The Wall Street Journal (September 14, 2007) contains the headline “Most Science Studies Appear to Be Tainted by Sloppy Analysis” The article is based on the work of epidemiologist John Ioannidis who has claimed that “most publish research findings are wrong (Why Most Published Research Findings Are False; PLoS medicine 2005;2:696-701). He notes that the probability of a true finding is related to the size of the study, the magnitude of the effect size difference, the number of tested relationships, the number of other studies on the same or a similar question, and the flexibility of design. The author notes that the prevalent practice of examining a study in isolation from all of the other work in the same area greatly increases the probability that a given result is due to random fluctuation.
David S. Smith, M.D., Ph.D.
These posts keep getting better and better..
Posted by: Term papers | November 05, 2009 at 12:28 AM