AP STATS CRAM SHEET

If you want to use this during the AP corresponding to categories ▪ Symmetric -- When the left half is
Exam, either print it out or save your approx a mirror image of the right
own copy. Don’t access the google doc
during the exam! Even if you don’t
have edit access, CB doesn’t want
people on shared documents.

To make a copy: Go to File→ Make a copy

To download: Go to File-->Download (pdf ▪ Skewed -- When extreme values in
recommended) o Dotplots -- Display where dots, one direction cause the graph to
representing observations, are placed on appear to have a “tail”
I’m removing access 15 minutes before a number line
the exam starts o Stemplots -- Display that shows how
individual values are distributed in data,
Exam Date: 1pm Central Time based on a “stem”. The stem is usually 1
Friday May 22nd digit place higher than the “leaf”
1|123
Exploring Data 2|456
▪ Skew direction is based on the
(20-30%) 3|789
Tens | Ones
●
direction of the tail.
SUMMARIZING DISTRIBUTIONS
● TYPES OF VARIABLES This set of data would be interpreted as
o Population -- Entire group of
o Categorical Variable – Variable that [11,12, 13, 24, 25, 26, 37, 38, 39]
individuals the study is examining
places the individual studies into one of o Histogram -- Display that acts like a
o Sample -- The part of the population
several groups (sex/eye color/favorite stemplot on its side. Bars represent
that is actually studied.
team/etc) frequencies over a certain stretch of data.
o W’s -- Who, what, where, when, why,
o Quantitative Variable -- Variable that is The frequency (on the y) can be relative
how, etc. Used for describing a
measured with numbers (with numbers) or percentage based.
distribution
(age/height/weight/IQ/etc)
o Unusual -- Anything that is not
● TABULAR METHODS (TABLES)
normal in a graph (outliers, gaps, large
o Frequency Distribution Table -- Table
fluctuations, etc)
giving all possible outcomes and their
o When describing a distribution, use
frequencies
“WCUSS”, meaning describe with
o n -- number of observations in a data set
W’s, Center, Unusual Aspects, Shape,
o Frequency (f) -- number of times that an
Spread
observation occurs
● MEASURES OF CENTRAL TENDENCY
o Relative Frequency (rf) -- ratio of
o Mean (Average) -- Sum of all values
frequency to total # of observations
o Cumulative Frequency Charts -- Graph divided by the number of values
o Cumulative Frequency – Number of
that shows frequency of one group plus o Median -- The point where 50% of
observations less than or equal to a
the frequencies of all groups of smaller the rest of the data is below said
specified value
observations point, and 50% is above
● GRAPHICAL METHODS
o Ogive -- Cumulative Frequency Graph o Mean is affected by outliers, while
o Bar Charts -- Graph with bars
the median is resistant to outliers
representing the frequency of a category
o Mean is pulled in the direction of
the skew
● MEASURES OF VARIATION
o Range -- Difference between largest
and smallest measurement
o IQR -- Interquartile Range, measure
of the middle 50% of the data.
● EXAMINING GRAPHS Calculated by Q3-Q1
o Pie Chart -- Display of frequency o Center -- The point around which data o SD -- Standard Deviation, measure of
distributions with “pie slices” points are spread (mean, median) variance which is affected by
o Spread -- Measure of how far data is outliers. See formula chart for
away from the center (SD, IQR, Range) equation.
o Shape -- Shows where most of the data ● Measures of Position
is o Percentiles -- Places where x percent
of data is below said value. For
example, the value at the 95th
 AP STATS CRAM SHEET
percentile means 95% of values fall ▪ Compare in context variation in y-values explained by the
below said value, and 5% are above o Make sure that you are comparing, IE LSRL. It is equal to the corr coeff
o Quartiles -- Division of data into 4 using comparison words such as squared.
equal parts using the 25th (Q1), 50th greater/more/less/lower/etc o Independent variable -- AKA
(Median), and 75th (Q3) percentiles ● BIVARIATE DATA (SCATTERPLOTS) explanatory variable, the variable that
o Standardized Scores (z-scores) -- o Used to compare relationships when changed, causes a change in the
Measure of the number of standard between two quantitative variables dependent.
deviations a value is away from the o When describing, use “FUDS*”, o Dependent variable -- AKA response
mean. meaning form, unusual, direction, variable, the variable that changes in
strength, and * response to a change in the
Negative z score = measurement lower o Form -- Is the scatterplot linear or independent
than mean. Positive z score = curved (nonlinear)? ● OUTLIERS
measurement higher than mean o Unusual -- Are there any unusual O Defined as something that is

● BOXPLOTS features (outliers)? surprisingly different from the rest of

O Graphical display based on measures
o Direction -- Is the relationship the data
of position positive (x ↑ y ↑) or negative (x ↑ y o Influential Observation -- An
O Also called a box and whisker, for
↓) observation that strongly affects a
there is a box in the middle and o Strength -- Strong relationships statistic. Some outliers will cause the
“whiskers” to the side follow a clear pattern, weak LSRL to change dramatically, making
relationships don’t follow as much of them influential points, while others
a pattern, and no relationship=no may not have any effect, making them
pattern. not influential.
o * -- As x increases, y seems to ● RESIDUAL PLOTS
generally increase/decrease as well O Residual -- The distance away an
OOutliers are marked with a dot or a x ● CORRELATION COEFFICIENTS actual value is away from a predicted
(see above). O The correlation coefficient (denoted value. Eqn:
O The end of the left whisker is the by r)is a measure of how linear a o Residual Plot -- A scatterplot of the
lowest NON-OUTLIER point. relationship is residuals vs the predicted values of Y.
o Left end of box=Q1 o r is always between (or equal to ) -1 o Residual Plots should look random
o Middle Line=Median and 1 and messy. If there are trends, then
o Right end of box=Q3 o Positive r = Positive direction. that indicates that a linear model may
o End of right whisker is the highest Negative r = Negative direction not be appropriate.
NON-OUTLIER point. o No correlation = ● ACHIEVING LINEARITY
o The ends of the whiskers are NOT Weak correlation = O If a plot is nonlinear, transformations
FENCES Strong correlation = can be used to make said plot linear.
O Fences -- Points that determine when Very Strong correlation = For example, you can graph ln (Y)
a value becomes an outlier. against X if it makes the plot linear.
▪ Lower fence -- Q1-1.5 IQR, Perfect correlation o Options
Anything lower is an outlier Note that this also works for negative ▪ Take Logarithms (New Y=ln (Y))
▪ Upper fence -- Q3+1.5 IQR correlation. ▪ Take Square roots (New Y=sqrtY)
Anything higher is an outlier ● LSRL (L EAST SQUARES REGRESSION ▪ Take Reciprocal (New Y=1/Y)
● CHANGING ALL VALUES OF A DATA SET LINE) ▪ Square (New Y=Y^2)
o Addition -- Adding/subtracting a O Defined as the line that gives a straight ● TWO WAY TABLES
constant to all data values will line relationship between 2 variables. O Contingency Table -- A table that
increase/decrease all measures of Also known as the line of best fit. plots categories of two different
center and position by said constant. o classifications against each other
o Multiplication -- Multiplying a Y is the response (dependent) variable See here for example
constant to all data values will X is the explanatory (independent) o Marginal Frequency -- Frequency
multiply all values (center, position, variable with which each category occurs. It's
and spread) by said constant. is the y-int the sum of all the frequencies in a
● COMPARING DISTRIBUTIONS is the slope row/column.
o Laundry List of Comparisons is the random error. You probably o Conditional Relative Frequency --
▪ Compare centers won’t see this much The relative frequency of one category
▪ Compare spreads o The predicted value of Y for a given given that another category has
▪ Compare clusters/gaps in value X using the LSRL is denoted by occurred.
measurements Ex. Given that student=female, what is
▪ Compare unusual features/outliers o Coefficient of Determination (R^2) -- prob of choosing engineering
▪ Compare shapes A measurement of the percent of
 AP STATS CRAM SHEET
o Association -- When one category o Response Bias -- When a question or the Ex. Sorting by age in a study on heart
influences the relative frequency of behavior of the interviewer changes a disease
another. person’s response Blocking may be used (not always
Sampling and o Nonresponse Bias -- When a person required) to similarly reduce variability

Experimentation selected to respond refuses to answer

o Undercoverage Bias -- When part of the
o Replication -- repeating the process of
giving treatment to reduce variation.
(10-15%) population is left out of the sampling
frame
Should be used.
o Diagram*
● DATA COLLECTION
o Population -- The entire group of o Wording Bias -- When the question o Explaining your experiment*
individuals that we are interested in asked may lead to an inaccurate response o *For the experiment explanation, scroll
o Frame (Sampling Frame) -- A list of all ● PLANNING AND CONDUCTING to the bottom since it’ll screw up the
the members from which the sample is EXPERIMENTS-TERMINOLOGY formatting if I write it here. For the
to be taken o Experimental Unit -- The smallest unit diagram, I’m not really sure how to put
o Sample -- The part of the population of the population to which treatment is it on here. Search it up or look at your
actually being examined applied (usually an individual person or notes I guess.
o Sample Survey -- Process of collecting item) ● MATCHED PAIR DESIGN
information from a sample o Factor -- A variable whose effect on the O Experiment where similar subjects are

o Census -- Collecting info from the entire response is the main aspect of the paired up so that they match as evenly
population, usually impractical experiment. as possible. Each subject is then
o Experiment -- A planned activity where If a doctor is trying to find a more randomly assigned to one of two
a variable is manipulated so that effective medication, then the treatments.
differences can be created and observed medication is the factor of interest. o Example: Design an experiment where
o Observational Study -- Where the ▪ Qualitative Factor -- Data in people drink coke and pepsi test test,
relationship between variables is non-numerical groups and figure out which one they prefer.
observed over time. No variables are ▪ Quantitative Factor -- Data can be Here what you would do is matched
manipulated. measured numerically pair where T1=coke then pepsi and
o Confounding Factors -- A external o Levels -- The values of a factor used in T2=pepsi then coke. Compare response
factor that may influence the results of a an experiment of preference.
study/experiment If a doctor is trying to find a more o For more examples of matched pair, see
● SAMPLING effective medication, and two companies here
o Simple Random Sample (SRS) -- offer new alternatives, then the 3 levels
Getting a sample from a population are current medication, medication from
Anticipating Patterns
where each person has the same chance
of being selected.
company 1, and medication form
company 2. (20-30%)
*SRS is either from a small population o Treatment -- The factor-level ● PROBABILITY
with replacement or from a large combinations used in the experiment. O Defined as the chance of something

population with/without replacement. In the doctor example, the 3 levels happening

o Systematic Sample -- The first item is become the three treatments, as there o Probability of an event A occurring is
selected using SRS and then every xth was only 1 factor of interest. represented as P(A)
person afterwards will be used in the o Control Group -- A group of o Observed Probability -- The
sample experimental units that is not given any probability that was observed after
o Stratified Random Sample -- Population treatment, used to establish a baseline of performing an experiment.
is divided into groups (strata) and SRS is comparison. o Expected Probability -- The probability
used on each group o Placebo Group -- A control group that calculated using math
o Proportional Sampling -- Stratified receives a placebo, or a treatment that o Sample Space -- A set of all possible
sampling but the number chosen is has no effect. outcomes of an experiment
proportional to the size of each strata o Single Blind -- Either the patient or the o Tree Diagram -- A diagram using
o Cluster Sampling -- An item is chosen researcher doesn’t know which treatment “branches” to represent probabilities of
via SRS, and all items around chosen was given. Reduces bias events occuring
item will be used in the sample. o Double Blind -- Both the patient and the ● PROBABILITY RULES
● BIAS AND INCORRECT SAMPLING researcher don’t know which treatment O P(A) must be between 0 and 1
O Judgmental Sampling -- When sample
was given. Reduces bias O Total sum of probabilities for all

is selected nonrandomly ● CONDUCTING EXPERIMENTS possible outcomes is always 1

o Convenience Sampling -- Easy to obtain o Randomization must be used to try to o Impossible event -- If an event can
samples that may not representative minimize variability never occur, P(A)=0
o Voluntary Response (Bias) -- When o Blocking -- sorting experimental units o Sure event -- If an event must occur
subjects choose to take part in a sample into “blocks”, where all units in a block every time
share similar characteristics ● PROBABILITY TERMINOLOGY
 AP STATS CRAM SHEET
O Complement -- The set of all outcomes OBinomial Distribution -- A distribution bounds, any number greater than 5 SDs
that do NOT lead to the event. of discrete random variables that has works as the boundary.
Complement=1-P(A) certain properties o To calculate the z score from a
o Disjoint events -- Events which have no The binomial variable X represents the # percentile, use inverse norm (invNorm)
outcome in common (can’t occur of successes in n trials on the percentile you want, specifically
together) ▪ There are n repeated trials of a invNorm(0.95, 0,1)
o Union -- The set of all possible number (fixed in advance) o Characteristics of a Normal Model
outcomes that lead to at least one of ▪ Each trial has two possible ▪ Continuous
two events. P(A or B) outcomes-success or fail ▪ Symmetric around mean
P(A or B)=P(A)+P(B)-P(A and B) ▪ All trials are identical and ▪ Bell shaped
o Intersection -- P(A and B) independent ▪ Mean=median=mode
P(A and B)=P(A)*P(B|A) o binompdf (n, p, x) -- The probability of ▪ Curve approaches x axis on both
o Conditional Event -- The probability of getting x units from n units given that sides
one event occurring given another has the probability of getting an x unit is p. ▪ Follows a 68-95-99.7 rule, meaning
occurred. P(A|B), reads “A given B” o binomcdf (n,p,x) -- The prob of getting that 68% of data is within 1 SD, 95%
P(A|B)=P(A or B)/P(B) x units or less from n units given that the is within 2 SDs, and 99.7% is within
o Independence -- When the occurrence prob of getting a x unit is p 3 SDs.
of one event doesn’t depend on another o Geometric Distribution -- A distribution ● SAMPLING DISTRIBUTIONS
Independence is proven if of discrete random variables with certain O Parameter -- Numerical measure of a
P(A or B)=P(A)*P(B) properties. population
● RANDOM VARIABLES The geometric rand variable X Ex. Your GPA, as GPA is computed
O Variable -- A quantity whose value represents the number of trials required using grades from all your courses
varies from subject to subject (ex. to obtain a first success Ex. The mean household income for
height, # of texts, hair color, etc) ▪ There are n repeated trials ALL households in CA
o Probability Experiment -- An ▪ 2 possible outcomes (suc/fail) o Statistic -- Numerical measure of a
experiment whose possible outcomes ▪ Trials are repeated until a set sample
may be known but the exact outcome is number of successes is reached Ex. The % of votes a candidate receives,
a random event ▪ All trials are identical and as not all eligible voters vote
o Random Variable -- A symbol that independent. Ex. The mean household income for 500
represents the (potential) outcome of an o geometpdf (p, x) -- The prob that the random households in CA
experiment first success is on x given that the o Sample Distribution -- The probability
▪ Discrete Random Variable -- A probability of success is p. distribution of all possible values of a
quantitative variable that takes a o geometcdf (p, x) -- The prob that the statistic
countable number of values (# of first success is on x or lower than x o Sampling Distribution -- The
texts per day, # of kids per class) given that the prob of success is p. distribution of sample statistics which
▪ Continuous Random Variable -- o Both binom and geom can be simulated form the theoretical distribution.
A quantitative variable that can with random digit tables. If you are Basically it describes how something
take all possible values within a asked to simulate, make sure to describe behaves from one sample distribution to
given range (ex. height, amount of every step and aspect in detail to get the another in a population.
rainfall, etc) most points. o Standard Error -- The standard
● PROBABILITY DISTRIBUTIONS ● NORMAL MODEL deviation of the distribution of a statistic
o Probability Distribution of a Discrete O Aka bell curve, defined as a model o Central Limit Theorem -- Regardless of
Random Variable -- A display giving all which closely approximates many the shape of the distribution, if the
possible values taken by a random different stats measurements. sample size is large (n>30), then the
variables and the probs for each value o Standard Normal -- The normal model distribution will be approx normal
o Expected value -- The mean of a with a mean of 0 and a SD of 1. Any ▪ Mean of CLT normal distribution:
discrete random variable X. See formula normal random variables can be made
chart for its equation (u=E(x)=___) into the standard normal model with z ▪ SD of CLT normal distribution
o Variance of a discrete random variable score formulas.
-- See formula chart o Z-Score -- The value given by the ▪ As n ↑, the distribution becomes
(SD^2=Summation…) equation . Indicates the # of more normal
● COMBINATIONS AND DISTRIBUTIONS SDs an object is away from the mean ▪ As n ↑, the center of the distribution
O COMBINATION -- # of ways r items can be
o The probability of a z score occurring, remains at u.
selected out of n items, given that the P(Z< or > X), can be calculated using z ▪ As n ↑, the spread of X ↓ and the
order of selection is NOT important scores. Enter in lower bound, upper distribution becomes more peaked
Denoted by , where bound, u and sigma values into the o Sampling Distribution of a Sample
O Factorial -- denoted as n!, and is normalcdf function to find the Proportion
probability. For graphs with infinite ▪
 AP STATS CRAM SHEET
● TESTING HYPOTHESES o The margin of error is the range of
▪ O Null Hypothesis -- A statement that is values above and below the sample stat
o Sampling Distribution of a Sample Mean assumed to be true unless proven in a confidence interval. Represented by
▪ otherwise. Denoted H0. ME
▪ o Alternate Hypothesis -- A statement o , where z* is the critical
o Sampling Distribution of a Difference about the parameter that must be true if value and SE is standard error. The
Between Two Independent Sample the null is false, and is usually a critical value is calculated using
Proportions proposition. Denoted HA. invNorm.
o Test Statistic -- A statistic computed o Population Proportion
▪
form the sampled data and used in
▪
testing a hypothesis. Varies between
o Sampling Distribution of a Difference o Population Mean
tests. For example, the test stat for a
Between Two Independent Sample
normal model is the z score.
Means
o P-Value -- The probability that the null ● T DISTRIBUTIONS
▪
hypothesis is true O The t distribution is different from the
▪ o Alpha Level -- The predetermined cutoff norm dist, and it is used when making
Statistical Inference point for what is considered to be
sufficient evidence against the null
estimates about population means. The
crit value becomes t* and invT is used.
(30-40%) hypothesis. For example, if the alpha
level is 0.05, then any p values lower
● CALCULATING CONFIDENCE INTERVALS
● ESTIMATION O Identify the type of confidence interval
o Estimate -- A specific guess or value than 0.05 will be enough evidence to to be used.
computed from a sample. Process of reject the null. o Check conditions, which will vary based
estimation is exactly what it sounds like ● ERRORS on interval.
O Type I Error -- Rejecting the null
▪ Point Estimation -- Giving a single o Set up the equation
point for an estimate hypothesis when it is true. Denoted by Confidence Interval for p is
▪ Interval Estimation -- Giving a Confidence Interval for u is
range of values as an estimate o Type II Error -- Failing to reject the null o State a correct conclusion in context. (I
▪ Point estimation should be either x when it is false. Denoted by am __% conf that the true population
bar or p hat. ___ is between ___ and ___)
▪ Point estimation for differences in (___% of all ___% conf intervals will
population means: contain the true population ___)
● TESTING HYPOTHESES
▪ Point estimation for differences in o State the hypotheses
population proportions: o Identify the test required and necessary
conditions
● SAMPLING DISTRIBUTIONS OF STATS o Provide the correct mechanics, done by
O Defined as the distribution of estimates copying everything from your calculator
o Power -- The prob of correctly rejecting as well as df if necessary
taken rom all possible samples of one the null when there really is a change,
size from a population. Usually o State your conclusion, and if you
Power=1-P(Type II Error) reject/fail to reject the H0 as well as why.
represented by histograms ▪ If Sample Size ↑ Then Power ↑
o Variability -- The degree of variation in ● LISTS OF TESTS
▪ If Type I Error ↑ Then Power ↑ o 1 Prop Z Test -- Used when testing if a
a statistic values ▪ If effect size (difference between
o Bias -- Indication that an estimate may proportion of one sample matches a
hypothesized and true) ↑ Then Power claimed proportion.
be wrong compared to the true answer ↑
o Inefficiency -- Indication that a guess is Ex. Bob claims that 80% of people like
o One Tailed Test-- The rejection region his products. You believe that his claim is
wrong but at irregular intervals. AKA (based on alpha lvl) is formed in 1 end
random variability inflated, and in fact less than 80% of
only people like his products. You get a sample
o Unbiasedness -- Idea that a stat is o Two Tailed Test -- The rejection region
expected to give values centered around of 200 people who bought Bob’s products
(based on alpha lvl) is formed on both to test this theory.
the unknown parameter ends
● INTERVALS o 1-Sample T Test -- Used when testing if
O Confidence Interval -- A range of two a claimed mean is true.
numbers between which we are Ex. A company claims that its mean
reasonably certain the true parameter cereal weight is 12oz. You think it's less
will fall and you take a random sample of cereal
o Confidence Level -- The percentage of boxes to test this theory.
intervals which will contain the true ● MARGINS OF ERROR o Two Prop Z Test -- Used when testing if
parameter value. there is a difference between proportions.
 AP STATS CRAM SHEET
Ex. A source claims that there is no evidence of an association between show the numbers 1-(# of subjects) to each
difference in the proportions of men and and gender? (unit in block A), write these numbers
women to vote right. You think that this is o Linreg t-test -- Used for testing if there is on identical sheets of paper, and place
wrong, and you find a randomly selected a significant relationship between x and y
them in my magic hat. Shaking
survey that gives values which state that variables. Does this by seeing if the
273/477 men voted republican and regression slope is different from 0.
between each draw, without
264/498 women voted republican, and Linreg t-tests can be determined usually replacement, the first x people will be
you use this to test your theory. when they give you explicit points or a assigned to treatment 1, with
o Two sample T Test -- Used when testing computer output. (description of treatment 1). The next
if there is a difference between means.
Ex. A source claims that all ethnic groups
Other x people will be assigned to treatment
2, with (description), the next x...etc.
are paid equally in a company. You think ● INTERPRETING COMPUTER OUTPUT
(Repeat processes for Block B, Block
that this is wrong, and you find a
C, etc) I will incorporate
randomly selected survey that gives
values which say the mean pay for whites
single/double blinding (*if necessary).
is $4/hr SD $2 and mean pay for The (experimental units) will have
minorities is $3.5/hr SD $5, and you use identical conditions (describe
this to test your theory. examples). At the end of (time
o Matched Pair T-Test -- Used for period), I will hire an expert, unbiased
calculating if there is a mean difference in (researcher type) to measure and
paired data. compare the response of (response
Ex. You want to find out if taking a
variable).
certain class will increase student’s SAT
Note that not all parts of the prompt may
scores. You compare means between
need to be used.
student scores before and after taking the
● IMPORTANT TERMS AND PROMPTS o Simulation with Random Digit Tables --
class.
O WCUSS -- W’s, Center, Unusual, Shape, Using (# of digits) numbers and random
o x^2 Goodness of Fit Test -- Used for
Spread. Used for describing distributions. digit tables, I will do a simulation to
testing if a sample matches a claim put
o FUDS*--Form, Unusual, Direction, (purpose of experiment, either to estimate
forth.
Strength, as x increases y inc/dec a mean or proportion). The numbers (a to
Ex. M&Ms claims that 50% of their
Used for describing scatterplots. b) will represent (Possible event 1), the
m&ms are red and 25% are blue, and
o r -- The correlation coefficient of ( r ) numbers (c to d) will represent (Possible
25% are green. You grab a large bag of
shows a (strength), (direction), (linearity) event 2), the numbers (d to e) will
m&ms and find that 200 are red, 97 are
relationship between ( x ) and ( y ) represent (Possible event 3)...etc. I will
blue, and 86 are green.
o r^2 -- ( R^2% ) of the variation in the y omit the numbers (y-z) (not always
o x^2 Test of homogeneity -- Used for
can be explained by a linear relationship needed). Starting from the left-most edge,
testing if two groups are the same/have
with the x taking 2 digits at a time, I will keep going
differences.
o y-int -- When ( x ) is 0 ( x units ), then the until I get (either a certain # of numbers
Ex. A TV program which took a random
predicted ( y ) is ( b0 ) ( y units ). for probability or a certain # of
sample of 100 boys and 200 girls found
o Slope -- For every 1 ( x unit ) increase in occurrences for means). This marks the
that 50 boys liked Thomas the Tank
the ( x ), there is a predicted ( b1) unit end of a trial. I will record the response
Engine, 30 liked Sesame Street, and 20
increase in the ( y ). variable of (if trail succeeded/failed for
liked Curious George. The study also
o Appropriate Linear Model -- Scatterplot probability or # of occurrences for means)
found that 100 girls liked Thomas the
is straight, Residual is messy in my table. I will do ___ trials. I will
Tank Engine, 20 liked Sesame Street, and
o Extrapolation -- When the value to be then find the (either probability of an
80 liked Curious George. Is there
solved for is beyond the boundary of event occurring or the average number of
evidence of a difference between
measurement occurrences).
preferences between boys and girls?
o Experimental Design Principles -- This simulation suggests that the
o x^2 Test of Independence -- Used for
Control, Randomize, Replicate, Block (if probability of (event) is ___.
testing of two groups are
needed) OR
independent/associated.
o Experimental Design Prompt -- I will This simulation suggests that the
Ex. Ex. A TV program which took a
design a completely randomized (occurance rate per trial) is ___.
random sample of 100 boys and 200 girls
experiment to test (explanatory var) o A note on all listed prompts - These were
found that 50 boys liked Thomas the Tank
on (response var). I have (context) prompts given to us by our teacher who
Engine, 30 liked Sesame Street, and 20
has a tendency to go overkill on them
liked Curious George. The study also given to me by (context) to act as my
(you probably don’t need to write that
found that 100 girls liked Thomas the experimental units. I will incorporate much to get credit)
Tank Engine, 20 liked Sesame Street, and blocking in this experiment based on MULTIPLE CHOICE STRATEGIES
80 liked Curious George. Is there (what is the block based on). Assign
 AP STATS CRAM SHEET
● Bubble as you go – you don’t want to run
out of time!
● Answer EVERY QUESTION – you don’t
lose points for guessing
o If you run out of time pick either B, C,
or D and bubble straight down.
● If you don’t recognize an answer choice –
it probably IS NOT THE ANSWER
● This is Stats, you can use common sense
on half of the questions
● If you get a really ridiculous answer (for
example, a z-score of 1756) you probably
did something wrong

FRQ STRATEGIES

ANSWER THE STUPID QUESTION!

● Don’t write in bullet points!
● Be VERY SPECIFIC, AS SPECIFIC
AS POSSIBLE and apply the answer to
the prompt
● Write fast, fast enough to still be legible

Created by AnonFox

Formatting and idea based on Imbesi’s

Psych Review Sheet

Notes based on Princeton, Barron’s, and

Class Notes

Category Distributions based on PrepScholar