Analysis: Studies With Large Findings Often Proved False

Medical studies that claim "very large" effects rarely hold up when tested by other research teams, according to a new statistical analysis published Wednesday in the Journal of the American Medical Association. The analysis also found that studies that claimed to have a very large effect tended to be small, with fewer than 100 participants. In addition, such studies typically measured intermediate effects, rather than more meaningful outcomes, such as incidence of disease or death, according to the analysis.

• "Medicine Rarely a Slam Dunk, Despite Splashy Studies" (Joelving, Reuters, 10/23).
• "Medical Studies With Striking Results Often Prove False" (Brown, Los Angeles Times, 10/24).