Formula sheet

2020 AP Statistics Exam

● Will Cover All 7 Units From the CED

● Students can expect at least part of a question to cover
○ (1)​ One-variable Data,
○ (2) Two-variable Data,
○ (3) Collecting Data,
○ (4) Probability,
○ (5) Sampling Distributions,
○ (6) ​Inference for Proportions, and
○ (7) Inference for Means.
● Of these units, the ones that are more likely to get ​2+ parts​ would be (1), (3), (6), and (7).

● Prediction
○ Question #1 will consist of 6 parts (25 minutes) 55% of overall score
○ Question #2 will consist of 5 parts (15 minutes) 45% of overall score

● Inference Questions (hypothesis, confidence) will look Different

○ In the past, students could expect to see one question where they constructed and
interpreted a confidence interval​ and one question where they performed a full
significance test. There is no room for this in 2020. Instead, the exam is likely to ask
about ​specific elements​ within each of these inference procedures​.
 Confidence Interval

● Define Parameter(s)
○ Proportion
■ Let ​p​ be the true proportion of smokers in the USA.
■ Let p​1 ​be the true proportion of students with internet access at EHS
■ Let p​2 be
​ the true proportion of students with internet access at WHS
● Let p​1 ​- p​2 ​be the true difference in proportion of students with internet access at EHS and WHS

○ Mean
■ Let be the true mean age of smokers in the USA
■ Let be the true mean age CA drivers
■ Let be the true age of AZ drivers
​

● Let - be the true difference in mean age between CA and AZ drivers.

● Name of Procedure
○ Proportion
■ One sample z interval for P (​FRQ#1​)
■ Two sample z interval for p​1 -​ p​2 ​ (​example​)
○ Mean
■ One sample t interval for (​FRQ#1​) [ used for ​paired data ​]
■ Two sample t interval for - ​ ​ (example ​page.2​)
● Check conditions
○ Proportion
■ Random ​- as stated in the problem
■ 10%​ - the sample size of ​30​_ M&Ms is clearly less than 10% of all M&Ms
■ Normality​ - n*p >10 and n*q > 10 (be sure to check both samples if there 2 samples)
○ Mean
■ Random ​- as stated in the problem
■ 10%​ - the sample size of ​30​_ M&Ms is clearly less than 10% of all M&Ms
■ Normality​ - Since n>30, the CLT applies;
○ If sample size is less than 30, must sketch a graph(dotplot, bar graph) to
show no strong skewed or outliers)
● Interpret a confidence level.
○ If we make many many 95% confidence intervals, we expect about 95% of those intervals to
capture the true mean/proportion of...

● Interpret a confidence interval.

○ We are 95% confident that the interval (a,b) captures the true mean/proportion of ….(context)

● Based on a confidence interval, do we have convincing evidence for a claim?

○ Since 0 is not in our 95% confidence interval, there is convincing evidence that there is a between.
○ Since our 95% interval contains 0, there might not be a difference between ...
 Significance Test

● Define Parameter(s)
○ Proportion
■ Let ​p​ be the true proportion of smokers in the USA.
■ Let p​1 be
​ the true proportion of students with internet access at EHS
■ Let p​2 be​ the true proportion of students with internet access at WHS
● Let p​1 ​- p​2 ​be the true difference in proportion of students with internet access at EHS and WHS

■
○ Mean
■ Let be the true mean age of smokers in the USA
■ Let be the true mean age CA drivers
■ Let ​be the true age of AZ drivers
● Let - be the true difference in mean age between CA and AZ drivers.
■

● Name of Procedure
○ Proportion
■ One sample z test for p (​page#2​)
■ Two sample z test for p​1 ​- p​2 ​ (​example​)
○ Mean
■ One sample t test for (​FRQ#1​) [ used for ​paired data ​]
■ Two sample t test for - ​ ​ (example ​page.2​)
● Check conditions
○ Proportion
■ Random ​- as stated in the problem
■ 10%​ - the sample size of ​30​_ M&Ms is clearly less than 10% of all M&Ms
■ Normality​ - n*p >10 and n*q > 10 (be sure to check both samples if there 2 samples)
○ Mean
■ Random ​- as stated in the problem
■ 10%​ - the sample size of ​30​_ M&Ms is clearly less than 10% of all M&Ms
■ Normality​ - Since n>30, the CLT applies;
○ If sample size is less than 30, must sketch a graph(dotplot, bar graph) to
show no strong skewed or outliers)
● Calculate p-value
○ Proportion - use calc: 2nd vars normalcdf( )
○ Mean - use calc: 2nd vars tcdf( )
● Interpret the P-Value
○ Assuming that the null is true ( add context) there is a (p value goes here ) probability of getting a
sample mean/proportion of (insert sample value) purely by chance.
● Conclusion of a significance Test
○ Since p-value is less than alpha, we reject the null. There is convincing evidence to support that
the true mean/proportion ( add context) is (insert the Ha here)
○ Since the p-value is greater than alpha, we fail to reject the Null. There is not convincing evidence
to support that the true mean/proportion ( add context) is (insert the Ha here)
 Deciding between a ​one sample or two sample​ T Test/Interval

To help decide between one sample and two sample t-procedures, students
should ask themselves:

■ HOW MANY SAMPLES DO I HAVE?

■ ONE​ ​-

● EX ​page 5​ ​page 4
■ TWO​ -

● ​example

(Usually) BUT watch out for two samples that pairs ​individuals

■ CAN ​ANY PIECE OF DATA​ IN THE FIRST GROUP BE COMPARED TO ​ANY PIECE OF
DATA​ SECOND GROUP

○ NO​ -

○ YES​ -

■ Yes​ page. 4

■ DO THEY REFERENCE PAIRING THE DATA

● YES -

○
● NO-

○
Final note (important)

1. Watch your language: use the right notation

a. Example: ​use P or P(hat) when talking about the mean.
2. Always include context in your responses.
3. Print Mr. Nguyen cheat sheet, and ​formula sheet
4. Charge your Calculator
5. DO NOT spend too much time on any part of the a question
6. Naked answer = no credit: write down numeral answer and answer the question in its context.
7. Less is more. Answer the question in context and shut up.
8. Don’t argue w/ the question writers.
9. Describe a Distribution
a. CUSS + BS
10. Compare Distribution
a. Use comparison words ( more than, greater than, smaller than,larger than)
11. Practice ​Test #1​ Answer: ​Key