In General: Describing/Comparing Distributions: Binomial Distribution: Geometric distribution:

B.S.: Be Specific; include C: Center-the median or mean B: Binary-either success or B: Binary-either success or
numbers where possible and U: Unusual features-gaps or outliers failure failure
refer back to the context of S: Spread- interquartile range (IQR), I: Independent trials I: Independent trials
the problem. standard deviation, or range N: Number of trials is fixed T: Trials until success
S: Shape-symmetrical, bimodal, or S: Success probability stays S: Success probability stays
Define what you’re talking skewed the same the same
about and show your work

Read the whole question Defining a Binomial Defining a Geometric

first. And ask: “what is this Distribution: X has a Distribution: X has a
question about?” binomial distribution with geometric distribution with
𝑛 = # and 𝑝 = # 𝑝=#
Check your answer when
you’re done to ensure it NSpire: Menu, 6: Stats, 5: NSpire: Menu, 6: Stats, 5:
makes sense. Distributions, A or B Distributions, H or I

Constructing a Confidence Performing a Hypothesis Test: Describing a Relationship: Describing a Scatterplot:

Interval: P: Define parameters, including STD uses correlation D: Direction- positive or
P: Define parameters, significance level coefficient, r negative
including confidence level H: Write your hypotheses S: Strength-strong, U: Unusual features-
A: State A: State assumptions/conditions moderate, or weak outliers
assumptions/conditions N: Name the test T: Trend- linear or nonlinear F: Form- linear or nonlinear
N: Name the interval T: Calculate the test statistic D: Direction-positive or S: Strength- strong,
I: Calculate the interval O: Obtain the p-value negative moderate, or weak
C: State the conclusion in M: Make a decision
context S: State the conclusion in context

1
Sentence Stems: Replace what is underlined
Interpret slope: As the x-variable increases by 1 unit, the predicted y-variable increases/decreases by slope units.
Interpret y-intercept: When the x-variable is 0 units, the predicted y-variable is y-intercept units. IF THIS NUMBER DOES NOT
MAKE SENSE (i.e. is negative), YOU MUST COMMENT ON THAT.

Describe the relationship: There is a strength, trend, direction relationship between x-variable and y-variable based on the
graph or correlation coefficient r.

Coefficient of Determination: R2% of the variability in the y-variable can be explained by a linear relationship with x-variable.
Interpret the mean: If many, many context are randomly selected, the average context will be about mean value.
Interpret standard deviation: The context typically vary from the mean value by about standard deviation value.
Interpret Confidence Interval (Conclusion): We are confidence level% confident that the true parameter in context is
between lower bound and upper bound in context.

Interpret Confidence Level: If we constructed many, many confidence intervals from random samples of size n, about
confidence level% of the intervals would capture the true population parameter in context.

Interpret the p-value: Assuming the H0 in context is true, there is a p-value% chance of getting a sample proportion/mean of
sample value or something more extreme by chance in random samples of size n.

𝒑 > 𝜶 Fail to Reject H0: We do not have convincing evidence for Ha in context.
𝒑 ≤ 𝜶 Reject H0: We do have convincing evidence for Ha in context.
Interpret the power: If the true population parameter in context is Ha, there is a power% probability of finding convincing
evidence to reject H0 in context.

2
Inference Table:
Name Assumptions Interval Test Hypothesis
1 Proportion  Random: sample must be
z randomly selected or randomly 𝑝̂ − 𝑝0
assigned 𝑝̂ (1 − 𝑝̂ ) 𝑧=
1 𝐶𝐼 = 𝑝̂ ± 𝑧 ∗√
√𝑝0 (1 − 𝑝0 ) 𝐻0 : 𝑝 =
 Independence: 𝑛 < 𝑁 𝑛 𝑛
10
 Normality: stated OR
𝑛𝑝 ≥ 10 NSpire: Menu, 6: Stats, 6: CI, 5: 1-Prop Z NSpire: Menu, 6: Stats, 7:
𝑛(1 − 𝑝) ≥ 10 Interval Stats Tests, 5: 1-Prop Z Test
2 Proportion MUST CHECK FOR BOTH
z SAMPLES 𝐶𝐼 𝑝
̂1 − 𝑝
̂2 𝐻0 : 𝑝1 = 𝑝2
 Random: sample must be 𝑧=
∗
𝑝̂(1
1 −𝑝̂)
1 𝑝
̂(1
2 −𝑝̂)
2 1 1
randomly selected or randomly = (𝑝
̂1 − 𝑝2 ±𝑧 √
̂) + √𝑝̂ (1 − 𝑝̂ )( + )
𝑛1 𝑛2 𝑛1 𝑛2 OR
assigned
1
 Independence: 𝑛 < 𝑁 NSpire: Menu, 6: Stats, 7:
10 𝐻0 : 𝑝1 − 𝑝2 = 0
 Normality: stated OR NSpire: Menu, 6: Stats, 6: CI, 6: 2-Prop Z Stats Tests, 6: 2-Prop Z Test
𝑛𝑝 ≥ 10 Interval
𝑛(1 − 𝑝) ≥ 10
1 Sample t  Random: sample must be
randomly selected or randomly 𝑠𝑥 𝑥̅ − 𝜇
assigned 𝐶𝐼 = 𝑥̅ ± 𝑡 ∗ 𝑡=
𝑆
√𝑛 𝐻0 : 𝜇 =
1
 Independence: 𝑛 < 𝑁 √𝑛
10
 Normality: stated OR
graph has no unusual features NSpire: Menu, 6: Stats, 6: CI, 2: t Interval NSpire: Menu, 6: Stats, 7:
OR 𝑛 ≥ 30 Stats Tests, 2: t Test
2 Sample t MUST CHECK FOR BOTH
SAMPLES 𝑥1 − ̅̅̅
̅̅̅ 𝑥2 𝐻0 : 𝜇1 = 𝜇2
(𝑆1 )2 (𝑆2 )2 𝑡=
 Random: sample must be
𝐶𝐼 = (𝑥 𝑥2 ± 𝑡 ∗ √
̅̅̅1 − ̅̅̅) + 2 2
randomly selected or randomly 𝑛1 𝑛2 √𝑆1 + 𝑆2 OR
assigned 𝑛1 𝑛2
1
 Independence: 𝑛 < 𝑁
10 𝐻0 : 𝜇1 − 𝜇2 = 0
 Normality: stated OR NSpire: Menu, 6: Stats, 7:
NSpire: Menu, 6: Stats, 6: CI, 4: 2 Sample Z
graph has no unusual features Stats Tests, 4: 2-Samp t Test
Interval
OR 𝑛 ≥ 30

3
 Is it a Test? /What Type of Test?

Yes/No Question:

Do that data provide…?

Is there convincing
evidence…?

Do we have evidence…?

Proportions vs. Means

Proportions: Categorical Data Means: Quantitative Data
“proportion” “mean” or “average”
Yes/No Answers Numerical Answers *Note: A question may only ask you to
Data given as fractions, Data points state the conclusion of a test. Look
out for a significance level 𝛼 for this.
includes phrase “out of”

4
Random Formulas not on the Formula Sheet:
𝐼𝑛𝑡𝑒𝑟𝑞𝑢𝑎𝑟𝑡𝑖𝑙𝑒 𝑟𝑎𝑛𝑔𝑒 (𝐼𝑄𝑅) = 𝑄3 − 𝑄1
𝐿𝑜𝑤/𝐿𝑒𝑓𝑡 𝑂𝑢𝑡𝑙𝑖𝑒𝑟𝑠 = 𝑄1 − 1.5(𝐼𝑄𝑅) 𝐻𝑖𝑔ℎ/𝑅𝑖𝑔ℎ𝑡 𝑂𝑢𝑡𝑙𝑖𝑒𝑟𝑠 = 𝑄3 + 1.5(𝐼𝑄𝑅)

𝑣𝑎𝑙𝑢𝑒 − 𝑚𝑒𝑎𝑛 𝑥−𝜇

𝑧 − 𝑠𝑐𝑜𝑟𝑒 = = To check if events are independent:
𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛 𝜎
Check the formula: 𝑃(𝐴) = 𝑃(𝐴|𝐵)

Combining Means of Random Variables: 𝜇𝑥±𝑦 = 𝜇𝑥 ± 𝜇𝑦  If they are equal, they are independent.
 If they are NOT equal, they are NOT independent.
Combining Standard Deviations of Random Variables: 𝜎𝑥±𝑦 = √𝜎𝑥 2 ± 𝜎𝑦 2

Transforming the Mean of Random Variables: 𝜇𝑎+𝑏𝑥 = 𝑎 + 𝑏𝜇𝑥

If A and B are independent events:
Transforming the Standard Deviation of Random Variables: 𝜎𝑎+𝑏𝑥 = |𝑏|𝜎
𝑃(𝐴 ∩ 𝐵) = 𝑃(𝐴) ∗ 𝑃(𝐵)
𝑅𝑒𝑠𝑖𝑑𝑢𝑎𝑙 = 𝑎𝑐𝑡𝑢𝑎𝑙 − 𝑝𝑟𝑒𝑑𝑖𝑐𝑡𝑒𝑑 = 𝑦 − 𝑦̂
𝑃(𝐴∩𝐵)
“Given”/Conditional Probability: 𝑃(𝐴|𝐵) =
𝑃(𝐵) ∪ = 𝑂𝑅

Least Squares regression line: 𝐿𝑆𝑅𝐿 = 𝑦̂ = 𝑚𝑥 + 𝑏 = 𝑎 + 𝑏𝑥 ∩ = 𝐴𝑁𝐷

𝐷𝑒𝑔𝑟𝑒𝑒𝑠 𝑜𝑓 𝑓𝑟𝑒𝑒𝑑𝑜𝑚 (𝑑. 𝑓. ) = 𝑛 − 1

General Things to Remember:

 Labeling a graph: Title, Label x-axis, Label y-axis, Scale x-axis, Scale y-axis
 PDF vs CDF
o PDF: used when looking for the probability of one specific number; uses equal signs; P(X=x)
o CDF: used when looking for the probability of a range of numbers; uses inequalities; P(X>x)

5
  Percentiles are read from left to right and represent the % of values that are less than or equal to a given value.

 Empirical Rule/68-95-99.7% Rule  Use normalcdf to find the probability/percent;

use invNorm to find the number given the
percent/probability
o Make sure to draw a curve with labels
to show work.

When a problem uses a normal distribution but wants

 Type I and Type II Errors you to find the probability given a sample size n or
says “sample proportion/mean”: it is a sampling
distribution.

Use the correct mean and standard deviation!

The Power of a Test – is the probability that the test will reject the null hypothesis when the null hypothesis is false assuming the null is true.

If you increase Type I error Type II error Power

 Increases Decreases Increases

n Same Decreases Increases

(0 – a) Same Decreases Increases

6
 Type of Pie Graph Bar Graph Segmented Bar Graph Histogram
Graph
Picture

Type of Stemplot/Stem-and- Boxplot/Box-and-whisker plot Dotplot Ogive/Cumulative Frequency

Graph leaf plot
Picture

Labeled Computer Output:

7
Sampling Method Description Pros Cons
Simple Random (SRS) One chooses the sample so that each unit (and Easy and unbiased Large variation, and must have a
every set of units) has an equal chance of being known population
selected
Examples: names in a hat, random number
generator/table

Stratified (Random) Divide the population into groups (strata) based More precise than an SRS and can be Difficult to divide into groups, more
on a similar characteristic, then use an SRS to cheaper if the groups are already complex than SRS, and must have
choose from EACH group available known population

Cluster Divide the population into groups (usually by Cost is reduced, is unbiased, and Sample may not be representative of
location), randomly select a group and sample don’t need to know entire population overall population
everything in THAT group

Systematic Use a system (every nth number) after choosing Unbiased, the sample is evenly Large variation and can be affected
randomly where to begin distributed across the population, by trends
and don’t need to know entire
population

Voluntary Sample is selected in a way that people do not Easy Highly unrepresentative of overall
have to respond population
Convenience Sample people who are easy or comfortable to Easy Highly unrepresentative of overall
collect information from population

Bias in Sampling Definition

Voluntary Response People choose themselves to participate
General Vocab Definition
Undercoverage Some groups are left out of the sample selection
process Observational Study Treatment IS NOT randomly assigned
Non-response Someone cannot or does not participate in sample Experiment Treatment IS randomly assigned
Response False/incorrect answers (can be intentional or not) Blocking Reduces variation; subjects are grouped based
on a characteristic, then treatments are
Wording of Questions Question is worded so that a certain response is
randomly assigned within the groups
given