A goal of this unit was to teach students that not all data are created equal, i.e., that it is important when analyzing data or making inference from the data analyses of others, to question how the data were collected. In particular, in a study to ascertain the efficacy of some treatment over another, are the conclusions based upon data from a well-designed experiment or an observational study, what are the threats to the validity of the study, and to what extent then should we believe the conclusions?
A primary text for the course was David Moore's Statistics: Concepts and Controversies (third edition, W.H. Freeman, 1991). This is an excellent, non-technical introductory statistics book and chapter one, on experimentation, is excellent and takes many examples from medicine. Another excellent introductory statistics textbook at a slightly higher level is Statistics , by Freedman, Pisani, Purves, and Adhikari (second edition, W.W. Norton, 1991) (FPPA). Chapters 1 and 2 are appropriate for this topic, covering experiments and observational studies through a series of interesting and real examples, a central example being the Salk vaccine field trial of 1954. For introductory material in this area these two sources are hard to beat: they are short, clear, readable, and full of interesting examples. I enclose a provocative article ``Ethics and Statisics in Randomized Clinical Trials'', by Richard M. Royall (Statistical Science, 1991, Vol. 6, No. 1, 52-58) in which the discussion is on the ethics of randomized trials. This and the rejoinders make for interesting reading, certainly good background material and perhaps worth a read by students, especially those who bring up the ethical questions themselves.
I enclose three articles that we used for class discussion: (1) ``An Aspirin a Day ...?'' by Greenhouse and Greenhouse (Chance, 1988, Vol.1, no.4, 24-31), (2) ``Statistics, Scientific Method, and Smoking'' , by B.W. Brown (from Statistics: A Guide to the Unknown, Holden-Day, 1972, 41-51. Note: first edition), and (3) various articles about Channel One News. These articles are simply a few of thousands that an instructor of this topic could choose. I selected ones that showed a variety of study designs and that were on subjects the students would find interesting.