I agree that forest plots do a good job of showing what they purport to. The main issue I see is that people tend to over-interpret subset results when this analysis is meant to be more "hypothesis-generating" for individual subgroups.

Comments