Wednesday, October 21, 2015

Cheating at UGA

Here's the tweet, and then the headline, from our local paper's story on cheating at UGA:
Let's put this into perspective.

These data are a measure of students caught cheating, not of how much cheating is going on at the UGA campus. You get weird blips in the data, like a bunch of students in one chem lab who cheated on the same assignment and got caught -- thus inflating the numbers.

The ABH's story kinda admits this -- in graf #5.
The one-year increase may not be quite as dramatic as it appears. More than 100 of the cases appear to come from a fall semester 2014 management information systems class in UGA’s Terry College of Business.
In other words, it's not unlike the rape story last week reported by both The Red & Black and, in a follow-up, the ABH, that noted the rise in reported rapes from six to 71. In both stories, the explanation was buried several grafs below -- the measurement system changed. Sigh.

Don't get me wrong. Students cheat at UGA, far more than are caught, but the "caught" numbers signify only that, how many were caught, not how much cheating is taking place. Has cheating increased? Find a good way to measure that and then get back to me.

Note: You can find various academic honesty reports here. Haven't found the one referenced in the news story yet. The paper should link to such stuff. That's how it's done on the InterWebs, folks.

No comments: