Thursday, June 29, 2006

Odds ratios, and statistical errors

Two articles I have noticed recently, rather than one article on both things, or linking both!

First,
a paper in Archives of Dermatology, which looked at statistics in papers in Arch Dermatol and in J Am Acad Derm. 364 original studies were found, published in 2003. 155 had statistics in them. 14 percent of these papers had errors in the statistical methods, 26.5 percent had errors in the presentation of results, and 2.6 percent had errors in both. (Information from the abstract). The authors conclude:

"Readers should critically analyze the methods and results of studies published in the dermatology literature."

Neville JA, Lang W, Fleischer AB. Errors in the Archives of Dermatology and the Journal of the American Academy of Dermatology from January through December 2003. Arch Dermatol 2006; 142: 737-740.

Then, a
paper in the same issue of the same journal, looking at the interpretation of odds ratios and their relation to relative risk. (I keep wanting to type "odds rations", for some reason).

Katz KA. The (relative) risks of using odds ratios. Arch Dermatol 2006; 142: 761-764

No comments: