[gchallen]

One thing to always note in these studies, particularly at the college level, is whether students were randomly-assigned to different sections. (I glanced at the study and couldn't confirm that this was how it was done, although they do reference studies that show similar results for random assignment.)

The reason is that registration at universities has become akin to airline boarding, comprising a complex schedule of priorities allowing certain privileged groups to register first—students in various honors programs, students enrolled in the same major as the course, athletes, and so on. As a result, more desirable class times (or instructors) will frequently be populated with more successful or well-prepared students compared with unfavorable times. (And 8AM is about as unfavorable as it gets.)

Whether or not registration prioritization makes any logical sense or serves any objectives other than making some students feel special is another question. (It's certainly not optimized around minimizing time to degree. Because why would we care about that? /s) And clearly it can have the effect of putting students who are already at risk into schedules that place them more at risk. But the effects of non-random assignment can also pollute studies like this.

Why we're holding classes at 8AM at all is another great question. You'd be tempted to answer "because space", but you might be surprised to find that many large classrooms are idled as early as mid-afternoon. Students seem to also dislike afternoon courses, but I suspect that faculty preferences are more the reason that those time slots don't get used.