STATGRAPHICS – Rev.

9/16/2013

Multiple Regression

Summary ......................................................................................................................................... 1
Data Input........................................................................................................................................ 2
Analysis Summary .......................................................................................................................... 3
Analysis Options ............................................................................................................................. 5
Component Effects Plot ................................................................................................................ 10
Conditional Sums of Squares ........................................................................................................ 11
Observed versus Predicted ............................................................................................................ 12
Residual Plots................................................................................................................................ 12
Unusual Residuals ......................................................................................................................... 15
Influential Points ........................................................................................................................... 15
Confidence Intervals ..................................................................................................................... 16
Correlation Matrix ........................................................................................................................ 16
Reports .......................................................................................................................................... 17
Interval Plots ................................................................................................................................. 18
Autocorrelated Data ...................................................................................................................... 20
Save Results .................................................................................................................................. 23

Summary
The Multiple Regression procedure is designed to construct a statistical model describing the
impact of a two or more quantitative factors X on a dependent variable Y. The procedure
includes an option to perform a stepwise regression, in which a subset of the X variables is
selected. The fitted model may be used to make predictions, including confidence limits and/or
prediction limits. Residuals may also be plotted and influential observations identified.

The procedure contains additional options for transforming the data using either a Box-Cox or
Cochrane-Orcutt transformation. The first option is useful for stabilizing the variability of the
data, while the second is useful for handling time series data in which the residuals exhibit serial
correlation.

Sample StatFolio: multiple reg.sgp

Sample Data
The file 93cars.sgd contains information on 26 variables for n = 93 models of automobiles, taken
from Lock (1993). The table below shows a partial list of 4 columns from that file:

Make Model MPG Highway Weight Horsepower Wheelbase Drivetrain

Acura Integra 31 2705 140 102 front
Acura Legend 25 3560 200 115 front
Audi 90 26 3375 172 102 front
Audi 100 26 3405 172 106 front
BMW 535i 30 3640 208 109 rear
Buick Century 31 2880 110 105 front

A model is desired that can predict MPG Highway from Weight, Horsepower, Wheelbase, and
Drivetrain.
 2013 by StatPoint Technologies, Inc. Multiple Regression - 1
 STATGRAPHICS – Rev. 9/16/2013

Data Input
The data input dialog box requests the names of the columns containing the dependent variable
Y and the independent variables X:

 Y: numeric column containing the n observations for the dependent variable Y.

 X: numeric columns containing the n values for the independent variables X. Either column
names or STATGRAPHICS expressions may be entered.

 Select: subset selection.

 Weight: an optional numeric column containing weights to be applied to the squared

residuals when performing a weighted least squares fit.

In the example, note the use of the expression Weight^2 to add a second-order term involving the
weight of the vehicle. This was added after examining an X-Y plot that showed significant
curvature with respect to Weight. The categorical factor Drivetrain has also be introduced into
the model through the Boolean expression Drivetrain=”front”, which sets up an indicator
variable that takes the value 1 if true and 0 if false. The model to be fit thus takes the form:

MPG Highway = 0 + 1Weight + 2Weight2 + 3Horsepower + 4Wheelbase + 5X5 (1)

where

 2013 by StatPoint Technologies, Inc. Multiple Regression - 2

 STATGRAPHICS – Rev. 9/16/2013
1 Drivetrain  front
X 5   if (2)
0 Drivetrain  rear

Analysis Summary
The Analysis Summary shows information about the fitted model.

Multiple Regression - MPG Highway

Dependent variable: MPG Highway (miles per gallon in highway driving)
Independent variables:
Weight (pounds)
Weight^2
Horsepower (maximum)
Wheelbase (inches)
Drive Train="front"
Number of observations: 93

Standard T
Parameter Estimate Error Statistic P-Value
CONSTANT 49.8458 10.5262 4.73539 0.0000
Weight -0.0273685 0.00530942 -5.1547 0.0000
Weight^2 0.00000261405 8.383E-7 3.11827 0.0025
Horsepower 0.0145764 0.009668 1.50769 0.1353
Wheelbase 0.338687 0.103479 3.273 0.0015
Drive Train="front" 0.632343 0.73879 0.855918 0.3944

Analysis of Variance
Source Sum of Squares Df Mean Square F-Ratio P-Value
Model 1902.18 5 380.435 46.41 0.0000
Residual 713.136 87 8.19696
Total (Corr.) 2615.31 92

R-squared = 72.7323 percent

R-squared (adjusted for d.f.) = 71.1652 percent
Standard Error of Est. = 2.86303
Mean absolute error = 2.13575
Durbin-Watson statistic = 1.685 (P=0.0601)
Lag 1 residual autocorrelation = 0.156111

Included in the output are:

 Variables: identification of the dependent variable. The general form of the model is

Y = 0 + 1X1 + 2X2 + … + kXk (3)

where k is the number of independent variables.

 Coefficients: the estimated coefficients, standard errors, t-statistics, and P values. The
estimates of the model coefficients can be used to write the fitted equation, which in the
example is

MPG Highway = 49.8458 - 0.0273685*Weight + 0.00000261405*Weight2

+ 0.0145764*Horsepower + 0.338687*Wheelbase
+ 0.632343*Drive Train="front" (4)

 2013 by StatPoint Technologies, Inc. Multiple Regression - 3

 STATGRAPHICS – Rev. 9/16/2013
The t-statistic tests the null hypothesis that the corresponding model parameter equals 0,
based on the Type 3 sums of squares (the extra sums of squares attributable to each variable
if it is entered into the model last). Large P-Values (greater than or equal to 0.05 if operating
at the 5% significance level) indicate that a term can be dropped without significantly
degrading the model provided all of the other variables remain in the model. In the current
case, both Horsepower and Drivetrain are not significant. Thus, either variable (but not
necessarily both) could be dropped from the model without hurting its predictive power
significantly.

 Analysis of Variance: decomposition of the variability of the dependent variable Y into a

model sum of squares and a residual or error sum of squares. Of particular interest is the F-
test and its associated P-value, which tests the statistical significance of the fitted model. A
small P-Value (less than 0.05 if operating at the 5% significance level) indicates that a
significant relationship of the form specified exists between Y and the independent variables.
In the sample data, the model is highly significant.

 Statistics: summary statistics for the fitted model, including:

R-squared - represents the percentage of the variability in Y which has been explained by the
fitted regression model, ranging from 0% to 100%. For the sample data, the regression has
accounted for about 72.7% of the variability in the miles per gallon. The remaining 27.3% is
attributable to deviations from the model, which may be due to other factors, to measurement
error, or to a failure of the current model to fit the data adequately.

Adjusted R-Squared – the R-squared statistic, adjusted for the number of coefficients in the
model. This value is often used to compare models with different numbers of coefficients.

Standard Error of Est. – the estimated standard deviation of the residuals (the deviations
around the model). This value is used to create prediction limits for new observations.

Mean Absolute Error – the average absolute value of the residuals.

Durbin-Watson Statistic – a measure of serial correlation in the residuals. If the residuals

vary randomly, this value should be close to 2. A small P-value indicates a non-random
pattern in the residuals. For data recorded over time, a small P-value could indicate that some
trend over time has not been accounted for. In the current example, the P-value is greater
than 0.05, so there is not a significant correlation at the 5% significance level.

Lag 1 Residual Autocorrelation – the estimated correlation between consecutive residuals, on

a scale of –1 to 1. Values far from 0 indicate that significant structure remains unaccounted
for by the model.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 4

 STATGRAPHICS – Rev. 9/16/2013

Analysis Options

 Fitting Procedure – specifies the method used to fit the regression model. The options are:

o Ordinary Least Squares – fits a model using all of the independent variables.

o Forward Stepwise Selection – performs a forward stepwise regression. Beginning with

a model that includes only a constant, the procedure brings in variables one at a time
provided that they will be statistically significant once added. Variables may also be
removed at later steps if they are no longer statistically significant.

o Backward Stepwise Selection – performs a backward stepwise regression. Beginning

with a model that includes all variables, the procedure removes variables one at a time
if they are not statistically significant. Removed variables may also be added to the
model at later steps if they become statistically significant.

o Box-Cox Optimization – fits a model involving all of the independent variables. The
dependent variable, however, is modified by raising it to a power. The method of Box
and Cox is used to determine the optimum power. Box-Cox transformations are a way
of dealing with situations in which the deviations from the regression model do not
have a constant variance.

o Cochrane-Orcutt Optimization – fits a model involving all of the independent

variables. However, the least squares procedure is modified to allow for autocorrelation
between successive residuals. The value of the lag 1 autocorrelation coefficient is
determined using the method of Cochrane and Orcutt. The Cochrane-Orcutt
transformation is a method for dealing with situations in which the model residuals are
not independent.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 5

 STATGRAPHICS – Rev. 9/16/2013
 Constant in model – If this option is not checked, the constant term 0 will be omitted from
the model. Removing the constant term allows for regression through the origin.

 Power – Specifies the power to which the dependent variable is raised. The default value of
1.0 implies no power transformation.

 Addend – Specifies an amount that is added to the dependent variable before it is raised to
the specified power.

 Autocorrelation – Specifies the lag 1 autocorrelation of the residuals. The default value of
0.0 implies that the residuals are assumed to be independent. If the Cochrane-Orcutt
procedure is used, this value provides the starting value for the procedure.

 Selection Criterion – If performing a forward or backward stepwise regression, this

specifies whether variable entry and removal should be based on the F-ratio or its associated
P-value.
=
 F-to-Enter - In a stepwise regression, variables will be entered into the model at a given step
if their F values are greater than or equal to the F-to-Enter value specified.

 F-to-Remove - In a stepwise regression, variables will be removed from the model at a given
step if their F values are less than the F-to-Remove value specified.

 P-to-Enter - In a stepwise regression, variables will be entered into the model at a given step
if their P values are less than or equal to the P-to-Enter value specified.

 P-to-Remove - In a stepwise regression, variables will be removed from the model at a given
step if their P values are greater than the P-to-Remove value specified.

 Max Steps – maximum number of steps permitted when doing a stepwise regression.

 Display – whether to display the results at each step when doing a stepwise regression.

 Force in… - after a model has been fit using stepwise selection, push this button to force one
or more variables that were not selected to be included in the model.

 Force out… - after a model has been fit using any method, push this button to forcibly
remove one or more variables from the model.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 6

 STATGRAPHICS – Rev. 9/16/2013
Example – Stepwise Regression
The model fit to the automobile data showed 2 insignificant variables. To remove them from the
model, Analysis Options may be used to perform either a forward stepwise selection or a
backward stepwise selection.

 Forward selection – Begins with a model involving only a constant term and enters
one variable at a time based on its statistical significance if added to the current
model. At each step, the algorithm brings into the model the variable that will be the
most statistically significant if entered. Selection of variables is based on either an F-
to-enter test or a P-to-enter test In the former case, as long as the most significant
variable has an F value greater or equal to that specified on the Analysis Summary
dialog box, it will be brought into the model. When no variable has a large enough F
value, variable selection stops. In addition, variables brought into the model early in
the procedure may be removed later if their F value falls below the F-to-remove
criterion.

 Backward selection – Begins with a model involving all the variables specified on
the data input dialog box and removes one variable at a time based on its statistical
significance in the current model. At each step, the algorithm removes from the
model the variable that is the least statistically significant. Removal of variables is
based on either an F-to-remove test or a P-to-enter test. In the former case, if the
least significant variable has an F value less than that specified on the Analysis
Summary dialog box, it will be removed from the model. When all remaining
variables have large F values, the procedure stops. In addition, variables removed
from the model early in the procedure may be re-entered later if their F values reach
the F-to-enter criterion.

In the current example, a backwards selection procedure yields the following:

Stepwise regression
Method: backward selection
F-to-enter: 4.0
F-to-remove: 4.0

Step 0:
5 variables in the model. 87 d.f. for error.
R-squared = 72.73% Adjusted R-squared = 71.17% MSE = 8.19696

Step 1:
Removing variable Drive Train="front" with F-to-remove =0.732595
4 variables in the model. 88 d.f. for error.
R-squared = 72.50% Adjusted R-squared = 71.25% MSE = 8.17206

Step 2:
Removing variable Horsepower with F-to-remove =2.22011
3 variables in the model. 89 d.f. for error.
R-squared = 71.81% Adjusted R-squared = 70.86% MSE = 8.28409

Final model selected.

In the first step, Drivetrain is removed since it is the least significant. At the second step,
Horsepower is removed. The algorithm then stops, since all remaining variables have F-to-
remove values greater than 4, and all previously removed variables have F-to-enter values less
than 4.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 7

 STATGRAPHICS – Rev. 9/16/2013
The reduced model is summarized below:

Multiple Regression - MPG Highway

Dependent variable: MPG Highway
Independent variables:
Weight
Weight^2
Horsepower
Wheelbase
Drive Train="front"
Number of observations: 93

Standard T
Parameter Estimate Error Statistic P-Value
CONSTANT 51.8628 10.2179 5.07569 0.0000
Weight -0.0245435 0.00506191 -4.84867 0.0000
Weight^2 0.00000236841 8.25606E-7 2.86869 0.0051
Wheelbase 0.28345 0.0899993 3.14947 0.0022

Analysis of Variance
Source Sum of Squares Df Mean Square F-Ratio P-Value
Model 1878.03 3 626.009 75.57 0.0000
Residual 737.284 89 8.28409
Total (Corr.) 2615.31 92

R-squared = 71.809 percent

R-squared (adjusted for d.f.) = 70.8587 percent
Standard Error of Est. = 2.87821
Mean absolute error = 2.19976
Durbin-Watson statistic = 1.67296 (P=0.0558)
Lag 1 residual autocorrelation = 0.162386

NOTE: from here forward in this document, the results will be based on the reduced model
without Drivetrain or Wheelbase.

Example – Box-Cox Transformation

If it is suspected that the variability of Y changes as its level changes, it is useful to consider
performing a transformation on Y. The Box-Cox transformations are of the general form

Y   Y  2  1

(5)

in which the data is raised to a power 1 after shifting it a certain amount 2. Often, the shift
parameter 2 is set equal to 0. This class includes square roots, logarithms, reciprocals, and other
common transformations, depending on the power. Examples include:

Power Transformation Description

 = 2 Y  Y 2 square
 = 1 Y  Y untransformed data
 = 0.5 Y  Y square root
 = 0.333 Y  3 Y cube root
 = 0 Y   ln(Y ) logarithm
 = -0.5 1 inverse square root
Y 
Y
 = -1 1 reciprocal
Y 
Y
 2013 by StatPoint Technologies, Inc. Multiple Regression - 8
 STATGRAPHICS – Rev. 9/16/2013

Using Analysis Options, you can specify the values for 1 or 2, or specify just 2 and have the
program find an optimal value for 1 using the methods proposed by Box and Cox (1964).

For the sample data, a plot of the residuals versus predicted values does show some change in
variability as the predicted value changes:

Residual Plot
4.3
Studentized residual

2.3

0.3

-1.7

-3.7
20 25 30 35 40 45 50
predicted MPG Highway

The smaller cars tend to be somewhat more variable than the larger cars. Asking the program to
optimize the Box-Cox transformation yields:

Multiple Regression - MPG Highway

Dependent variable: MPG Highway
Independent variables:
Weight
Weight^2
Wheelbase
Number of observations: 93

Box-Cox transformation applied: power = -0.440625 shift = 0.0

Standard T
Parameter Estimate Error Statistic P-Value
CONSTANT 230.703 9.37335 24.6126 0.0000
Weight -0.0129299 0.00464353 -2.78451 0.0065
Weight^2 6.18885E-7 7.57367E-7 0.817153 0.4160
Wheelbase 0.229684 0.0825606 2.782 0.0066

Analysis of Variance
Source Sum of Squares Df Mean Square F-Ratio P-Value
Model 1568.28 3 522.761 74.99 0.0000
Residual 620.444 89 6.97128
Total (Corr.) 2188.73 92

R-squared = 71.6528 percent

R-squared (adjusted for d.f.) = 70.6972 percent
Standard Error of Est. = 2.64032
Mean absolute error = 2.08197
Durbin-Watson statistic = 1.70034 (P=0.0727)
Lag 1 residual autocorrelation = 0.148826

Apparently, an inverse square root of MPG Highway improves the properties of the residuals, as
illustrated in the new residual plot:

 2013 by StatPoint Technologies, Inc. Multiple Regression - 9

 STATGRAPHICS – Rev. 9/16/2013

Residual Plot
4.4
Studentized residual
2.4

0.4

-1.6

-3.6
20 25 30 35 40 45 50
predicted MPG Highway

Note: some caution is necessary here, however, since the transformation may be heavily
influenced by one or two outliers. To simplify the discussion that follows, however, the rest of
this document will work with the untransformed model.

Component Effects Plot

Plotting a multiple regression model is not as easy as plotting a simple regression model, since
the space of the X variables is multi-dimensional. One useful way to illustrate the results is
through the Component Effects Plot, which plots of the portion of the fitted regression model
corresponding to any single variable.

Component+Residual Plot for MPG Highway

12
8
component effect

4
0
-4
-8
-12
90 95 100 105 110 115 120
Wheelbase

The line on the plot is defined by

̂ j x j  x j  (6)

where ˆ j is the estimated regression coefficient for variable j, xj represents the value of variable
j as plotted on the horizontal axis, and x j is the average value of the selected independent
variable amongst the n observations used to fit the model. You can judge the importance of a
factor by noting how much the component effect changes over the range of the selected variable.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 10

 STATGRAPHICS – Rev. 9/16/2013
For example, as Wheelbase changes from 90 to 120, the component effect changes from about –4
to +4. This implies that differences in Wheelbase account for a swing of about 8 miles per gallon.

The points on the above plot represent each of the n = 93 automobiles in the dataset. The vertical
positions are equal to the component effect plus the residual from the fitted model. This allows
you to gage the relative importance of a factor compared to the residuals. In the above plot, some
of the residuals are as large if not larger than the effect of Wheelbase, indicating that other
important factors may be missing from the model.

Pane Options

 Plot versus: the factor used to define the component effect.

Conditional Sums of Squares

The Conditional Sums of Squares pane displays a table showing the statistical significance of
each coefficient in the model as it added to the fit:

Further ANOVA for Variables in the Order Fitted

Source Sum of Squares Df Mean Square F-Ratio P-Value
Weight 1718.7 1 1718.7 207.47 0.0000
Weight^2 77.1615 1 77.1615 9.31 0.0030
Wheelbase 82.1713 1 82.1713 9.92 0.0022
Model 1878.03 3

The table decomposes the model sum of squares SSR into contributions due to each coefficient
by showing the increase in SSR as each term is added to the model. These sums of squares are
often called Type I sums of squares. The F-Ratios compare the mean square for each term to the
MSE of the fitted model. These sums of squares are useful when fitting polynomial models, as
discussed in the Polynomial Regression documentation.

In the above table, all variables are statistically significant at the 1% significance level since
their P-Values are well below 0.01.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 11

 STATGRAPHICS – Rev. 9/16/2013
Observed versus Predicted
The Observed versus Predicted plot shows the observed values of Y on the vertical axis and the
predicted values Yˆ on the horizontal axis.

Plot of MPG Highway

50
45
40
observed

35
30
25
20
20 25 30 35 40 45 50
predicted

If the model fits well, the points should be randomly scattered around the diagonal line. Any
change in variability from low values of Y to high values of Y might indicate the need to
transform the dependent variable before fitting a model to the data. In the above plot, the
variability appears to increase somewhat as the predicted values get large.

Residual Plots
As with all statistical models, it is good practice to examine the residuals. In a regression, the
residuals are defined by

ei  yi  yˆ i (7)

i.e., the residuals are the differences between the observed data values and the fitted model.

The Multiple Regression procedure creates 3 residual plots:

1. versus X.
2. versus predicted value Yˆ .
3. versus row number.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 12

 STATGRAPHICS – Rev. 9/16/2013
Residuals versus X
This plot is helpful in visualizing any missed curvature with respect to a selected variable.

Residual Plot
4.3
Studentized residual

2.3

0.3

-1.7

-3.7
1600 2100 2600 3100 3600 4100 4600
Weight

No obvious curvature is detectable.

Residuals versus Predicted

This plot is helpful in detecting any heteroscedasticity in the data.

Residual Plot
4.3
Studentized residual

2.3

0.3

-1.7

-3.7
20 25 30 35 40 45 50
predicted MPG Highway

Heteroscedasticity occurs when the variability of the data changes as the mean changes, and
might necessitate transforming the data before fitting the regression model. It is usually
evidenced by a funnel-shaped pattern in the residual plot. In the plot above, some increased
variability in miles per gallon can be seen at high predicted values, which corresponds to the
smaller cars. For the smaller cars, the miles per gallon appears to vary more than for the larger
cars.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 13

 STATGRAPHICS – Rev. 9/16/2013
Residuals versus Observation
This plot shows the residuals versus row number in the datasheet:

Residual Plot
4.3
Studentized residual

2.3

0.3

-1.7

-3.7
0 20 40 60 80 100
row number

If the data are arranged in chronological order, any pattern in the data might indicate an outside
influence. In the above plot, no obvious trend is present, although there is a standardized residual
in excess of 3.5, indicating that it is more than 3.5 standard deviations from the fitted curve.

Pane Options

 Plot: The following residuals may be plotted on each residual plot:

1. Residuals – the residuals from the least squares fit.

2. Studentized residuals – the difference between the observed values yi and the predicted
values ŷ i when the model is fit using all observations except the i-th, divided by the
estimated standard error. These residuals are sometimes called externally deleted
residuals, since they measure how far each value is from the fitted model when that
model is fit using all of the data except the point being considered. This is important,
since a large outlier might otherwise affect the model so much that it would not appear to
be unusually far away from the line.

 Plot versus: the independent variable to plot on the horizontal axis, if relevant.
 2013 by StatPoint Technologies, Inc. Multiple Regression - 14
 STATGRAPHICS – Rev. 9/16/2013

Unusual Residuals
Once the model has been fit, it is useful to study the residuals to determine whether any outliers
exist that should be removed from the data. The Unusual Residuals pane lists all observations
that have Studentized residuals of 2.0 or greater in absolute value.

Unusual Residuals
Predicted Studentized
Row Y Y Residual Residual
31 33.0 40.1526 -7.15265 -2.81
36 20.0 26.9631 -6.96309 -2.62
39 50.0 43.4269 6.5731 2.72
42 46.0 36.4604 9.53958 3.66
60 26.0 32.8753 -6.8753 -2.50
73 41.0 35.3266 5.67338 2.04

Studentized residuals greater than 3 in absolute value correspond to points more than 3 standard
deviations from the fitted model, which is a very rare event for a normal distribution. In the
sample data, row #42 is more 3.5 standard deviations out. Row #42 is a Honda Civic, which was
listed in the dataset as achieving 46 miles per gallon, while the model predicts less than 37.

Points can be removed from the fit while examining any of the residual plots by clicking on a
point and then pressing the Exclude/Include button on the analysis toolbar.

Influential Points
In fitting a regression model, all observations do not have an equal influence on the parameter
estimates in the fitted model. Those with unusual values of the independent variables tend to
have more influence than the others. The Influential Points pane displays any observations that
have high influence on the fitted model:

Influential Points
Mahalanobis
Row Leverage Distance DFITS
19 0.139122 13.5555 0.18502
28 0.246158 28.3994 0.685044
31 0.156066 15.6544 -1.225
36 0.0961585 8.58597 -0.849931
39 0.250016 29.0136 1.89821
60 0.0298891 1.78389 -0.463748
73 0.0352144 2.29596 0.451735
83 0.102406 9.27903 0.573505
Average leverage of single data point = 0.0434783

Points are placed on this list for one of the following reasons:

 Leverage – measures how distant an observation is from the mean of all n observations in
the space of the independent variables. The higher the leverage, the greater the impact of the
point on the fitted values ŷ. Points are placed on the list if their leverage is more than 3 times
that of an average data point.

 Mahalanobis Distance – measures the distance of a point from the center of the collection of
points in the multivariate space of the independent variables. Since this distance is related to
leverage, it is not used to select points for the table.
 2013 by StatPoint Technologies, Inc. Multiple Regression - 15
 STATGRAPHICS – Rev. 9/16/2013

 DFITS – measures the difference between the predicted values ŷ i when the model is fit with
and without the i-th data point. Points are placed on the list if the absolute value of DFITS
exceeds 2 p / n , where p is the number of coefficients in the fitted model.

In the sample data, rows #28 and #39 show a leverage value of nearly 6 times that of an average
data point. Rows #31 and #39 have the largest values of DFITS. Removing high influence points
is not recommended on a routine basis. However, it is important to be aware of their impact on
the estimated model.

Confidence Intervals
The Confidence Intervals pane shows the potential estimation error associated with each
coefficient in the model.

95.0% confidence intervals for coefficient estimates

Standard
Parameter Estimate Error Lower Limit Upper Limit
CONSTANT 55.0336 9.61121 35.9333 74.1339
Weight -0.023276 0.00475462 -0.0327248 -0.0138271
Weight^2 0.00000230174 7.73643E-7 7.64284E-7 0.0000038392
Wheelbase 0.220693 0.0860352 0.0497156 0.39167

Pane Options

 Confidence Level: percentage level for the confidence intervals.

Correlation Matrix
The Correlation Matrix displays estimates of the correlation between the estimated coefficients.

Correlation matrix for coefficient estimates

CONSTANT Weight Weight^2 Wheelbase
CONSTANT 1.0000 -0.6247 0.7456 -0.6847
Weight -0.6247 1.0000 -0.9776 -0.1349
Weight^2 0.7456 -0.9776 1.0000 -0.0508
Wheelbase -0.6847 -0.1349 -0.0508 1.0000

This table can be helpful in determining how well the effects of different independent variables
have been separated from each other. Note the high correlation between the coefficients of
Weight and Weight2. This is normal whenever fitting a non-centered polynomial and simply
means that the coefficients could change dramatically if a different order polynomial was

 2013 by StatPoint Technologies, Inc. Multiple Regression - 16

 STATGRAPHICS – Rev. 9/16/2013
selected. The fact that the correlation between the coefficients of Weight and Wheelbase is small
is more interesting, since it implies that there is little confounding between the estimated effects
of those variables. Confounding or intermixing of the effects of two variables is a common
problem when attempting to interpret models estimated from data that was not collected from a
designed experiment.

Reports
The Reports pane creates predictions using the fitted least squares model. By default, the table
includes a line for each row in the datasheet that has complete information on the X variables
and a missing value for the Y variable. This allows you to add rows to the bottom of the
datasheet corresponding to levels at which you want predictions without affecting the fitted
model.

For example, suppose a prediction is desired for a car with a Weight of 3500 and a Wheelbase of
105. In row #94 of the datasheet, these values would be added but the MPG Highway column
would be left blank. The resulting table is shown below:

Regression Results for MPG Highway

Fitted Stnd. Error Lower 95.0% Upper 95.0% Lower 95.0% Upper 95.0%
Row Value CL for Forecast CL for Forecast CL for Forecast CL for Mean CL for Mean
94 24.7357 2.91778 18.9381 30.5333 23.7842 25.6872

Included in the table are:

 Row - the row number in the datasheet.

 Fitted Value - the predicted value of the dependent variable using the fitted model.

 Standard Error for Forecast - the estimated standard error for predicting a single new
observation.

 Confidence Limits for Forecast - prediction limits for new observations at the selected
level of confidence.

 Confidence Limits for Mean - confidence limits for the mean value of Y at the selected
level of confidence.

For row #94, the predicted miles per gallon is 24.7. Models with those features can be expected
to achieve between 18.9 and 30.5 miles per gallon in highway driving.

Using Pane Options, additional information about the predicted values and residuals for the data
used to fit the model can also be included in the table.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 17

 STATGRAPHICS – Rev. 9/16/2013
Pane Options

You may include:

 Observed Y – the observed values of the dependent variable.

 Fitted Y – the predicted values from the fitted model.
 Residuals – the ordinary residuals (observed minus predicted).
 Studentized Residuals – the Studentized deleted residuals as described earlier.
 Standard Errors for Forecasts – the standard errors for new observations at values of the
independent variables corresponding to each row of the datasheet.
 Confidence Limits for Individual Forecasts – confidence intervals for new observations.
 Confidence Limits for Forecast Means – confidence intervals for the mean value of Y at
values of the independent variables corresponding to each row of the datasheet.

Interval Plots
The Intervals Plots pane can create a number of interesting types of plots. The plot below shows
how precisely the miles per gallon of an automobile can be predicted.

Plot of MPG Highway with Predicted Values

50
45
MPG Highway

40
35
30
25

20
15
20 25 30 35 40 45 50
predicted value

 2013 by StatPoint Technologies, Inc. Multiple Regression - 18

 STATGRAPHICS – Rev. 9/16/2013
An interval is drawn on the plot for each observation in the dataset, showing the 95% prediction
limits for a new observation at the corresponding predicted value.

Pane Options

 Plot Limits For: type of limits to be included. Predicted Values plots prediction limits at
settings of the independent variables corresponding to each of the n observations used to fit
the model. Means plots confidence limits for the mean value of Y corresponding to each of
the n observations. Forecasts plots prediction limits for rows of the datasheet that have
missing values for Y. Forecast Means plots confidence limits for the mean value of Y
corresponding to each row in the datasheet with a missing value for Y.

 Plot Versus: the value to plot on the horizontal axis.

 Confidence Level: the confidence percentage used for the intervals.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 19

 STATGRAPHICS – Rev. 9/16/2013

Autocorrelated Data

When regression models are used to fit data that is recorded over time, the deviations from the
fitted model are often not independent. This can lead to inefficient estimates of the underlying
regression model coefficients and P-values that overstate the statistical significance of the fitted
model.

As an illustration, consider the following data from Neter et al. (1996), contained in the file
company.sgd:

Year and quarter Company Sales Industry Sales

($ millions) ($ millions)
1983: Q1 20.96 127.3
1983: Q2 21.40 130.0
1983: Q3 21.96 132.7
1983: Q4 21.52 129.4
1984: Q1 22.39 135.0
1984: Q2 22.76 137.1
1984: Q3 23.48 141.2
1984: Q4 23.66 142.8
1985: Q1 24.10 145.5
1985: Q2 24.01 145.3
1985: Q3 24.54 148.3
1985: Q4 24.30 146.4
1986: Q1 25.00 150.2
1986: Q2 25.64 153.1
1986: Q3 26.36 157.3
1986: Q4 26.98 160.7
1987: Q1 27.52 164.2
1987: Q2 27.78 165.6
1987: Q3 28.24 168.7
1987: Q4 28.78 171.7

Regressing company sales against industry sales resulting in a very good linear fit, with a very
high R-squared:

 2013 by StatPoint Technologies, Inc. Multiple Regression - 20

 STATGRAPHICS – Rev. 9/16/2013

Multiple Regression - company sales

Dependent variable: company sales
Independent variables:
industry sales
Number of observations: 20

Standard T
Parameter Estimate Error Statistic P-Value
CONSTANT -1.45475 0.214146 -6.79326 0.0000
industry sales 0.176283 0.00144474 122.017 0.0000

Analysis of Variance
Source Sum of Squares Df Mean Square F-Ratio P-Value
Model 110.257 1 110.257 14888.14 0.0000
Residual 0.133302 18 0.00740568
Total (Corr.) 110.39 19

R-squared = 99.8792 percent

R-squared (adjusted for d.f.) = 99.8725 percent
Standard Error of Est. = 0.0860563
Mean absolute error = 0.0691186
Durbin-Watson statistic = 0.734726 (P=0.0002)
Lag 1 residual autocorrelation = 0.626005

However, the Durbin-Watson statistic is very significant, and the estimated lag 1 residual
autocorrelation equals 0.626. A plot of the residuals versus row number shows marked swings
around zero:

Residual Plot
2.7
Studentized residual

1.7

0.7

-0.3

-1.3

-2.3
0 4 8 12 16 20
row number

Clearly, the residuals are not randomly distributed around the regression line.

To account for the autocorrelations of the deviations from the regression line, a more
complicated error structure can be assumed. A logical extension of the random error model is to
let the errors have a first-order autoregressive structure, in which the deviation at time t is
dependent upon the deviation at time t-1 in the following manner:

y t   0   1 xt   t (8)

 t   t 1  ut (9)

 2013 by StatPoint Technologies, Inc. Multiple Regression - 21

 STATGRAPHICS – Rev. 9/16/2013
where || < 1 and ut are independent samples from a normal distribution with mean 0 and
standard deviation . In such a case, transforming both the dependent variable and independent
variable according to

yt  yt  yt 1 (10)

xt  xt  xt 1 (11)

leads to the model

yt   0 1     1 xt  ut (12)

which is a linear regression with random error terms.

The Analysis Options dialog box allows you to fit a model of the above form using the
Cochrane-Orcutt procedure:

You may either specify the value of  in the Autocorrelation field and select Ordinary Least
Squares, or select Cochrane-Orcutt Optimization and let the value of  will be determined
iteratively using the specified value as a starting point. In the latter case, the following procedure
is used:

Step 1: The model is fit using transformed values of the variables based on the initial
value of .

Step 2: The value of  is re-estimated using the values of t obtained from the fit in Step
1.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 22

 STATGRAPHICS – Rev. 9/16/2013
Step 3: Steps 1 and 2 are repeated between 4 and 25 times until the change in the
derived value of  compared to the previous step is less than 0.01.

The results are summarized below using the sample data:

Multiple Regression - company sales

Dependent variable: company sales
Independent variables:
industry sales
Number of observations: 19

Cochrane-Orcutt transformation applied: autocorrelation = 0.765941

Standard T
Parameter Estimate Error Statistic P-Value
CONSTANT -0.64188 0.642688 -0.998744 0.3319
industry sales 0.171085 0.00408958 41.8343 0.0000

Analysis of Variance
Source Sum of Squares Df Mean Square F-Ratio P-Value
Model 7.7425 1 7.7425 1750.11 0.0000
Residual 0.0752083 17 0.00442402
Total (Corr.) 7.81771 18

R-squared = 99.038 percent

R-squared (adjusted for d.f.) = 98.9814 percent
Standard Error of Est. = 0.0665133
Mean absolute error = 0.0531731
Durbin-Watson statistic = 1.7354
Lag 1 residual autocorrelation = 0.0990169

The above output shows that, at the final value of  = 0.766, the Durbin-Watson statistic and the
lag 1 residual autocorrelation, computed using the residuals from the regression involving the
transformed variables, are much more in line with that expected if the errors are random. The
model also changed somewhat.

Save Results
The following results may be saved to the datasheet:

1. Predicted Values – the predicted value of Y corresponding to each of the n observations.

2. Standard Errors of Predictions – the standard errors for the n predicted values.
3. Lower Limits for Predictions – the lower prediction limits for each predicted value.
4. Upper Limits for Predictions – the upper prediction limits for each predicted value.
5. Standard Errors of Means – the standard errors for the mean value of Y at each of the n
values of X.
6. Lower Limits for Forecast Means – the lower confidence limits for the mean value of Y
at each of the n values of X.
7. Upper Limits for Forecast Means– the upper confidence limits for the mean value of Y at
each of the n values of X.
8. Residuals – the n residuals.
9. Studentized Residuals – the n Studentized residuals.
10. Leverages – the leverage values corresponding to the n values of X.
11. DFITS Statistics – the value of the DFITS statistic corresponding to the n values of X.
12. Mahalanobis Distances – the Mahalanobis distance corresponding to the n values of X.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 23

 STATGRAPHICS – Rev. 9/16/2013

Calculations

Regression Model

Y   0  1 X 1   2 X 2  ...   k X k (13)

Error Sum of Squares

 
n 2

Unweighted: SSE   yi  ˆ 0  ˆ1 xi  ˆ 2 x2  ...  ˆ k xk (14)

i 1

 
n 2

Weighted: SSE   wi yi  ˆ 0  ˆ1 xi  ˆ 2 x2  ...  ˆ k xk (15)

i 1

Coefficient Estimates

̂   X WX 1  X WY  (16)


s 2 ˆ  MSE X WX 
1
(17)

SSE
MSE  (18)
n p

where ˆ is a column vector containing the estimated regression coefficients, X is an (n, p)

matrix containing a 1 in the first column (if the model contains a constant term) and the settings
of the k predictor variables in the other columns, Y is a column vector with the values of the
dependent variable, and W is an (n, n) diagonal matrix containing the weights wi on the diagonal
for a weighted regression or 1’s on the diagonal if weights are not specified. A modified sweep
algorithm is used to solve the equations after centering and rescaling of the independent
variables.

 2013 by StatPoint Technologies, Inc. Multiple Regression - 24

 STATGRAPHICS – Rev. 9/16/2013
Analysis of Variance

With constant term:

Source Sum of Squares Df Mean F-Ratio

Square
 n 
2 SSR MSR
MSR  F
Model   wi yi  k k MSE
SSR  bX WY   i 1 n 

 wi
i 1

SSE
Residual SSE  Y WY  bX WY n-k-1 MSE 
n  k 1

n 2

Total SSTO   wi  yi  y  n-1

i 1
(corr.)

Without constant term:

Source Sum of Squares Df Mean F-Ratio

Square
SSR MSR
MSR  F
Model SSR  bX WY k k MSE
SSE
Residual SSE  Y WY  bX WY n-k MSE 
nk

Total SSTO  Y WY n

R-Squared

 SSR 
R 2  100 % (19)
 SSR  SSE 

Adjusted R-Squared

  n  1  SSE 
2
Radj  1001    % (20)
  n  p  SSR  SSE 

 2013 by StatPoint Technologies, Inc. Multiple Regression - 25

 STATGRAPHICS – Rev. 9/16/2013
Stnd.Error of Est.

̂  MSE (21)

Residuals

ei  yi  ˆ o  ˆ1 x1  ...  ˆ k xk (22)

Mean Absolute Error

w e i i
MAE  i 1
n
(23)
w
i 1
i

Durbin-Watson Statistic
n

 e i  ei 1 
2

D i 2
n
(24)
e
i 1
2
i

If n > 500, then

D2
D*  (25)
4/ n

is compared to a standard normal distribution. For 100 < n ≤ 500, D/4 is compared to a beta
distribution with parameters

n 1
  (26)
2

For smaller sample sizes, D/4 is compared to a beta distribution with parameters which are based
on the trace of certain matrices related to the X matrix, as described by Durbin and Watson
(1951) in section 4 of their classic paper.

Lag 1 Residual Autocorrelation

n

e e i i 1
r1  i 2
n
(27)
e i 1
2
i

 2013 by StatPoint Technologies, Inc. Multiple Regression - 26

 STATGRAPHICS – Rev. 9/16/2013
Leverage


hi  diag X i X WX  X i wi
1
 (28)

p
h (29)
n

Studentized Residuals

ei wi
di  (30)
MSEi 1  hi 

Mahalanobis Distance

 n

 hi  wi /  wi 
MDi   i 1  n(n  2) (31)
 1  hi  n 1
 
 

DFITS

di  hi 
DFITS i    (32)
wi  1  hi 

Standard Error for Forecast

sYh( new)   MSE1  X h  X WX  X h 

 1
(33)
 

Confidence Limit for Forecast

Yˆh  t / 2,n  p sYh ( new )  (34)

Confidence Limit for Mean


Yˆh  t / 2,n p MSE X h  X WX  X h
1
 (35)

 2013 by StatPoint Technologies, Inc. Multiple Regression - 27