Wednesday, August 06, 2014

Battlefield Decisions

Every August, this happens.  It’s time to start making decisions about which sections are likely to enroll enough students to run, and which need to be cancelled.

If you’ve never been in on a meeting like that, you may picture it something like this:

Dr. Evil (stroking cat): Prof. Doe really loves this class, but he looked at me cross-eyed once.  Screw him!  Bwa-ha-ha-ha-ha!

Group: SCREW HIM!  (maniacal laughter ensues, punctuated by flashes of lightning)

The truth is more pedestrian.  

Dean: This is the only daytime section of this class.  

VP: Is it a program requirement?

Dean: Yes, for the Nuclear Basketweaving option.

Academic Advisor: That timeslot usually fills in late August.

VP: Okay.

These decisions are never fun.  You don’t appreciate how many variables are in play until you get down to cases.  Suddenly, every bright-line rule has an extenuating circumstance, mandates conflict with each other, and much hinges on educated guesses about enrollment during the rest of August.  Could the students in the evening section move to online?  Is this part of someone’s full-time load?  Is the course required for the major?  What if it doesn’t run?  Do other courses fulfilling the same requirement still have seats?  Does it have prerequisites that make a ninth-inning rally unlikely?  Does the student population for that class tend to register late?  What’s the pattern over the last few years?

All of which makes perfect sense, as long as it isn’t your class.

We don’t cut off enrollments until right before classes start, so the enrollment figures we have now won’t be the final ones.  That necessarily involves some projection.  Will the twelve students become fifteen, or is the section stuck at twelve?  Since we can’t replay the scenarios later, there’s no way of knowing whether a section we cancelled early would have rallied at the last minute.  We probably get some of them wrong, but there’s no way of knowing the ones that would have rallied.   We only see the inevitable few that we thought would rally, and didn’t.

When I’ve discussed this sort of thing in the past, people have responded that Data Analytics would save us.  I’m a fan of carefully-applied data, but it works much better in the aggregate than in individual cases.  If enrollment is down, say, two percent across the college as a whole, that two percent is not evenly distributed.  And last year’s pattern may not match this year’s.  (The major countertrend is the move to online enrollments.  That magnifies the decline in onsite enrollments.)  Analytics might tell me that there will be a general downward drift in, say, English, but it won’t tell me whether the 2:30 section on Tuesday will rally or not, any more than the latest public opinion poll will tell me what Dave thinks.  

And then there are the qualitative issues.  If a particular, though small, cohort of students has been jerked around for a couple of semesters, there’s an argument for giving the benefit of the doubt this time.  A brand new program will often need a little time to find its constituency.  And there are always the small end-of-sequence classes that students in a given major need for graduation.  We’re getting better at handling those, but they’ll probably never go away completely.

In a more perfect world, we’d have a long gap between the close of enrollment and the start of classes, so we’d be working with fixed numbers.   Alternately, enrollments would be easily and consistently predictable from year to year, preferably with sustained and gradual growth.  Getting really utopian, we’d have enough money, staff, and space that we could base these decisions entirely and only on academic considerations.  Alas, no.

The good news, having been through this at three different colleges now, is that I’ve honestly never seen this process used to “get” anybody.  That may be small solace when it’s your class that’s on the chopping block, but it’s true.  Battlefield decisions aren’t perfect, but we know which side we’re on.