Sometimes students need to understand what difference it makes (whether correlations can be interpreted as being causal or not). If correlations allow for accurate predictions, why does it matter what causes what? I like to mention that knowing causal direction is the only way to have an impact on outcomes. For example, say self-esteem is correlated with academic achievement. What difference does the direction of the relationship matter? It is true that as markers and predictors, we can still make predictions from the correlation between the variables. However, increasing self-esteem will only increase academic achievement if the causal direction goes from self-esteem to achievement. If it goes from achievement to self-esteem or if there is some third factor involved, it will be a waste of time to try to increase academic achievement by increasing self-esteem (as some educators evidently found out after squandering millions of dollars).
My colleague David Cater and I have been investigating psychological factors that predict retention. The first step was to give incoming freshman a survey containing a number of different psychological tests and then follow-up the next year to see if there were any differences between those who returned to school the next year and those who didn't. Even though we found a difference between those who came back and those who didn't (and even though we compared means of groups instead of correlating the two variables), this was an ex post facto study so we couldn't draw causal conclusions about the differences. However, once we found out that there seemed to be a relationship between certain study habits and retention, it was possible to design a study where some freshman were randomly assigned to receive training in those study habits while others were not. If the study habit training group was then retained at a higher level, I could be more confident in concluding that it was the study habits causing the retention and not some third factor related to both study habits and retention. If the training had no effect, the problem could be with the training being ineffective or possibly the fact that study habits did not have a causal impact on the retention. Rick Dr. Rick Froman Professor of Psychology John Brown University 2000 W. University Siloam Springs, AR 72761 [EMAIL PROTECTED] (479) 524-7295 http://www.jbu.edu/academics/sbs/faculty/rfroman.asp -----Original Message----- From: Gerald Peterson [mailto:[EMAIL PROTECTED] Sent: Wednesday, March 08, 2006 3:14 PM To: Teaching in the Psychological Sciences Subject: effects of correlational thinking Michael aside, I frequently find psych authors themselves using causal-type language about correlational research. However, I do think we need to help students explore the issue more rather than just provide the mantra of never inferring causation from correlation. --- You are currently subscribed to tips as: [email protected] To unsubscribe send a blank email to [EMAIL PROTECTED]
