I worked at a large online university for a few back.
Somewhere along the way I got keys to an AWS account and few interesting diverse data sets (attendance, grades, interactivity, etc). I was just having fun learning ElasticMapReduce until I stumbled upon some very interesting results. We had awesome attendance data, since all the classes were online. If you graphed attendance by number of class logins per week (or to a lesser degree minutes spent in class, lectures attended, etc.) you'd notice it dropped over the course of the term for all final letter grades as well as drop's and incomplete's. This makes sense, particularly for the demographic (lower end of higher ed, received credits from a few other schools). We considered ourselves the school of last resort, for student's who'd bounced around other institutions without earning a degree.
What was startling was that the downward slope of the drop in attendance over the course of a term was directly proportional to the outcome (grade), with one fun exception (D's where there was a dead cat bounce as the end of the terms).
It got to the point I could tell whether someone was going to pass a class on the number of days attended in the first week of class. This pattern held across all courses, but was especially pronounced in intro level ones. We turned this into a gamification effort and had some pretty nice results, but the academics weren't as interested as one would have thought (perhaps because I am not one?).
Someone above mentioned government mandates, and I'd wager I wasn't the first to notice this (either with data or via experience). While the slashdot crowd might have done fine skipping class, most simply aren't. If Uncle Sam is footing most of the educational bill, shouldn't we as taxpayers demand steps are taken to ensure methods to maximize the return on investment (courses passed, degrees earned, etc)?
I'd also like to reiterate that we, slashdotters, are not the norm. Sure I skipped class senior year to code Netscape's LiveWire at work, but that if far from the norm.
Some interesting data here:
https://nces.ed.gov/programs/c...