Important Note

Tufts ended funding for its Open Courseware initiative in 2014. We are now planning to retire this site on June 30, 2018. Content will be available for Tufts contributors after that date. If you have any questions about this please write to

Tufts OpenCourseware
  • To enhance the students’ understanding of case control studies
  • To introduce students to the process of critiquing a case control study

Outside Preparation: Due for Small Group #4


An investigator has recorded data, including salary and the happiness index score, for 1500 randomly selected middle-aged adults. Both of these variables are continuous with normal distributions. She wants to determine if knowing salary helps to predict the happiness index score. Using a linear regression model, she calculates a coefficient of determination of 49%.

  1. Write a sentence interpreting the 49% coefficient of determination.
  2. The investigator then wonders if her data indicate that salary, age, gender and happiness index score can predict whether or not hypertension has been diagnosed. What statistical analysis should she use?
  3. When preparing her analysis for # 2 above, she thinks that daily exercise might be a potential confounder. What should she do in her analysis to address this concern?
  4. The investigator then wonders if there is a relationship between gender and the happiness index score. She wants to calculate a Pearson's correlation coefficient? Is she correct? Explain your answer.
  5. A medical student working with the investigator calculates a Pearson's correlation coefficient for salary and the happiness index score. He places the happiness index score on the Y axis and salary on the X axis. He then reverses this such that the happiness index score is on the X axis and salary is on the Y axis. Would you expect the r value to be different based on which method is used? Explain your answer.


  1. Review Lecture 2 - Observational Studies notes on case control studies
  2. Read the paper: A Case-Control Study of Baldness In Relation To Myocardial Infarction In Men; JAMA, February 24, 1993, Vol 269, No 8, pages 998 - 1003; bring a copy of the article to small group

Approximate Class Schedule:

30 minutes Instructor review of key concepts from Lectures 9 and 10
15 minutes Review of homework assignment
45 minutes The class will be divided into four small groups to answer the questions
30 minutes Class discussion of questions

To Be Completed In Class

Answer the following question regarding the paper A Case-Control Study of Baldness in Relation to Myocardial Infarction in Men:

  1. Write the definition of a case and the definition of a control.
    1. Case:
    2. Control:
  2. Why did the authors do this study?
  3. Do you think there could be information bias, i.e. incorrect information as it pertains to the type and extent of baldness?
  4. Were the interviewers blinded? If yes, explain why. If no, indicate how this could have lead to interviewer bias, i.e. inaccurate information collected by the interviewers?
  5. Do you think it is a potential source of bias that some interviews were conducted by telephone and some in person? If no, explain your answer.
  6. In the objective section, the authors talk about the risk of myocardial infarction. In the data collection section they talk about cases having a first MI. In the conclusion section they talk about coronary artery disease. Is this permissible? Explain your answer.