Chapter 9 Test
Chapter 9 - Day 12
Questions to be Sure to Include
A multiple choice question that gives four different ways of calculating the z-test statistic for a test for a proportion.
A multiple choice question where students have to identify the possible ways of increasing the power of a test.
On the free response, you should have two questions. One question asking students to do a significance test for a proportion, and one question asking students to do a significance test. Use old AP questions.
Consider having a follow up question part (b) for each significance test. Here are some options for your follow-up question:
Given your conclusion in part (a), which kind of mistake—a Type I error or a Type II error—could you have made? Explain what this mistake would mean in context.
Why do we check the (random, 10%, Large Counts) condition?
A simulation based estimate of the P-value (seriously 2009B #5 is the best question ever….if your students do well on this question, then you did a great job teaching Chapter 9!)
Give students the confidence interval for the data in part (a) and ask them to explain why the confidence interval leads to the same conclusion as the significance test.
Ask students if the significance test reveals a causal relationship. If the data comes from an observational study, then we cannot infer causation.
The power of this test to detect the alternative (Ha) is _______. Interpret this value.
Tips to Give Your Students
Close reading and careful writing are critical to your success this year.
Be sure to answer all parts of each question.
There are a lot of formulas in this Chapter. Don’t memorize them. Understand them. Use the (general formula, specific formula, plug numbers in, find answer) approach.
When reading a free response question, ask yourself “Is this context about a proportion or about a mean”. Your notation, conditions, and formulas depend on the answer to this question. Pro tip: variables that are categorical can be measured in proportions and variables that are quantitative can be measured with means.