Students will explore the Smoking Behavior database to analyze many queries and generate a new hypothesis, which they will integrate into a new hypothetical case control study.
Class Time: 100 minutes (2 class days)
|Students will understand how to query the Smoking Behavior database to look for patterns among case and control subjects to generate hypotheses for future studies||Students demonstrate that they can 1) interpret data in the “Hypothesis Generation” view of the database by recognizing differences between cases and controls; 2) define exposed and not-exposed to achieve the strongest association; and 3) generate new hypotheses based on the patterns they observe in the data|
Section A. Reviewing Hypothesis Testing and Hypothesis Generation
1. If needed review Figure 6.1- Comparison of Hypothesis Testing and Hypothesis Generation to review the difference between Hypothesis Testing and Hypothesis Generation.
Remind students that when they were testing their specific hypotheses, they were limited to using four questions only (due to the problem of multiple comparisons), and they had to define their specific hypotheses, exposed and not-exposed, and the study population before looking at the data.
Explain that in hypothesis generation these constraints don’t apply because they aren’t going to make any scientific claims based on their queries. Instead, they are going to look for trends in the data and propose a new hypothesis that they would want to test in a future experiment. They will also provide some information about the design of the future experiment.
Section B. Modeling How to Explore the Data
Complete Student Sheet 7.1. Using Hypothesis Generation to Explore Data as a class.
1. Project the database on the screen for your students to view it: https://gsoutreach.gs.washington.edu/database2/
2. Select Step 1.4 Hypothesis Generation, and then Question 12, “How old were you when you first tried a cigarette, even a single puff?”
3. Guide the class in selecting how they will define exposed and not-exposed for Queries 1, 2, and 3. Ask students to record the data from the three queries on Student Sheet 7.1. You may want to use Figure 7.1. Using graphs to look for trends in the data to help students think about how to get the highest odds ratio.
4. Have students work individually or in pairs to answer Questions 1 and 2 at the bottom of Student Sheet 7.1. Then hold a class discussion about the trends they observe as they change the exposed/not-exposed and study population. Students should notice that having more cases than controls who are exposed, and more controls than cases who are not-exposed leads to a high odds ratio. They should also notice that when the study size decreases by selecting males or females for the study population, the 95% confidence interval becomes broader, making the results less accurate.
|Figure 7.2. Using the 2×2 table to look for trends in the data shows some tips for getting the largest possible odds ratio for any one question. Students may need some help in understanding these tips and linking them to the observations they made in Student Sheet 7.1.|
Section C. Using the Database to Generate a New Hypothesis, and Proposing a New Case Control Study
This activity guides research teams in considering research topics and questions they might want to explore in order to generate a new hypothesis and propose a new case control study.
- Ask students to work in their teams to Blank_ Research Project Page_Lesson 7a. Planning your Hypothesis Generation Research . After completing this sheet, students should form pairs and divide the questions they plan to query between the two pairs. Point out that several questions, with multiple response answers, cannot be used to calculate an odds ratio. For these questions (purple questions from Student Sheet 4.3) students should graphically interpret the data.
- Direct the student pairs to complete their queries using the database and record their data in Blank_Research Project Pages_Lesson 7b. Conducting Queries in Hypothesis Generation, following the instructions in Directions for RPP. Students will need an additional copy of RPP-14 if they do more than four queries.
- When teams have completed their queries, they should get together with their teammates to complete Blank_Research Project Pages_Lesson 7c. Hypothesis Generation and Proposing a New Case Control Study , which guides them to generate a new hypothesis and propose a new case control study in which to test it.