3 ms·
> Neither is perfect (both operate at about 0.5 level in validation studies), So not much better than a coin flip?
by woodchuck64 14y ago
> Neither is perfect (both operate at about 0.5 level in validation studies),
So not much better than a coin flip?
- yummyfajitas 14y agoSuppose you flip a coin on 100 candidates, one of whom is qualified. You'll get about 50 "hire" answers, and the odds are 50% that one of those hires is actually qualified. So no, a coin flip is far worse.
- woodchuck64 14y agoThe article is referring only to employees that were hired and were also given the GMA, so qualification was not an issue. However, using a coin flip rather than a GMA score on a prospective employee would imply a score of either 0 (missed 'em all) or 100 (perfect), and thus would likely have 0 correlation with the actual employee performance, rather than .51. So your conclusion is right even if your analysis is not.