« On Circulation and Truth | Main | West Nile »
September 06, 2007
biostatistics
Yesterday's JAMA (Vol 298, No. 9, pp. 1010-22) had an interesting study on the lack of knowledge and understanding of biostatistics and results among medical residents. (One can only assume that surgery residents were not included for some favorable reason.)
Though 95% believed that these concepts were important, only a quarter had reasonable confidence in their understanding, and even that confidence was overestimated.
Of course the authors opine the need for more effective training in biostatistics. But like the studies of which this study studied the study, the real question is not statistical significance but meaningful significance. If these medicine residents can become effective clinicians without understanding statistical significance as well as they have been taught to think they should, perhaps statistical significance is less significant than meaningful significance. In other words, doctors in training realize that real effectiveness is not learned from the numbers but from practice with real patients, who are statistically underrepresented in studies.
| By Robert Maddox | 10:53 PM