Generate and test a series of related hypotheses

Project directions:

The project goal is to generate and test a series of related hypotheses. Your final product will be to generate a table 1 that shows your results. The educational objectives are to learn how to clean a data set, define variables in a meaningful way, apply your skills in hypothesis testing, and collaborate with a colleague on statistical analysis.

Using the NHANES data set provided to you, answer the following research question: “Are people with a higher education level healthier than those with a lower education level."

Here are some guidelines to assist you in your endeavor:

  1. You will work with a partner
  2. You have been given a data dictionary with the data set
  3. To create a table 1, you will need to dichotomize education level
  4. Generate specific hypotheses to test:

Please use the following variables – beef eaten/month as continuous, BMI as continuous, elevated cholesterol. This will allow us all to compare our approaches to each other. However pick another 2-4 variables that you are interested in testing as well.

  1. Describe your decision making process for why you picked variables for health, why you defined them they way you did as opposed to other options, and how you decided which tests to include
  2. You table should contain some measure related to probability (P value or CI) and effect size for each variable.

Dr. Callas or Pinckey will set aside time at the end of our class lectures to work on the projects in class. We are there to assist you and challenge you as you work on your projects. You should feel welcome to get regular feedback on your project.

If you decide to use your own data set instead, you should do the following:

  • Approve this with Callas or Pinckney.
  • You'll need a dichotomous variable to create your table columns
  • Analyze at least 2 continuous variables, and at least 5 variables total.