Monday, April 17, 2017


Stakes Aren’t Cheap

Emily Hanford has a good piece in the Washington Monthly on the seemingly inexplicable survival of Accuplacer.  Check it out.

It’s an attempt to answer the question of why a universally criticized placement test manages to survive, despite ample data suggesting its flaws.  

The Accuplacer is a placement test often administered to students during the admission process.  It has several components, each addressing a different skill: reading, writing, arithmetic, and algebra. It typically takes a few hours, and students often take it “cold” or very nearly so.  Colleges use the score on Accuplacer to decide whether students need remediation, and if so, how much.  

As Hanford notes, studies of students who “disobey” their Accuplacer score suggest that despite its name, it isn’t terribly accurate.  More students tend to get shunted into remediation than need it.

That was one of the shocks when I moved from the for-profit world to the community college world.  For the for-profits, retention was a survival issue, so they saved remediation for the very most desperate students.  It was very much the exception.  At most community colleges, including my own, the overwhelming majority of students (usually ⅔ or so) require at least one developmental course.

States and colleges that have experimented with multifactor placement have been able to reduce that proportion and to increase completion rates.  So why don’t we all do that?

In a word, cost.  Both money and time.

The Accuplacer may be deeply flawed, but it’s fast and cheap.  We can get a lot of scores quickly.  For a sector in which many students make up their minds at the last possible moment, and for which budgets have been cut for years, that’s no small thing.

Multifactor placement -- in other words, looking at high school course selection and GPA -- is much more labor-intensive upfront.  It requires getting transcripts, and evaluating them intelligently and quickly enough to get students lined up for the start of classes.  Getting high school transcripts in July or August can be an issue; students who apply to the more selective schools that require transcripts typically do so by April.  Then, even if we get them, we have to compare them.  Selective institutions spend money on staff to do that; historically, most community colleges haven’t, because there was no need to.  Given enough money, we could, but given enough money, we could do a lot of things.  When any full-time hires are at a premium, this need tends to fall to the bottom of the list.

There’s also a real, if somewhat knee-jerk, sense among many faculty that any attempt to allow more students to bypass remediation constitutes a lowering of standards.  Even with data, many resist the attempt on ideological grounds, often accompanied by anecdotes about students who were obviously overmatched in the past.  I understand the impulse: it’s frustrating to get students in 100-level classes who don’t seem to have mastered high school skills.  But it’s also frustrating to see so many students walk away, disgusted at paying for courses that don’t count, when we know statistically that many of them shouldn’t have had to in the first place.

That sense of holding the line on standards -- even arbitrary ones -- gets a perverse boost from legislative fiats in other states.  When Connecticut or Florida decides to restrict remediation legislatively, it just feeds the narrative that barbarians are at the gates. It’s hard to advocate for thoughtful reforms when folks are wondering if they’re Trojan horses for thoughtless ones.  People connect dots, even when they aren’t really connected.

So the Accuplacer lives by a sort of default.  It isn’t terribly accurate, but it’s fast, it’s cheap, and the political battles have already been fought.  Moving beyond it (or tests like it) makes a lot of sense on the policy level, but it requires resources.  You don’t kill zombies without stakes.

You want the dots connected? There is a proposal to cut CC funding further and use it to provide more support for "job creating" research at the flagship universities while trying to cut taxes. They don't know, or don't care, or don't care to know, that a substantial fraction of students in those research fields started at community colleges. The money will come from "just in time" remediation efforts that help future STEM students catch up from their abysmal HS math classes.

Apropos your lead observation, the tragedy of Florida is that the legislature made it illegal to give the state placement test to those HS grads exempted from remediation, even for pure advising or research reasons. Imagine if we could have had detailed data on the success at various levels of math coupled to data on placement scores for students who skipped remediation. This would be data about everyone, not just the clever ones who found a way to evade the system. Fine grained data like that would identify just how bad the tests were and help set new cut scores for both true remediation and additional supportive classes. But, no, they didn't fix it even when it was pointed out to them.

As for multifactor placement, I consider it outrageous that this cannot be done with computers -- right now, today. HS transcipts should arrive electronically, already coded so people don't have to read them. (Or, at minimum, so ones from your own state don't require human interpretation.) Within one year, people in the state education bureaucracy could generate the heuristics to automate advising and placement for math and English based on HS classes, SAT or ACT scores, HS exit exams, and a placement test. The only cost would be replacing lawyers in the state bureaucracy with people who can do things, like think and use computers.
You said " For the for-profits, retention was a survival issue, so they saved remediation for the very most desperate students."

Could you expand on this? I'm confused about how for-profits would want to do less remediation. The two things I"m confused about:
1) If the student needs remediation it's because they lack the skills/knowledge to move into a higher level course (at least, in a perfect world this should be true).
If they're bumped into a higher level course doesn't that just mean that they'll fail at a higher rate there?

2) Don't for-profits want people to pay for as many courses as possible?
Wouldn't it be great for their bottom line to have students take more (remedial) courses?

I'm genuinely confused about these and would love to know where I'm going wrong.
Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?