Thursday, September 2, 2010

Correlation & Causation: NAU Lumberjacks Edition

Earlier this week NPR featured a story about stimulus money used at Northern Arizona University to install electronic scanners to track attendance at large lecture classes on campus.  Attendance is important, the university argues, because attendance improves grades and....(University Attendance Scanners Make Some Uneasy : NPR)
Ms. KAREN PUGLIESE (Vice provost for academic affairs, Northern Arizona University): The stronger a students grade performance in the first year, the far more likely they are to persist at Northern Arizona University and graduate.
I would hope that the time constraint of a morning radio program prevented NPR from featuring the compelling evidence to back up the Vice provost's claim.  Otherwise, I can only conclude that a high ranking official at one our nation's universities has fallen victim to an all too common mistake: confusing correlation with causation.

It's quite possible Ms. Pugliese has the relationship exactly backwards here.  Students who are more likely to persist at NAU for whatever reason (personal work ethic, family pressure, the desire to pursue a certain career, they hate the idea of having a job, the hope of higher pay once they do start working) may be much more likely to have higher grade performance that first year for those very same reasons.

After all, if first year grades were all that mattered, NAU could just hand out A's and save the stimulus money.

For more on correlation vs. causation be sure to check out Economists Do It With Models!

No comments: