Question: Refer to Exercise. Pick three of the possible problems with meta-analysis listed in Section 25.4 and discuss how they might apply to the reviews in the Cochrane Database.
Exercise: According to a report in New Scientist (Vine, 21 January 1995), the UK Cochrane Centre in Oxford is launching the Cochrane Database of Systematic Reviews, which will "focus on a number of diseases, drawing conclusions about which treatments work and which do not from all the available randomized controlled trials" (p. 14). The reviews will be made available electronically to practicing physicians and will contain meta-analyses of various areas of medical research.
a. Why do you think they only planned to include "randomized controlled trials" (that is, randomized experiments) and not observational studies?
b. Do you think they should include studies that did not find statistically significant results when they do their meta-analyses? Why or why not?