In: Statistics and Probability
Pick one public policy of interest to you and describe how you might use a t-test to provide evidence about the policy's effectiveness. In your post, please address the following questions:
1. What policy are you talking about? (For example: the Affordable Care Act, a major health care initiative that helped people get health insurance)
2. What data might you feasibly get? (For example: survey data on whether people have insurance, collected both the year before and the year after the policy change)
3. What variable would you use? (For example: a variable indicating whether people have health insurance)
4. What would be your null and alternative hypotheses? Why did you choose a one-tailed or a two-tailed test? (I'm going to stop my example now, because this is the part that is covered in the book/lectures)
5. What significance level would you use?
6. Suppose that your t-statistic is just below the t* for your chosen significance level. (For example, you chose the 95% level, and have a sample size large enough to make t* 1.96. You get a t-stat of 1.85.) What would you conclude? What would Ludwig and Phillips want to think about here?