Correlation, not causality
Published October 22, 2008
As the University of Minnesota strives to become one of the top three public research universities in the country, it would appear this effort has not trickled down to more accurate reporting of research in The Minnesota Daily. MondayâÄôs article on a Boynton Health Services survey concerning stress and student grades reported the study as if it were evidence that higher stress causes lower grades. However, from what I can discern of the methodology from the article, researchers demonstrated no more than a correlation between GPA and stress. Even a poor student of social science methodology can usually remember that âÄúa correlation does not establish causality.âÄù A correlation cannot tell us if high stress is causing poor grades or âÄî the very (if not more) plausible alternative âÄî if poor grades are causing high stress. I have to hope the error was in the reporting of the Daily, and not that the BHS individuals conducting the study were this poorly versed in the basics of social science research. Either way, we seem to be a long way from our goal of becoming one of the top three public research institutions in the world. Chris Miller University student