elterrible78 wrote:There are a couple problems with this, actually. One is that the GPA ranges are enormous, and another is that aside from numbers (which are only very loosely controlled for), the only other factor controlled for is URM status. Based on my analysis (which controls much more strictly for numbers, as well as for URM and nontrad status, gender, ED vs RD, and timing of the application), there actually IS a very significant boost for splitters who apply as early as possible, which might (along with the whole GPA thing) be accounting for the differences in those charts.
I'm not suggesting those charts prove anything one way or another. The ranges are quite large, and were they to narrow the sample size would be extraordinarily small.
But if you're saying that your analysis of the data suggests that splitters who ED early fare better than they otherwise would, it sounds like we agree.