The study is based on the Dunedin Multidisciplinary Health and Development Study which is following a cohort of just over 1000 babies born in 1972 in Dunedin.
This video requires a web browser capable of HTML5 or Flash video.
TasksThe file Herpes data.xls has the data for 1890 men recorded in nine variables. The file Herpes variables.xls names the nine variables and defines the classification categories of each variable.
The data used for this study on whether childhood circumcision reduces the risk of acquiring genital herpes has been modified for teaching purposes by supposing a larger cohort of 1890 men. The true results, however, are given at the end of the video. (Note that the concept of relative risk is in the school syllabus).
- Task 1: Using Herpes data.xls calculate the cross tabulations presented in the video.
- Task 2: Calculate the prevalences (proportions) at age 26 of herpes and construct the confidence intervals for each proportion as well as the confidence intervals for differences between the proportions. Note any significant differences.
- Task 3: Calculate the relative risks. (The confidence intervals for the relative risks are outside the school syllabus. They are not symmetrical about the point estimate but have an interpretation which is similar to other confidence intervals. The value 1 for a relative risk is the crucial decision value because then the two risks are equal).
- Task 4: Discuss the importance of a designed study as opposed to an observational study. The discussion should cover adjustment for confounders either by stratification described in the video or by constructing a designed (clinical) trial to control confounders.
- Task 5: For the groups stratified by the number of sexual partners calculate the cross tabulations from the data, and the confidence intervals for proportions and differences between proportions. Discuss any significant differences and compare with results from Task 2.