Spot the bias type: 1. To treat breast cancer, women often undergo either a mastectomy or a less intrusive procedure called a lumpectomy. You want to know whether there is a difference in the two procedures in terms of the chance that the cancer will go into remission. You compare the remission rates of women who have had each procedure. 2. You want to know if providing prisoners with access to high school and college degree programs while in prison will reduce the chances that they return to prison once they are released. You read a report on this, which concludes: "Our study found that prisoners who enrolled in educational degree programs in prisons were 75 percent less likely to return to prison than were those who didn't." 3. A study concludes that "marriage increases longevity, as married men on average live 5 years longer than unmarried men." 4. "A four-year college degree pays off, as among 40 year olds, those with a college degree earn on average $20,000/year more than do those with only an AA degree. 5. "Our program offers proven results: After only three months enrolled in our program, participants lose an average of 25 pounds." In-depth questions: 1. What sorts of biases were present in the example of how Houston School District used school test scores? (see article linked in materials for Thursday) 2. Describe how one might design a randomized control experiment that has the objective of figuring out if taking calculus in the 11th grade improves English grades in the 12th grade. Make sure you are clear about: how the sample, control and treatment groups are selected; what the treatment is (how control vs treatment groups differ); and what outcome you would measure and how.