Two articles in the Chance Wiki (September 6, 2005).
This page is moving to a new website.
I submitted two articles in the July-August 2005 Chance Wiki. These articles, which originally had separate links are now consolidated in
The first article, "Can you get fired over the wording of a questionnaire?" was mentioned on my weblog several times before I drafted up something for the Chance Wiki
- Publication in the ChanceWiki (August 2, 2005)
- Allegations of scientific misconduct (updated August 1, 2005)
- Interesting web links and quotes for the month of May (updated May 23, 2005)
The second article "Racial Profiling" is an interesting case study, but because of the politics involved, I felt it was safer not to mention it on this weblog. I don't want to abuse the good will of Children's Mercy Hospital by using the web site they provide me for something that could be interpreted as a political rant.
Interestingly, both articles received minor editing by other contributors, and all of these edits improved the final product, in my opinion.
There are several articles already submitted for the September-October issue,
including a very good one "Why Medical Studies are Often Wrong." This entry in the Chance Wiki comments on John Allen Paulos's commentary on a recent JAMA article:
- Contradicted and initially stronger effects in highly cited clinical research. Ioannidis JP. Jama 2005: 294(2); 218-28. [Medline]
There's another interesting article published by JAMA with Ioannidis as a co-author:
- Relative citation impact of various study designs in the health sciences. Patsopoulos NA, Analatos AA, Ioannidis JP. Jama 2005: 293(19); 2362-6. [Medline]
and Ioannidis was co-author on a couple of fascinating papers which examined the hierarchy of evidence:
- Large trials vs meta-analysis of smaller trials: how do their results compare? Cappelleri JC, Ioannidis JP, Schmid CH, de Ferranti SD, Aubert M, Chalmers TC, Lau J. Jama 1996: 276(16); 1332-8. [Medline]
- Comparison of evidence of treatment effects in randomized and nonrandomized studies. Ioannidis JP, Haidich AB, Pappa M, Pantazis N, Kokori SI, Tektonidou MG, Contopoulos-Ioannidis DG, Lau J. Jama 2001: 286(7); 821-30. [Medline] [Abstract]
The first study looked at a large number of meta-analyses that included one large study and compared the result of the large study to a summary of the remaining small studies. Although there are some differences between the large study and the meta-analysis of the smaller studies,
Clinically important differences without a potential explanation are extremely uncommon.
The second study examined 45 diverse topics for which both randomized and and nonrandomized studies had been performed. In this study, the authors concluded that
Despite good correlation between randomized trials and nonrandomized studies-in particular, prospective studies-discrepancies beyond chance do occur and differences in estimated magnitude of treatment effect are very common.
[On June 7, 2006, I fixed some minor typos]