Electronic Letters to:
|
|
eLetters: Electronic letters published:
-
![[Read eLetter]](/icons/misc/sectionDOWN.gif)
understanding correlations
- Ian Shrier (19 December 2006)
|
|
|||
|
Ian Shrier, physician, researcher McGill University
Send letter to journal:
ian.shrier{at}mcgill.ca Ian Shrier
|
Dear Editor The authors of this study suggest that the 2d:4d ratio is an important predictor of sporting ability in women. The interpretation is based on the beta-coefficients and p-values from regression analyses of different sports and the authors cite several works that hypothesize about biological mechanisms. I have several questions about the methods. First, if one is a high level running athlete, then they might only play recreational soccer. Did the authors include an elite runner as a recreational soccer player? It would also be helpful for the authors to clarify whether subjects who didn't participate in a sport were excluded from that particular analysis, and if so, they should indicate the number of subjects for each sport. Another concern regards the statistical analysis. Unfortunately, one needs to know more than the beta coefficient and p-value from a correlation to appropriately interpret the results. At the least, one needs to know the r-squared value, and even more appropriate would be for the authors to indicate that they conducted the appropriate tests showing that the data fulfills the assumptions of linear regression. Actually seeing the data in a scatter plot for each sport (or one scatter plot with different symbols for each sport) would also be helpful but with so many sports, this may have been limited by journal space. Even assuming all of these are correct, one should examine the relationships between sports. Why do soccer and running have large negative correlations but gymnastics does not? How is golf similar to soccer and running, but not cricket? How come there was no correlation for skiing in this analysis whereas there was a correlation in the Manning article cited by the authors (reference #11)? The requirements of badminton are similar to those of squash but the correlations were quite different. Given the proposed mechanisms by the authors, why would Martial Arts have a reverse correlation? I look forward to reading the authors' response. Ian Shrier MD, PhD, Dip Sport Med, FACSM Centre for Clinical Epidemiology and Community Studies Lady Davis Institute for Medical Research, SMBD-Jewish General Hospital, McGill University |
|||
