
PNG
The .PNG format is great for posting graphs on web sites. PNG files use
screen resolution, which is fine for displaying on screen but is too little
resolution for submission to printed journals. There are no options of color
depth.
Prism Windows can encode transparent colors in PNG export. Prism Mac
cannot, and any transparent colors in your graphs become solid in the
exported PNG file.
BMP
When you submit figures to journals for publication, you have to decide
which format to use. Each journal has their own rules, and these
sometimes change. But here are some general notes on preparing Prism
graphs for publication.
Exporting as a TIFF file
Exporting as TIFF, our customers tell us, usually works well. Notes:
· In most cases, you will want the page background in Prism to be white,
and to check the option in the Export dialog to include that color in the
file.Transparent backgrounds can sometimes cause problems.
· Choose the color model 516 requested by your journal. Different journal
have different requirements.It seems that RGB is now more popular
than CMYK, so choose RGB if you don't know what your journal wants. If
your graph is entirely black on white, choose the monochrome or gray
scale color model.
· For size, choose to make the exported file the exact width the journal
wants. More. 518
EPS files (encapsulated postscript) files encode the graphs as vectors and
fonts, so have infinite resolution (except for any embedded pictures).
They are more compact and sharper than TIFF or other bitmap files. Even
if the journal you are submitting to doesn't list EPS as an acceptable
format, ask them, since their printed guidelines may be out of date.
The problem of fonts in EPS files
Prism Windows (but not Mac) offers an option to convert fonts to
outlines 517 when exporting to an EPS file.
The default is to save the text as characters, with references to the font.
If you have used only standard fonts, this format lets the journal
producers edit the text, or change the font and size. The only drawback is
that if the file is opened on a computer that does not have the exact fonts
you used, another font will be used and the look of the graph (or layout)
will change. If you use standard fonts, we recommend this approach.
If you check the option to convert all text to outlines or glyphs, the look
of your graph or layout will be preserved on any computer, even one that
does not have the fonts you used. The disadvantage is that the text no
longer is encoded as text, so the people processing your files at the
journal will not be able to edit the text, change the point size or switch to
a different font. We only recommend this approach if you use any unusual
fonts in your figure, and are not sure if the journal computers will contain
those fonts.
Prism Mac always converts fonts to outlines.
Adobe products offer a third way to deal with fonts that combines the
best of the two approaches above, but Adobe has not provided the hooks
our programmers would need to implement this approach. If you have
access to Adobe Illustrator, here is an approach that is likely to work well.
Export the graphs from Prism Windows without converting the text to
outlines. Or try copy and paste. Then export from Illustrator in EPS
format. We do not use Illustrator, so can't give you any detailed
information.
Beware of the term embedded font. If your journal asks for that, clarify
what they mean. Previous versions of Prism and some other programs
used the term "embed" to describe converting fonts to outlines. Adobe,
and some journals, use the term "embed" to the describe the third
approach above, an approach that Prism does not offer.
Choosing a color model
If your graph is entirely black on white, choose the monochrome or
grayscale color model.
If your graph includes color, choose the color model 516 requested by your
journal. Different journal have different requirements. It seems that RGB
is now more popular than CMYK, so choose RGB if your journal has not
specified.
Previewing EPS files
Once you've exported an EPS file from Prism, you'll probably want to see
it to make sure it is correct. With Macs, this is no problem. Just double-
click on the file and it will be opened in Preview, a program that is part of
the OSX system. What actually happens is that Preview converts from
EPS to PDF, and displays the pdf image.
Windows, however, has no built-in program that can view EPS files.
Adobe Illustrator seems to be the standard that many journals use, so
import the EPS files into Illustrator if you can. If you don't have access to
Illustrator, you'll have to find another program. Try double-clicking on the
file to see if you already have a program that can open EPS files. Also try
the free program GhostScript, which works OK. Another option is to
export from Prism as a pdf file, and view that. This works because eps
and pdf files contain mostly the same information, with a different
"wrapper".
The Export dialog requires you to choose between RGB and CMYK color
models when you choose to export a graph (or layout) in the PDF, EPS,
JPG or TIF format.
RGB (Red, Green, Blue) is based on projecting. Red light plus Green light
plus Blue light all projected together create white. Black is encoded as
the absence of any color. Colors encoded as RGB are sometimes called
spot colors.
Even though it uses one more number to encode a color, the CMYK
scheme actually encodes a smaller "color space" than does RGB. That
means that the appearance of some colors will change when converted
from RGB to CMYK. Most noticeably, bright colors in RGB (left side of
figure below) will look duller and darker in CMYK as shown in the right
side of the figure below.
Learn more about the differences between RGB and CMYK here and here.
Prism Windows (but not Mac) offers two choices about how to handle
fonts when exporting to EPS format.
The default choice is to choice saves the text into the EPS file as
characters, with references to the fonts you used. If you have used only
standard fonts, this format lets others edit the text, and change the font
and size. The only drawback is that if the file is opened on a computer
that does not have the exact fonts you used, another font will be used
and the appearance can change drastically.
Prism Windows also offers an option to convert all text to outlines, so that
each letter is described as a set of graphic objects without reference to
any font file. Prism Mac always exports EPS in this format. The advantage
of this approach is that the look of your graph or layout will be preserved
on any computer, even one that does not have the fonts you used. The
disadvantage is that the text no longer is encoded as text. That means it
is impossible to opens the .eps file and edit the text, change the point
size or switch to a different font.
Adobe products offer a third way to deal with fonts that combines the
best of the two approaches above. The font definitions themselves are
included in the file, so the text remains as text, but will render perfectly
on any computer. As far as we can tell, this choice only works within
Adobe products and Adobe has not provided the hooks our programmers
would need to implement this approach.
When exporting a graph to submit to a journal, you will often want to set
the exact width of the graph, and any other objects (text, legends,
drawing, pictures, results tables) on the graph page. There are two ways
to change the size of the graph in Prism:
· On the graph page. Drop the Change menu and choose Resize Graph.
We recommend the second approach above (change the size of the graph
in Prism), as you can then check that nothing is too small. One guideline
(from Cadmus, who handles production for many journals) is to make no
font be smaller than 6pt, and no line be thinner than 1/2 point.
It is easy to find out how large an image is in pixels, and its dpi:
· Mac. Double-click to open in Preview and then drop the Tools menu and
choose Show Inspector.
· Windows: Double-click to open in Photo Viewer, then drop the File menu
and choose Properties. (This works for Windows 7; other versions are
similar.)
When you paste or export a Prism graph into another program, a clear
background will let colors and objects show through, so Clear is not the
same as White.
The problem is that not all export formats support the concept of
transparency, and Prism doesn't support all formats. The JPG, and BMP
formats simply have no concept of a clear background, so those file
formats simply don't allow for the possibility of clear backgrounds.
Note that Prism lets you set two separate colors for the background of a
graph. What Prism calls the background is the color of the 'page'. The
plotting area color is the background color within the plotting area defined
by the axes. Both can be set by right clicking (control-click for mac) on
the graph, or by dropping the color wheel icon. Layouts, of course, only
have the page background property.
Which export formats allow you to choose a clear background?
With Prism Windows, you can choose a clear background when you export
in PDF, EMF or WMF formats. Prism Windows cannot export TIFF images
with clear backgrounds (but Prism Mac can).
With Prism Mac and Windows, you can choose a clear background when
you export to PDF, TIFF, and EPS format.
This choice is in the Exporting options part of the Export Graph dialog,
where you can choose to include a background color or export with a clear
background.
Copying graphs and layouts to the clipboard
Choose on the File & Printer tab of Preferences whether you want Prism to
copy graphs and layouts to the clipboard with the background color you
specified in Prism or with a clear background.
Sending to PowerPoint
Rather than copy and paste, we urge you to use the Send-to-PowerPoint
command (a button at the right end of the toolbar, or a command in the
File menu. The choice you make for sending background color (vs.
When you copy a graph or layout from Prism Mac, two versions are placed
onto the clipboard. The PDF version supports transparent background and
plotting area. Whether the background is clear or not depends on an
option you set in Prism's Preference dialog (bottom of File & Printer tab).
The bitmap version does not support a clear background. If you copy and
paste, which version pastes depends on your version of PowerPoint and
on whether you use Paste Special. Avoid the uncertainty by using the
Send-to-PowerPoint command.
The Send-to-Word command copies the current graph or layout (or a set
of graphs or layouts selected on the gallery 144 ), and pastes it into Word.
Position your insertion point in Word before using this button.
Open the Send to MS Office tab of the Prism Preferences 153 dialog.
When sending to Word, choose whether to display full colors, black and
white, or gray scale. The latter two options simply set a setting within
Word to display and print the graph in black and white or gray. If you
embed or link, and double-click to edit in Prism, you'll see the graph in its
original colors.
Problems sending to Word or PowerPoint? EMF(old) vs. EMF+ vs. WMF
Windows supports three similar formats for describing images using
vectors and fonts. The WMF (Windows Metafile) format is older. The
EMF+ (Enhanced metafile) format is newer, and in theory a bit better.
Prism can also use an older version of EMF. In fact, one format works
better for some programs and the other format for others. It seems
impossible to predict.
Choose which format to place on the clipboard (and to use when sending
graphs to PowerPoint) at the bottom of the File & Printer tab of the
preferences dialog. If you're having problems with pasted graphs looking
funny, try another format.
This page applies only to Prism Windows. Object Linking and Embedding
is not possible on the Mac. We wish it were.
The difference between pasting an object, a link or a picture
There are three ways to copy a Prism graph or layout and paste it into
many Windows programs, including Word or PowerPoint.
Paste a graph or layout as a Prism object:
Prism will paste a picture of the graph, plus a copy of all the data,
analyses and graphical settings needed to create that graph (or layout).
Double-click the object in the other program to edit it within Prism, where
you can edit data, change analysis settings, and format the graph. When
you exit Prism, the revised graph will appear in Word.
A Prism object only includes the data tables and results that are linked to
the graph (or layout) that you pasted. Other parts of the Prism project
are not included.
A picture may be worth a thousand words, but it doesn't let you edit data,
change analysis choices or reformat the graph. When you double-click on
a pasted picture, you cannot edit it within Prism. Instead, you'll open up a
generic graphic or photo editor.
An advantage of this approach (for some) is that people you send the file
to won't be able see your data or edit your graphs. Because there is no
need to store the data and settings, this approach doesn't take up so
much disk space.
Paste a graph or layout as a link
Prism can paste a picture along with a link to the Prism file. When you
double-click, Prism will be launched and open the linked Prism project.
You can then edit the data, analysis settings, or graph in Prism. When
you exit, the revised graph will be updated in Word or PowerPoint.
One advantage of this method (over pasting an object) is that your Word
or PowerPoint files will be smaller. Another advantage (for some) is that it
keeps your data in one place for easy editing. The disadvantage is that
the link really is just a link to a file with a particular name in a particular
place. If the linked file has been deleted, renamed or moved, you won’t
be able to edit it. If you move the Word or PowerPoint to another
computer, you won't be able to open the linked file unless it too is moved.
Choose the Insert Object command from the other program's Insert menu
(or in Office 2007, the right side of the Insert tab). Choose to create
object from a file, and choose your Prism file.
The advantage of this method is that it includes the entire Prism project.
This might be useful if you send a document to a colleague and want to
include all supporting details.
When you copy a graph (or layout), Prism put all the necessary
information on the clipboard so the pasting program can choose if it
wants to paste a picture, an object, or a linked picture. Microsoft Word
and Excel and older versions of PowerPoint paste an embedded object.
PowerPoint 2007 and 2010 are different. By default, they paste a picture.
To override this, drop the clipboard menu, choose Paste Special, and
then choose to paste a Prism object.
Instead of pasting (and letting the program choose), use the Paste
Special command, so you can choose which format is pasted. With most
programs, Paste Special is a choice on the Edit menu. With Office 2007
and 2010, it is a choice on the Paste menu which is on the left side of the
Home tab of the ribbon.
The Paste Special dialog lets you choose to paste an object ("GraphPad
Prism 7 Project Object"), or a Picture ('Enhanced metafile"). You can also
paste a bitmap, which might be useful if you want this Word or
PowerPoint document to look nice on a Mac.
If you check this option to paste a link, the scrolling list gives you no
choices. This option will not be available if you copy a graph from an
unnamed Prism project, because linking requires that the Prism file has a
name.
Choose which format to place on the clipboard (and to use when sending
graphs to PowerPoint) at the bottom of the File & Printer tab of the
How to post
To post a graph or layout to an FTP server, drop the File menu and choose
Post to Web, and choose to send the current sheet, all graphs or all
layouts.
Prism will post the graphs (or layouts) on your FTP server and create an
index page. Immediately after posting the graphs, Prism will launch your
browser, and take you to that page. Copy the URL of the created page
and email it to others so they can view your graphs.
Note that Prism posts only pictures of your graphs or layouts for others to
view. It does not upload the Prism file itself.
Specifying the FTP server
Before you can post to a Web site or FTP server, you must configure
Prism to do so in the Internet tab of the Preferences dialog. This dialog
also will come up the first time you click on the Post to a Web site button
in the Send section of the Prism toolbar.
Enter the ftp URL and your account name and password (if required). You
must also enter the URL to use to view these files (Prism will append the
folder and file name). If your server is behind a firewall that prevents
incoming FTP connections, check the option to use passive mode so Prism
will initiate data connections rather than asking the server to connect
back to your computer. You may want to ask your webmaster for help in
configuring these FTP settings.
17 Printing
The Print section of Prism’s toolbar has two buttons for printing.
-Print the current sheet, or multiple sheets selected in the Gallery 144 ,
using your default Print settings.
-Bring up the Print dialog to choose a printer and which pages to print.
Change the font size to control the size of the print on the page, and thus
the number of pages needed to print a table. To do so, click the Zoom
buttons. Or use the keyboard shortcuts: Use Smaller Font (Ctrl-J), Use
Larger Font (Ctrl-M)
Note these buttons work differently in data and results tables than they
do in graphs and layouts. When you are working with a table, the zoom
buttons change the actual size of the font used, which will then change
the size of the printed table. When working with a graph or layout, these
buttons simply zoom what you see on screen, but do not change the
actual size of the graph or layout when printed.
While you can change graphs and layouts from portrait to landscape
orientation, you cannot o so on data or results tables.
Printing preferences
These options on the File & Printer tab of the Preferences dialog affect
printing of tables.
· Print grid lines on tables -- Check this box to include the horizontal and
vertical lines on data and results tables. Omit it for a cleaner look.
· Print row and column labels -- When printing data tables, print the row
numbers on the left of every page and the column letters on top of
every page.
· Print the sheet name as header -- Print the sheet name, file name, and
date on each page as a header. You'll probably want this header on
pages placed in your notebook, but not on graphs submitted for
publication.
Click the rotate button in the Change toolbar and choose Flip.
Flipping the page orientation does not change the aspect ratio of the
graph. You won't see the change on screen unless you zoom 103 down to
see the entire page.
Use the Flip button in the Change section of the Prism toolbar to change
the orientation of a layout.
Printing preferences
Two options shown on the File & Printer tab of the Preferences dialog
affect printing of graphs and layouts.
· Print the sheet name as header -- Print the sheet name, file name, and
date on each page as a header. You'll probably want this header on
pages placed in your notebook, but not on graphs submitted for
publication.
· Print no grays and no colors -- Convert all colors to black. Don’t print
colors as shades of gray.
Printing posters
Prism can print on very large pages (up to 1.5 by 2 meters) so can create
a poster directly.
Choose the paper size by dropping the File menu and choosing Printer
setup (Windows) or Page Setup (Mac). The details vary with each printer
driver.
Since you can only set the page size once for the entire project, we
suggest that you create your graphs in one project (or several projects)
using a standard page size. Then make your poster in a separate project
created just for that purpose. Set a large page size and then go to the
layout section and create a layout sheet. Arrange the placeholders, and
assign graphs from your other projects.
Open a Gallery 101 page by clicking on the section title in the Navigator
(Data tables, Results, Graphs, etc.). You'll see thumbnails of all the
sheets in an array.
Print selected sheets: Select the sheets you want to print. Then
select the Print command or click on one of the Print buttons on the
Prism toolbar.
Print the gallery page itself: Click on the Print button in the Prism
toolbar and click the Gallery option in the Print dialog. Printing the
gallery might be useful to create a table of contents for a lab notebook,
or to see an overview of your work.
18 Scripts
You can use a Prism script to automate the analysis of a large number of
files, or simulate many data sets for Monte Carlo analyses. Prism's
scripting language lets you open and close Prism files, import data, and
export or print results and graphs. You can run Prism scripts from within
Prism, or launch Prism from another program.
Note that Prism scripts do not tell Prism how to analyze or graph your
data. That is built-in to the Prism project (or template). Use a script to
import data into an existing project and to specify which results and
graphs you want to export or print.
To learn about scripts, first look over the examples we present in the next
section. These examples show you the kinds of things that Prism scripts
can do and give you a feel for how easy it is. Then read on to learn the
details of Prism syntax. The final sections of this chapter explain how you
can launch a Prism script from another program.
Can a Prism script control how Prism analyzes and graphs data?
No. A Prism script can open a Prism project file (or template), import
data, and export results and graphs. But the details of how data are
analyzed and graphed are stored in the Prism file. Note that when you
change data, Prism updates the results and the graphs. This includes
updating analysis constants and paste-linked data. For this reason, Prism
scripts tend to be very short and easy to write.
Does Prism have an API that can be called from other programs?
No. But other programs can launch a Prism script. Read the details about
how this works with the Windows 571 and Mac 569 versions of Prism.
Yes 534 .
Prism can save its project files in two formats. The PZFX format 118
includes all the data and info tables in plain text XML that other programs
can view, edit or replace.
While other programs can't really create Prism files, they can start with a
Prism file you have created, and replace the data. When a user then
opens this file, the results and graphs will be updated.
To get an idea about how this approach works, follow these steps:
1. Create a Prism file, with data, analyses and graphs as you want them.
3. Then open the file with a text editor (notepad). The analysis and
graphing stuff is all encrypted in a way you can't access. But the data
and info are in plain text, structured, XML.
4. Edit the data (and perhaps the info constants), and resave.
5. Double click on the file to open in Prism. You'll see that graphs and
analyses are updated to analyze and graph the new data.
Using this approach, you can effectively (from the user's point of view)
save a Prism file from your application (or instrument). No scripts. No
need to use your software and Prism on the same computer at the same
time. No need to even use the same platform (Win vs. Mac) for Prism and
your program. All you have to do is replace the data portion of a .PZFX
file, and resave it.
The easiest way to learn Prism's scripting language is to follow some examples.
The examples are all for Prism Windows. The only difference on a
Macintosh is that you use a colon instead of a backslash to specify disk
names, folder names, and files. For example, the first line of the first
example would read:
Script example 5. Import data and info constants from the script
file 539
Script example 6. Import data into several tables 540
This is an example of Monte Carlo analysis. We'll simulate 500 data sets
with random variation and fit a dose-response curve to each. This script
creates a table containing the logEC50 and HillSlope from all 500 fits.
Before running this script, you must create a file that has the first data
table formatted for a one-grouping variable data set (no X, No Y
subcolumns). The first results sheet must be one that simulates data with
error. The second results sheet must be a nonlinear regression that fits a
dose-response curve to the data.
Script command Comments
Wtable “Hill Slope”, Write the value from the first Y column
6,1 in row 6 to the next (second) column of
the output table, and label that column
“Hill Slope”.
When the script is done, table 1 will have two columns and 500 rows.
Each row will have the logEC50 and Hill Slope from one of the simulated
experiments.
This example uses the script to import several files onto one data table.
While Example 4 would not be useful when run from Prism, it would be
useful as a way to launch Prism.
For 64 bit Windows: Create a shortcut icon with a target something like
this:
“C:\Program Files (x86)\GraphPad\Prism 7\prism.exe” @”C:\Users\(current user)\Docume
Double-click that icon to launch Prism and run the script doit.pzc.
For Mac: You cannot create an icon that launches Prism running a certain
script. All you need to do with Prism Mac is double-click on the script icon
and that will launch Prism and run the script.
18.3.5 Script example 5. Import data and info constants from the script file
Script command Comments
4.000 115.3119
</DATA>
SetInfo "Dan Brown",3, Set the value of the third info constant.
1
It would not make sense to create this script by hand, but it would be
very useful if created by another program.
Before running this script, create a Prism file with 9 data tables each
linked to a graph, and all the graphs placed on one layout.
Next Loop.
Each script command begins with the command name. Most commands
then have one or several arguments. Separate the command name from
the arguments with a space, and separate the arguments from each
other with spaces or commas. You can use upper and lower case letters
interchangeably, and can add extra spaces to improve readability.
In the rest of this section, required arguments are shown in square
brackets [like this] and optional arguments are shown in angled
Explanation
Example: Description:
Notes:
· Prism numbers the sheets so the first sheet is #1. This seems normal to
everyone except computer programmers, who expect the first one to be
#0.
· Don't forget to include a space (or comma) between the results and the
view. The command "GoTo R4V2" goes to the default (current view) of
results sheet 4, and the view part of the command is ignored.
Column numbering
· Column A, is column 1.
Row numbering
· The Column Title row, if there is one, can be referred to either as row 0
or -1 (to be consistent with column titles).
· The other rows follow the row number shown on the table. So the first
data row is #1. This seems obvious to most, but people familiar with
other computer languages may expect the first row to be #0.
Commands
//single comment line
/*
Several lines
of comments.
*/
ForEach [# of loops]
ForEach [filespec, i. e. ”C:\data\DR*.txt”]
ForEachSheet
Next
Regenerate
Explanation
The heart of many scripts is a loop. You can loop a certain number of
times (for simulations), over all matching files (repeated importing), or
over all sheets in a section (to print or export…). End the loop with a Next
statement. You cannot nest loops.
Starting and ending the loop
There are three ways to start a loop:
For all three kinds of loops, the Next statement ends the loop.
Next command to loop again. See script example 2 536 at the beginning of
this chapter.
The Prism scripting language lets you use these four variables in any
command that exports text. They are mostly useful inside of loops 545 .
· %F. The name of the file. This is useful in ForEach loops that import a
set of files. %F is the file name before the period.
· %S. The name of the current sheet. This is useful when looping over all
sheets in a file. Starting with Prism 7, you may also use this variable
outside of a loop.
· %N. The loop number. This is useful in simulations, which loop a set
number of times.
These variables can be preceded and/or followed by other text. They can
only be used within a script loop. Note there are additional variables 559
you can use when writing text.
DuplicateDataTable <CLEAR>
Explanation
DuplicateDataTable
For the most part, script commands navigate among sheets you have
already created by hand. You can’t create new data tables, new analyses
or new graphs in a script except by duplicating existing ones. When you
duplicate a data table, all you get is a new data table. You can import
data into it, but not analyze or graph within the script.
DuplicateFamily
Duplicates a sheet as well as all linked sheets. Import new data to replace
the existing data, and all analyses and graphs will update.
All new sheets include the prepended text in front of the sheet name.
Enter %F as the prepend text, to use the name of the most recently
imported file.
CreateLayout
Create a new layout with an array of placeholders with the specified
number of rows and columns on a page of the specified orientation
"portrait" or "landscape" (portrait is the default). Assign graphs starting
with the specified one. Specify by rank order in the navigator, with "1" as
the first graph. If you specify GraphsCount, Prism will assign that many
graphs only. If you omit this parameter, Prism will stop when it has
reached either the last placeholder on your layout or the last graph in the
project
SetYReplicates [#]
ClearTable <row>,<col>,<subcol>
DeleteAllDataSets <CLEAR>
Explanation
You must create graphs manually (or by duplicating). You can’t change
most graph properties via scripts, but you can apply a color scheme,
change the limits of an axis, and change the graph or axis titles.
When entering the graph or axis title, you can only enter plain text, and
Prism does not provide tools to enter Greek letters, math symbols,
superscripts or subscripts. But note that Windows lets you enter some
special characters by holding Alt and typing its four digit code. For
example, alt-0181 enters the Greek letter mu.
Each use of the SetAxisLimits command lets you change the range
(minimum and maximum values) or interval of an axis. Each use of the
command only lets you change one of those values. If you want to change
the Y axis (left Y) so it ranges from 100 to 200 with an interval of 10, you
need to use this command three times (in any order):
SetAxisLimits Y top 200
SetAxisLimits Y bottom 100
SetAxisLimits Y interval 10
If the Y axis was split into two segments, you need to specify which
segment to change. The first (left or bottom) segment is '0', and the
second segment is '1'. If you wanted to change the bottom segment, use
these commands. Note that it is not enough to specify the segment
number; you also need to include the word 'segment'.
SetAxisLimits Y segment 0 top 200
SetAxisLimits Y segment 0 bottom 100
SetAxisLimits Y segment 0 interval 10
After you change a graph using a script, linked layouts won't update
unless you write script commands to go to that layout. Here is a script
that flips through all the layouts, so ensures they are all up to date.
Goto L
ForEachSheet
Next
Explanation
All the commands that read and write files use the path set in the
SetPath command unless you specify a different drive or folder in the
other commands. For Windows, use these examples:
SetPath “C:\Users\(current user)
\Documents\007\goldfinger\oddjob”
SetPath “(local drive):\DataFiles\August2005\”
SetPath “\\LabServer\DDrive\PrismFiles"
The last example opens a file on another computer named “LabServer”.
In this case you don’t use the “C:” notation for hard drives, but rather
refer to the name of the drive (“DDrive” in this example.)
On the Mac, use a format like this:
SetPath “Hard Drive:Prism 4:Data:June21”.
If you prefer to specify the path within each individual command, you can
still use the same formats as shown above.
Open [“filename”]
Close
Save <”filename”>
Explanation
If you don’t include a disk and directory, Prism will look in the directory
(folder) specified by the SetPath command. Include the file extension
Unlike the Import command, the Open command by itself will not
automatically open the next file file in a ForEach loop. To do so, use this
syntax: Open %F.PZFX. That opens the next file in the loop (%F) with
the extension .PZFX (which you may want to change to .PZF).
The Close command does not save the file, so issue the Save command
first if you want to keep the changes. Include a file name with the Save
command to create a new file. Omit the file name to save changes using
the same name.
Explanation
The Import command imports data from a text file into a data table. The
ImportLink command replaces data in a data object file using existing
choices for filtering and rearranging
If you omit the filename and are in a ForEach file loop, Prism will use the
current loop’s file. If you include a file name, don’t precede it with a path,
Prism will use the path defined in the SetPath statement if there is one.
the X column, and column 1 is the first Y column. Column -1 contains the
row titles. It is not possible to import subcolumn titles.
If there is only one data object on the table, don’t specify row, column, or
subcolumn. Prism simply uses the only data object on the table.. If there
is more than one data object you need to specify the row, column, and
subcolumn of the upper-left corner of the data object you wish to import
into.
Note that the ImportLink command updates a Prism data object linked to
a text file. If you Paste Link data from Excel, you cannot update the link
using this command. You'd need to write VBA code in Excel to change the
data in Excel.
Explanation
If you create the script file from within another program, you can include
data and info constants within the same file as the script commands.
The InsertData command imports data included within the script file, into
the current data table, starting at the specified position. The data are
included right in the script file enclosed between the tags <Data> and
</Data>. The rest of the script continues right after the </Data> tag.
InsertData
<Data>
1, 2, 4
5, 7, 9
</Data>
The InsertDataLink command imports data included within the script file,
into a data object whose upper left position is at the indicated position.
This works similarly to the ImportLink 551 command.
Copy
Copy [#rows] [#columns] [first row] [first col] <first subcol>
Paste [first row] [first col] <first subcol>
Paste
Explanation
From a data, info, or results table, copy selected values to the clipboard.
If you use the Copy command with nothing else, Prism copies the entire
sheet to the clipboard.
You can also specify which range to copy. Note that #columns includes
subcolumns. So if you want to paste triplicate values into two adjacent
columns, you need to enter #columns as 6 (or 7 if you also paste an X
column).
If you use the Paste command with no arguments, Prism will paste the
entire table starting at the current position of the insertion point.
Copy and paste use the Windows or Mac system clipboard, so will
overwrite any other information copied to the clipboard.
Explanation
SetInfo 5.3, 3
Setinfo "Notebook 25, page 34", 5
Setinfo "Experimenter", 3, 0
Setinfo "Dennis", 3, 1
Setinfo "Notebook 2007H, page 304", 1, 2
The first example sets the third constant equal to 5.3. The second
example sets the fifth constant to “Notebook 25, page 34”. The third
example names (or renames) the third constant to “Experimenter”. The
fourth example assigns the value “Dennis” to that constant. The fifth
example adds some text to the notes portion of the info page.
With the SetValueToInfo command, the value comes from a data, results
or info table. You must first navigate to that table using the Goto script
command. Then specify the row and column (and perhaps subcolumn) of
the cell with the value you want to use, and which row of which info table
you want to ‘paste’ that value into. Example:
SetValueToInfo 2, 4, 14, 0
From the table that a prior Goto script command has taken it to, find the
X value (column 0) in row 14. Put a copy of that value in the fourth row of
the second info table.
Explanation
For the formats that let you set resolution (dots per inch), enter 100,
300, 600 or 1200. You cannot enter other values. The default is 300 dpi.
For EPS or PDF export, set the mode to “RGB” (256 colors, default),
“CMYK” or “Grayscale”
For TIFF export, set the mode to “RGBlow” (256 colors, 8 bits),
“RGBhi” (millions of colors, 24 bits), “CMYK” (default), “GRAY” or “MONO”.
Note that “MONO” means black and white only, with no grays. CMYK and
RGB are two ways of denoting colors.
For TIF export, include the "compress" option to save the file using
"PackBits" compression, which will make the file smaller but perhaps less
compatible. Or use the "LZW" option (Windows only) to compress using
the Lempel-Ziv-Welch algorithm, which doesn't compress quite as well as
PackBits, but may be more compatible with some programs.
For EPS export, choose nofonts to not include fonts. This makes the EPS
file smaller and more compatible, but requires that the same fonts be
present on the computer used to import the file.
If your page has a colored background, you will see that background in
the exported file, unless you include the parameter "CLEAR" to export the
graph or layout with no background color (only allowed for some
formats).
For all export commands, Prism will use the SetPath location if you don't
specify disk and directory.
Explanation
If used from a data or results table, this command exports the entire
table. The format will depend on the extension you give the file name.
Choose from tab-delimited text (.txt), comma-delimited text ( .csv), or
Prism XML ( .xml).
If used from an info table, and the filename is a .txt file, this command
exports only the notes (and not the info constants). If the file name is an
.xml file, then this command exports both info constants and notes. If you
want to export only info constants, use the WSheet command 559 instead.
If you choose the XML format, and include the ALL option, all data and
info tables will be exported.
You can also export the values in a data table or results sheet to a table
created by a script. 557
Explanation
Explanation
Prism offers two methods for writing results from a script. The previous
section explains how to write tables. A more general method, writing to a
text file, is explained here. Generally, you’ll find it easier to write to
tables (previous section) but may find uses for the extra flexibility of
writing to text files.
The first step is to use the OpenOutput command to open (or create) a
file to contain the text the script will output. If you don't precede the file
with disk and directory, Prism will use the path defined in SetPath. If
there is no SetPath command, then Prism assumes that you want to use
the directory that contains the script. If you include a delimiter, say “,” it
will appear between entries. Otherwise the output file will be delimited
with tabs.
If the output file already exists, Prism appends to that file unless you
have included CLEAR.
Results will be tab delimited unless you include an alternative delimiter in
the OpenOutput statement.
Examples:
OpenOutput C:\gp25\output.txt, Clear
OpenOutput C:\data\july97, “,”
You can write (export) data and results to the text file using any of the
commands that begin with W (or by creating a text or CSV table as
described in the next section).
If you end any of the W commands with a semicolon, the next W
command will add text to the same line, following a tab. If you end with
an ampersand (&), the next W command will add text to the same line,
without a tab. If you omit the semicolon or ampersand, Prism inserts the
codes for carriage return and line feed so the next W command will write
to the next line in the file.
The most useful W command is WCell which writes out the contents of
one cell in a data or results table. You need to specify the row and
column of the cell to write. Row 0 refers to the row of column titles, and
row 1 is the first row of data. Column 0 refers to the X column, and
column 1 is the first Y column (column A).
You can also write out entire rows or columns, or the entire sheet, at
once. The commands that begin with WT transpose the data (so each
row becomes a column) before writing to the text file.
WVersion writes the version of Prism you are using, to fully document
your results.
Note that the WTable command is distinct 557 from the set of commands
described here, despite the fact that it begins with the letter W.
· %S Sheet name
· %V Prism version
These commands were introduced in GraphPad Prism 6.04 and 6.0e. The
idea is to specify a block (range) of cells to write to a text file already
specified in an OpenOutput command.
Rows
The parameter for rows can designate a single row, a row down to a blank
row, a range of rows , or a list of rows which can be in any order. You
must have used the OpenOutput script command first to specify where
the results will be written.
What you
How Prism interprets it
enter
Only row 7. Of course, row 7 is the row labeled with a 7 in
7
front of it, so row 0 is the column titles.
“7” Only row 7
“7-“ Start at row 7 and keep going until you hit a blank row
Start at row 7 and end at row 12. Note that the second
“7-12”
number is the ending row, not the number of rows.
“7, 11, 9,
Specified rows in that order.
8”
Columns
What you
How Prism interprets it
enter
Only column 3. The column of row titles (if it exists) is column
3 0. Column 1 is column A, but for some results there are no A,
B, C column labels.
Only column 3. The column with titles is 0, so in the example
“3-3”
above, this is the “significant” column
“3” Only column 3.
Start at column 3 and keep going right until a column is empty
“3-“
(or the end of the table).
“3, 4” Columns 3 and 4
“3-6” Columns 3, 4, 5 and 6.
Columns 3, then 8 then 4. Note that the order in the results
“3, 8, 4” will not match the order in the Prism table. Also note that I
included spaces after the commas which are ignored.
“3,8,4” “
Row label
The parameter row label would must be quoted text. That text will appear
as the first item in every row output. If the WBlock command ends with a
semicolon, then there is no delimeter (no tab, no space, no comma) after
that text. If there is no semicolon, then the row label text is separated
from the next column just like all the columns are separated (tab unless
an alternative is specified in OpenOutput command). If there is no row
label (as it is an optional parameter), then a semicolon at the end of the
command is ignored and has no impact.
Examples
Here are results of one way ANOVA followed by Dunnett's test from Prism
6.04 (the results show two new columns that were not in prior releases):
The results are shown here as tables. In fact, they will be tab-delimited
text (or comma delimeted, if that is set in the OpenOutput command).
Example 1
The results will be as shown below. Here it is a table. In fact it will be tab
delimited text. (The delimeter is set in the OpenOutput script command,
and defaults to tabs).
Psummary B ****
Psummary C ns
Psummary D *
Psummary E ns
Significant? B Yes
Significant? C No
Significant? D Yes
Significant? E No
Example 2
The results will be as shown below. Note the semicolon after the labels,
so in both cases the label and the columnID column run together.
PsummaryB ****
PsummaryC ns
PsummaryD *
PsummaryE ns
Significant?B Yes
Significant?C No
Significant?D Yes
Significant?E No
Example 3
***
Psummary Treated
*
Treated+Anta
Psummary ns
gonist
Treated+Agon
Psummary *
ist
Psummary Solvent Blank ns
Significant? Treated Yes
Treated+Anta
Significant? No
gonist
Treated+Agon
Significant? Yes
ist
Significant? Solvent Blank No
Example 4
Treated+Antagonis
Treated Treated+Agonist Solvent Blanks
t
Yes No Yes No
*** Ns * ns
Example 5
Yes No Yes No
Significant?
Prism file
SendToWord
SendToPP
Explanation
CloseScriptDialog
Print
Exit
Explanation
The Exit command closes all open files, without saving and without
prompting to save, and exits Prism. It works on Windows only.
While testing and developing a Prism script, run it from Prism. Once you
perfect it, you may want to run it from the command line or from another
program as explained later.
Creating a script
4 Enter or edit the script on the left side of the GraphPad Prism Script dialog. Use
. the right side to review script syntax.
5 Repeatedly click ‘1 step’ to step through the script. You can watch the result in
. Prism, as well as observe the script log.
6 Click “12 steps’ if you are feeling lucky.
.
7 Click ‘Run” when you are convinced the script will run without error. The Script
. dialog will minimize, as will Prism, while the script runs.
Editing a script
1 Click the Prism button (or drop file menu) and choose Run Script
.
2 Choose one of the scripts listed in the menu. Or choose Open Script file to open a
. script in an external file.
3 Follow steps 4 through 7 above.
.
If you use OSX 10.4 (“Tiger”) or later, Prism installs an Automator Action
that lets you launch a Prism script. The inputs to this Action is either a file
containing a Prism script, or the text of the script itself. The output is
either the word “Done” or an error message.
Run a script from Microsoft Office using MacScript
MacScript commands let you launch Prism from Word or Excel. These
examples assume you are running Visual Basic for Applications (VBA)
from within Excel.
Launch Prism with this MacScript command.
As soon as Visual Basic has launched Prism, it will continue to the next
statement in the Visual Basic program or macro. It will NOT wait for Prism
to complete before continuing. If you don't take the extra steps described
below, your Visual Basic program will try to read a file containing Prism’s
results before Prism has finished creating that file. To avoid this, make
your Visual Basic program pause until Prism creates a file.
When you write the Prism script, include lines at the end of the script to
create a file that will tell Visual Basic you are done. The example below
creates done.txt.
OpenOutput "done.txt"
WText "done"
CloseOutput
Your Visual Basic code should first delete the file done.txt (if it exists from
a previous run) and then launch Prism. Immediately following, include
these lines that make your program loop until done.txt is created.
Here is a sample AppleScript that launches Prism and tells it to run the
script named DR2 located in the P4scripts folder on a disk named HD.
One way to launch a script is from within Prism (drop the File menu and
choose Run Script). You can also launch a Prism script from outside
Prism. When you launch Prism this way, if the script closes all open Prism
files, then Prism will shut down when the script ends. Otherwise Prism will
remain open.
Launching a Prism script from a desktop icon
The shell command has to specify both the full location of prism.exe and
the script that Prism will launch. Include the @ symbol in front of the
script name.
· Prism will run invisibly. Even when it is done, Prism will not show on
screen.
· As soon as Visual Basic has launched Prism, it will continue to the next
statement in the Visual Basic program or macro. It will NOT wait for
Prism to complete before continuing. If you don't take the extra steps
described below, your Visual Basic program will try to read a file
containing Prism’s results before Prism has finished creating that file. To
avoid this, make your Visual Basic program pause until Prism creates a
file, as explained below.
When you write the Prism script, include lines at the end of the script to
create a file that will tell Visual Basic you are done. The example below
creates done.txt.
OpenOutput "done.txt"
WText "done"
CloseOutput
Your Visual Basic code should first delete the file done.txt (if it exists from
a previous run) and then launch Prism. Immediately following, include
these lines that make your program loop until done.txt is created.
Do Until Dir$(“C:\prism 6\done.txt") > ""
Application.Wait Now + TimeValue("00:00:1")
Loop
The first line in the example above checks whether the file done.txt
exists. Change the path and file name as needed. If the file exists (Prism
is done), Visual Basic continues beyond the loop with any code that
follows. If the file doesn't exist yet, Visual Basic waits 1 second, and then
loops back to test again whether the file exists.
The first line creates a Prism command object. You may give the object
variable any name you like (on the left side of the = sign), but it is
convenient to name it Prism. You don’t have to tell the program the
location of prism.exe. Windows takes care of this automatically via the
registry.
The second line, which is optional, tells Prism which folder to use. This
replaces the need for a SetPath statement within the Prism script. When
running a script, Prism looks for data files and templates in this directory
(folder) first. Place the name of the folder in quotation marks or use a
text variable.
The third line tells Prism to show its progress dialog. Once you have
tested your code, remove this line so Prism will run invisibly (show no
dialogs).
The fourth line launches the Prism script. You must include the full path of
the script file; the directory used in the SetPath message (above) is not
used automatically.
The final two lines cause Prism to exit, and free the memory used by the
Prism object.
Shell command vs. VBA automation
There are two advantages to using OLE automation, rather than the Shell
statement. First, with OLE automation Visual Basic and Prism run
sequentially, with no possibility of Visual Basic trying to obtain results
that haven't been created yet. Second, with OLE automation, your Visual
Basic program can direct Prism to import data from a particular folder
(Prism.SetPath command) without changing the Prism script.
Launching a script from an intranet web page (Windows)
First, create a batch file on your web server that calls (using the Call
command) the program and individual configuration files.
Or this:
Call F:\Prism\Prism.exe @"F:\PRISM\PRISMSCRIPTS\DOIT.PZC"
Next edit your web page to include a hyperlink to the batch file. For
example, HTML link the text "Prism" to the file StartPrism.bat
Although Prism can open Excel files, this is slow. Instead, write a Visual
Basic macro to export selected values as a text file. The macro below,
run from Excel, exports the upper left corner of worksheet 1 (12 rows, 6
columns) to a comma-separated value (CSV) file, which Prism can
import.
Open "C:\data\july97\dr2.csv" For Output As #1
For I = 1 To 12
For J= 1 to 6
Print #1, Worksheets(1).Cells(I,J).Value; ",";
Next
Write #1,
Next
Close #1
Follow this with code to launch Prism, run the Prism script, and perhaps
import results from Prism back to Excel.
While creating a system to run Prism from Excel, you need to edit both
the Excel VBA statements and the Prism PZC script. One way to simplify
your work is to write Visual Basic code that exports the PZC file from
Excel. Then you only have to edit one file. For example, you could use
the following VBA code:
Open "C:\Prism4\auto.pzc" For Output As #1
Print #1, “Shortlog”
Print #1, “SetPath C:\gp\prism25\”
Print #1, “Open generate.pzm”
Print #1, “openOutput fit.txt”
Print #1, “ ForEach 500”
Print #1, “ GoTo R 1”
Print #1, “ Regenerate”
Print #1, “ GoTo R 2”
Print #1, “ WCell 5,1”
Print #1, “ Next”
Close #1
Shell ("C:\program files\prism5\prism.exe @C:
\prism5\auto.pzc")
The first line starts a new file called auto.pzc and designates it as file
number 1. The next 10 lines, export Prism script commands you have
written into that file. The next line closes the script file, which is
launched by the final line shown above. Further lines would import the
results exported by Prism.
Putting all the commands into one file doesn't help conceptually. You still
have to distinguish the VBA code within Excel from the script commands
of Prism. The advantage is that you can do all the editing within one file,
rather than two. The disadvantage is that writing the script is a bit
harder, as you have to place each line in quotations and precede it with
"Print #1,".
The maximum size of a script is not limited by number of lines, but rather
by number of characters. The limit is 64,000 characters. If you hit this
limit, consider removing comments and blank lines. Also consider
removing any parts of the script that are not needed. For example, most
script commands have default parameters and if these are correct, you
don't need to hard code parameters.
Prism offers many choices on its Preference dialog. It allows you to set
defaults for color schemes, fonts, symbols sizes, graph units, and more.
Beyond that, the Prism Preferences file remembers which files you have
recently opened or closed, user-defined equations you have created, color
schemes you have created, autocomplete text, and more. All that is
saved in a preference file that customizes Prism for you.
Windows XP:
[system disk]:\Documents and Settings\[user] \Application data\GraphPad Software\Pris
Mac:
~[local user]\Library\Preferences\ com.GraphPad.Prism.plist
With older versions of OSX, you can locate the file using Finder and drag
it to the trash. Since Lion, the file is invisible. To delete it, launch the
Terminal application, and enter this command:
rm ~/Library/Preferences/com.GraphPad.Prism.plist
When Prism is used in a teaching lab, you want Prism to startup the same
way each time. When a student uses Prism, you don't want preferences
set by other students to be used, as that would make it seem that Prism
acts unpredictably.
[system disk]:\Users\[user]\AppData\Roaming\GraphPad
Software\Prism\7.0
3. Set any preferences as you want them to appear for each user.
4. Create any color schemes you want everyone to have access to, as well
as user-defined equations you want everyone to access.
6. You might wish to write-protect the preference file to make sure it isn't
changed in the future.
The second step is to add /C to the Prism command line 587 , to tell Prism
not to resave the preference file when it exits.
Individual users will still be able to make changes to the Preference file,
and these will be used during that session. But those changes won't get
saved and so won't be seen by the next user.
The method explained above for Windows doesn't work, because there
are no command line switches in OSX. Instead, create and write-protect
the preference file.
~[local user]:Library:Preferences
3. Set any preferences as you want them to appear for each user. Also
create any color schemes or user defined equations you want everyone
to access.
4. Quit Prism.
: ~[local user]:Library:Preferences
Beware that Macs have multiple Library folders. Pick the one within the
local user that each student will log in as, not the overall Library folder.
Repeat for the file that contains the list of autocomplete text entries:
com.GraphPad.Prism.autocomplete.plist
Prism Windows stores preferences in the file Prism6.cfg. Prism Mac stores
preferences in a file named “Prism6 Preferences”. The location of these
files depends on which version of Windows or the Mac OS you use, so
search for them.
If several people use the same Windows computer, each may wish to
maintain their own preferences and equations. To do this, create a
different shortcut icon for each user. In the target portion of the
properties dialog for the shortcut, enter the full path name of the Prism
program followed by “-O” (letter O, not zero) and then the full name of
the configuration file you want to use. If the file doesn't exist, Prism will
create it. For example:
C:\Program Files\PRISM7\PRISM.EXE -OC:\JERRYS\JSPRISM.CNF
If you are running a network license, you may place the Prism6.cfg file
into the program folder on the shared (server) drive. When Prism is first
used on a new computer, it will use the settings in this shared
configuration file. Changes will be saved on the individual user’s hard
drive. This means that all users can begin with the same settings, but
then customize them..
You can ensure that everyone using Prism sees the same templates 138 ,
methods 136 and examples 135 . Create them on one computer, and then
move the files to the other computers you manage. To do that, you need
to know where the files are stored.
Templates
Templates 138 are complete Prism files that you'll see listed on the Open
File tab of the Welcome dialog. You can tell Prism where to save
templates in the File Locations tab of the Preferences dialog. The default
locations are listed below.
Methods
A method file 136 contains a data table and linked analyses and graphs.
When you apply a method, using the Wand (Wizard to Analyze Data),
those analyses and graphs are attached to a new data table.
Examples
An example file is a Prism graph with linked data and analyses. You
clone 135 an example from the Clone A Graph tab of the Welcome dialog.
Saved examples
Prism Windows XP saves examples in:
[system disk]:\Documents and Settings\[User]\Application Data\GraphPad Software\Prism
Shared examples
The Welcome dialog cloning tab offers a separate tab for shared
examples. These are intended to be examples created in a lab to be
shared with everyone in a lab. The only way to show graphs on this tab is
to manually move filed from the saved example folder to the shared
example folder.
Prism Mac also looks for shared examples within the application bundle:
[Prism.app]/Contents/SharedSupport/Examples
When you save the template or method, enter a brief description. If you
want to provide more extensive help to your users, also create a web
(.htm) file that explains use of the template or method. Give it the same
name as the template or method, but with the extension .htm, and place
it in the same folder. For example, if your template is doseresponse.pzt,
create a help file named doseresponse.htm. When someone uses the
template or method, Prism will pop up the dialog with the brief
instructions. This dialog also has a button “Help for this template”. When
your uses click this button, Prism will launch a web browser to display the
.htm file you created.
Prism makes it easy to add your own equations, and these will appear in
the "User defined equations" group of equations. But if you want to share
a set of user-defined equations with others in your lab, you can create a
new group of equations that will appear underneath the ones we provide.
1. Start a new Prism project. Go to preferences, file and printer tab, and
make sure that you are saving in .pzf format and that the "Save
compact" option is not checked.
2. For each equation, either fit a data set with nonlinear regression or
simulate a theoretical curve (plot a function). Reorder the results pages
to the order you want the equations displayed in the equation list.
3. If a user picks your equation and clicks 'Details' they will see a small
preview of the graph that is linked to the equation. Make sure this
graph is easy to comprehend at small sizes.
5. When saving your equation, use a name something like this: "[Smith
Lab] Flow cytometry equations", so it is very clear to anyone using
Prism that these equations did not come from GraphPad (we won't be
able to answer questions about them).
6. Note that each file you create will become a "folder" of equations in
Prism. You'll want each file, therefore, to include several equations.
· Windows: Move (or save) the files to the Equations folder within the
Prism program folder.
· Mac: Find the Prism.app file. Right (or control) click and choose Show
Package Contents. You'll see the "folders" within the package. Drill down
to Contents: SharedSupport:Equations. Move your equation files there.
Note that by changing the prism.app file (bundle) you have changed its
· Both platforms: Please do not delete any equation files that came with
Prism. They are "built in" so should always be available (to prevent
confusion later).
If Prism is running, quit (exit) it, then restart it. You'll now see your list of
equations in the Fit tab of nonlinear regression.
Editing those equations is a bit tricky. Quit Prism, and move the file out of
the Equations folder. Now restart Prism, open the file and edit it. When
done, save your changes and quit Prism. Finally, move the file back to the
Equations folder, and restart Prism.
For every "built-in" equation file, Prism creates a .pze file. These files
contain, essentially, compiled equations and so let Prism load the
equations more quickly. With Windows, you'll find these deep in the
Application Data folder. With Mac, they are deep in the Application
Support folder and also in the Cache folder in the Equations folder in the
Prism bundle.
You don't need to manage these files, and shouldn't try to. Prism will
automatically create and update the needed file when necessary.
If you want to use a local server to host LabArchives files, you can do so.
Details here.
department.
Prism 7 for Windows runs under either the 32- or 64-bit versions of
Windows Vista, 7, 8 or 10. Prism will run in a screen as small as 800 x
540. While it runs fine on 64 bit versions of Windows, it is a 32 bit
program.
Prism tries to connect to the internet with every launch, and must
succeed at least every 30 days or 20 launches (whichever comes first) to
conform the validity of the license. Once it hits those limits, it displays:
"You are currently offline. Prism must connect to the internet at least
once every 30 days to validate your license. Please connect to the
internet and launch Prism again."
If you are getting this message even though you computer is connected
to the internet, perhaps you have a firewall that prevents Prism from
communicating with the internet. Prism uses port 443, which is the same
port that web browsers use when browsing securely (https:), but it is
possible that your IT people configured your firewall so that web browsers
(and email programs...) but not Prism can use that port. If that is the
case, they will need to change the firewall settings.
When Prism starts up, you can give it some special instructions in the
command line. In order to add command line instructions, edit the
Prism.exe start icon. Right click on the Prism desktop icon and In
Properties change the target to add the command after the last quote
mark. Leave a space between the quote mark and the slash, so the target
looks like this:
"C:\Program Files\GraphPad\Prism5\prism.exe" /U
All of the options shown below also work if you use the Shell command to
launch Prism from another program using Visual Basic for Applications
(VBA).
When Prism (Windows) first starts, it tries to check for updates via the
web. If it can't connect, it normally just keeps going. But in very rare
cases, it tries to check for updates and then can't get past this step so
never gets to the Welcome dialog. This is quite rare, but easy to work
around. One way to get Prism to stop check for updates is to use the /U
switch when starting Prism.
Edit the Prism start icon so that the target looks something like this (with
quotes around the path, but not around the /U, and a space beteween the
quotation mark and the slash):
"C:\Program Files\GraphPad\Prism5\prism.exe" /U
The "built-in" equations are actually installed as a set of Prism files. You
can even add your own "built-in" equations. The first time Prism runs, it
compiles these equations to a form (.PZE files) that it can access and
compute very quickly. It should never be necessary, but you can force
Prism to compile all the equations again. To do so, edit the Prism start
icon to add the /E switch so the target looks something like this (with
quotes around the path, but not around the /E, and a space beteween the
quotation mark and the slash):
"C:\Program Files\GraphPad\Prism5\prism.exe" /E
You can create an icon that will open a Prism file and print all pages. Edit
the Prism start icon to add the /P switch so the target looks like this.
Please note space between /P and [file_path]. Prism prints silently,
without Prism appearing in the taskbar.
"C:\Program Files\GraphPad\Prism 6\prism.exe" /P c:\test.pzf
Personal support
Drop the Help menu and choose Personal support command. Prism will
instruct your email program to create a new email, automatically
including your version and serial number, and attaching (if you choose)
the file you are working on.
If the Personal support command does not work on your computer (or you
use a Mac) email us at [email protected] .
Please send:
· The version of Prism you are running and if you are using Prism for
Windows or for Mac. To find the version number, drop the Help menu
(Windows) or Prism menu (Mac OS X) and choose About Prism.
Windows versions have two digits after the decimal point (i. e. 7.01).
Mac versions have a single digit after the decimal followed by a letter (i.
e. 7.0b). Send us the actual version number (which you can see on the
Welcome dialog when you start Prism. Don't say something ambiguous
like "Prism 7" or "the latest version".
· Send screen shots when that would help us understand the problem.
· Don't send raw data in the text of the email and expect us to paste this
into Prism, reproduce what we think you might have done, and then
comment on what we see which may or may not be what you see.
Similarly don't send data in Excel files.
No phone support
· Our customers live all around the world, but we're only here during
business hours on the West coast.
· When we return your phone call, you might be at lunch, away from your
desk, or gone for the day. If we leave a message, it might start a round
of telephone tag that could last for days.
Citing Prism
When citing analyses performed by the program, include the name of the
analysis, and Prism version number (including the second number after
the decimal point). Use this example as a guide:
To find the full version number, pull down the Help menu (Windows) or
the Prism menu (Mac OS X). Then choose About Prism.
Include the name of the page, the date you accessed it and the URL. Use
this example as a guide:
"What you can conclude when two error bars overlap (or don't)?"
GraphPad Software, Inc. Accessed 5 November 2008.
http://www.graphpad.com/faq/viewfaq.cfm?faq=1362
Include the name of the calculator and the date you used it. Here is an
example:
GraphPad if you violate the terms of this Agreement. A personal subscription may be
installed and activated twice, on computers used solely by the subscriber. Upgrades to
newer versions during the term of the license are included. You acknowledge that (i)
your right to install and use the Software is limited to the term of your subscription, (ii)
the Software is designed to cease to function upon expiration of the term of your
subscription if the subscription is not renewed, and (iii) you may generally renew your
subscription, but the terms upon which the Software may be licensed upon a renewal
are subject to change by GraphPad at its sole discretion. Any attempt to defeat the
time-control disabling function in the Software is a material breach of this Agreement
and a violation of intellectual property law.
(d) A group subscription license lasts for the period of time for which you subscribed as
indicated on your invoice for the Software, and may only be earlier terminated by
GraphPad if you violate the terms of this Agreement. Your subscription is to activate
the software on an authorized number of computers. Upgrades to newer versions
during the term of the subscription are included. You acknowledge that (i) your right to
install and use the Software is limited to the term of your subscription, (ii) the Software
is designed to cease to function upon expiration of the term of your subscription if the
subscription is not renewed, and (iii) you may generally renew your subscription, but
the terms upon which the Software may be licensed upon a renewal are subject to
change by GraphPad at its sole discretion. Any attempt to defeat the time-control
disabling function in the Software is a material breach of this Agreement and a violation
of intellectual property law.
USE OF PRIOR VERSIONS IF YOU UPGRADED. If you purchased your license as an
upgrade from a prior version, you not transfer your license for the prior version to
another computer or to anyone else.
INSTALLATION AND USAGE. You may install and use the Software on up to the
Authorized Number of Computers. THIS IS NOT A CONCURRENT USER LICENSE. The
Authorized Number of Computers includes each Computer that at any time is or has
been given access to the Software, regardless of whether such Computer is running the
Software at a particular point in time.
AUDITS. To ensure compliance with this Agreement, you agree that upon reasonable
notice, GraphPad or GraphPad’s authorized representative shall have the right to
inspect and audit your use of the Software. Any such inspection or audit shall be
conducted during regular business hours at your facilities or electronically, either by
GraphPad or by representatives authorized by GraphPad for this purpose. If such
inspections or audits disclose that you have installed or permitted access to the
Software in a manner that is not permitted under this Agreement, then (i) you are liable
to pay for any unpaid license fees as well as the reasonable costs of the audit; and (ii)
the license grant with respect to the Software will terminate immediately. Any
information obtained by GraphPad or GraphPad’s authorized representative during the
course of such inspection and audit will be used and disclosed by GraphPad solely for
purposes of such inspection and audit and for enforcement of GraphPad’s rights under
this Agreement and applicable law, unless other uses or disclosures are required under
applicable law. Nothing in this Section shall be deemed to limit any legal or equitable
remedies available to GraphPad for violation of this Agreement or applicable law.
MANDATORY ACTIVATION AND LICENSE CONFIRMATION. Activation associates the
use of the software with a specific computer. You must activate Prism before using it
the first time. Prism will confirm the activation each time you launch it. During
activation (and activation confirmation), the software will send information about the
software and the device to GraphPad. This information includes the version, the serial
number, the Internet protocol address of the device, and information derived from the
hardware configuration of the device. BY USING THE SOFTWARE, YOU CONSENT TO
THE TRANSMISSION OF THIS INFORMATION. UNLESS THE SOFTWARE IS ACTIVATED,
YOU HAVE NO RIGHT TO USE THE SOFTWARE. This is to prevent its unlicensed use.
YOU ARE NOT PERMITTED TO BYPASS OR CIRCUMVENT ACTIVATION. You must
activate every computer from which Prism is used.
GENERAL RESTRICTIONS. You agree and acknowledge that, unless enforcement is
prohibited by applicable law (and then only to the extent prohibited by applicable law),
the following actions are expressly prohibited:
(a) You may not (and may not permit any third party to) decompile, disassemble or
reverse engineer the Software.
(b) You may not modify, translate, adapt, arrange or create derivative works of the
Software or the User Documentation.
(c) You may not sell, transfer, rent, lease, loan, or otherwise distribute all or any portion
of the Software, the User Documentation or any other rights granted to you in this
Agreement.
(d) Except as otherwise expressly provided in this Agreement, you may not allow access
to the Software or the User Documentation over the Internet, including, without
limitation, in connection with a web hosting, commercial time sharing, service bureau,
or similar service.
(e) You may not remove, alter or obscure any copyright or other proprietary notices,
labels or marks from the Software or the User Documentation.
(f) You may not otherwise install, access or otherwise use or copy the Software or the
User Documentation other than in strict compliance with the terms of this Agreement.
(g)You may not use the Software in medical diagnosis or treatment, or in applications
or systems where the Software's failure to perform can reasonably be expected to
result in significant physical injury, property damage, or loss of life.
(h) You may not attempt to to reduce the number of licenses that you need by use
multiplexing or pooling (hardware or software to pool connections, reroute
information, or reduce the number of devices or users that directly access or use the
software).
(i) You may not install the Software on network servers, and may not use any kind of
remote access software to control and view the software running on one computer from
another computer or terminal.
ALL RIGHTS RESERVED. Title to and ownership of the Software, the User
Documentation and all related intellectual property are retained by GraphPad. The
Software is confidential and the Software and the User Documentation are copyrighted.
The Software and the User Documentation are licensed to you, not sold. All rights of
every kind that are not expressly granted to you in this Agreement are entirely and
exclusively reserved to GraphPad.
LIMITED WARRANTY. GraphPad warrants that, as of the date on which the Software is
delivered to you and for thirty (30) days thereafter, the Software will provide the
features and functions generally described in the User Documentation and that the
media on which the Software is furnished, if any, will be free from defects in materials
and workmanship under normal use. Except for the foregoing, the Software and the
User Documentation are provided "AS IS." All warranty claims must be made during
such thirty (30) day period. GraphPad’s entire liability and your exclusive remedy will
be, at GraphPad’s option, to attempt to correct or work around errors, to replace the
defective media on which the Software is furnished, if any, or to refund the license fees,
if any, paid by you and terminate this Agreement.
DISCLAIMER OF WARRANTY. EXCEPT AS SPECIFIED IN THIS AGREEMENT, GRAPHPAD
MAKES, AND YOU RECEIVE, NO WARRANTIES, EXPRESS OR IMPLIED WITH RESPECT
TO THE SOFTWARE AND THE USER DOCUMENTATION. ANY STATEMENTS OR
REPRESENTATIONS ABOUT THE SOFTWARE AND ITS FUNCTIONALITY IN THE USER
DOCUMENTATION OR ANY COMMUNICATION WITH YOU CONSTITUTE TECHNICAL
INFORMATION AND NOT AN EXPRESS WARRANTY OR GUARANTEE. IN ADDITION,
GRAPHPAD SPECIFICALLY DISCLAIMS ANY OTHER WARRANTY INCLUDING, WITHOUT
LIMITATION, THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
PARTICULAR PURPOSE, AND NONINFRINGEMENT. WITHOUT LIMITING THE
FOREGOING, GRAPHPAD DOES NOT WARRANT THAT THE OPERATION OF THE
SOFTWARE WILL BE UNINTERRUPTED OR ERROR FREE. SOME JURISDICTIONS DO
NOT ALLOW LIMITATIONS ON IMPLIED WARRANTIES, SO THE FOREGOING MAY NOT
APPLY TO YOU. THIS LIMITED WARRANTY GIVES YOU SPECIFIC RIGHTS. OTHER
RIGHTS, WHICH VARY FROM JURISDICTION TO JURISDICTION, MAY APPLY TO YOU.
LIMITATION OF LIABILITY. IN NO EVENT SHALL GRAPHPAD HAVE ANY LIABILITY FOR
ANY INCIDENTAL, SPECIAL, INDIRECT, OR CONSEQUENTIAL DAMAGES; LOSS OF
PROFITS, REVENUE, OR DATA; BUSINESS INTERRUPTION, OR COST OF COVER. IN
ADDITION, IN NO EVENT SHALL THE LIABILITY OF GRAPHPAD FOR ANY DAMAGES
ARISING OUT OF OR IN CONNECTION WITH THE SOFTWARE, USER DOCUMENTATION,
OR THIS AGREEMENT EXCEED THE AMOUNT PAID OR PAYABLE BY YOU FOR THE
SOFTWARE. THE LIMITATIONS OF LIABILITY IN THIS SECTION SHALL APPLY TO THE
20.8 Training
-X-
X and Y, transposing 300
X values, regular sequence 185
XLS files, importing 239
XML 118
XY data tables 165
-Y-
Y axis, adding one on right 327
Yellow highlighter 151
-Z-
Zooming 103, 500