A long time ago, some enrollment consultant told me something: Students who visit campus enroll at a higher rate than those who do not. I already knew this, of course, as do most people who do admissions for ten minutes.
But then, he came up with this gem: “So the goal of your admissions efforts should be to get as many students to visit campus as possible.” Even then (this was probably 1984 or so), before I had a graduate degree, and after only a couple of undergraduate courses in statistics, I told him he was crazy: Visiting campus doesn’t cause students to enroll; visiting campus is a strong signal of propensity to enroll. If you randomly selected a hundred students from three states away and bused them to campus against their will, it’s unlikely you’d get one to enroll, let alone the 68 we might have predicted (remember those yield rates when students applied to three colleges?)
It was the first thing that popped into my mind when I heard the latest suggestion that all students should be required to take the SAT or ACT in high school, because, when more students take the test, more students–especially those who are largely invisible to colleges–go to college.
Third-grade me says, “duh.”
Of course, post hoc, ergo propter hoc does not always lead to a faulty conclusion, but in this case, I’d suggest there is even less causality here than meets the eye, even though the paper doesn’t explicitly suggest cause and effect. However, I’m not sure the average state legislator or member of the local school board has the perspective or the training to make these fine distinctions. The New York Times has now run with it, and it’s likely to be in the inbox of large numbers of policy makers on Monday morning.
(Whenever I hear the word “selective” in these discussions about getting more students to colleges, it reminds me of the ridiculous work on “undermatching,” that I’ve written about before. People tend to focus on the things that are important to them, and if you went to a selective university, it’s apparently impossible for you to envision how not everyone necessarily wants that for themselves.)
But I digress.
If you work in admissions, you have almost certainly already figured out why kids who take a college entrance exam who otherwise might not have done so end up going to college. It’s not, I’d suggest, the taking of the test: It’s the fact that the student’s name gets sold over and over to colleges who are searching for students. The colleges pick up the slack in the process, because many of them believe low-income students can’t be ruled out because of income. Additionally, the budgets of most colleges and universities are tuition driven, and against decreasing populations of high school graduates, they are generally eager to generate enrollment to keep themselves operating.
If more students who take the test go to college, it’s probably because dozens, if not hundreds, of colleges rush in to tell them they can go to college, and that they should consider it, and that aid can make it possible. That’s a powerful message.
If you don’t believe it, talk to any high school counselor about the institutions that like to generate lots of easily denied applications from students just to be more selective. Admissions offices who infrequently admit students with SATs below 1400 coddle and encourage those with scores of 1100, and no amount of discouragement can persuade the students they don’t have a chance. In the end, the students become grist for the college rejection mill, but the point is, this information barrage works.
So in some sense, requiring students to take the ACT or SAT might be a good faux solution to the larger problem, but it brings with it problems and costs people overlook. And mostly, I think, it fuels our mis-directed obsession with standardized tests as a measure of something important. We’ve had lots of lawsuits in admissions that could have been avoided if the public didn’t believe that an SAT was an “academic qualification.”
Study after study has shown that the ACT and SAT add virtually nothing beyond what is already in the file to the predictive equation when trying to understand who will do well in college. At my institution, tests uniquely explain about 2% of the variance in first-year performance, a figure similar to many other studies.
(An aside here: People like tests for a lot of reasons, as Robert Sternberg has pointed out. To his list, I’d add the very low rate of false positives on the tests. That is, if you score very well on a test, you clearly have a certain kind of intelligence: The kind that allows you to pick the “right” answer from the four given. College, at its best, is not like this; neither is life, where you sometimes have to figure out the answer without knowing whether the question is right. Tests alone do predict something, that’s only because tests effectively amplify the signal of high school performance; once you account for high school GPA, saying the same thing again doesn’t make the outcome much clearer.)
The unintended costs of requiring tests would be high, though: More money and instructional time spent on test prep, more teaching to college admissions tests, more benchmarking districts on tests that still measure mostly wealth and ethnicity. The districts and schools with lower percentages of test takers end up looking the worst (regression to the mean), which kicks off all sorts of shitstorms (a phrase, I’ve discovered, even the Germans have adapted from English given the perfect metaphorical image it creates.) And worst of all, it creates more opportunities for the testing agencies to get their tentacles even farther into curriculum development at the middle-, and even elementary-school levels. The relationship between curriculum development and testing is a dubious one, at best.
I like to remind everyone that no one appointed these companies to do this on our behalf; no government agency appointed them. They’ve largely done this on their own. These are private companies who develop marginally effective testing services they sell to the public, and who then sell the information they collect to colleges.
Given the state of politics and education, I’m not sure we can actually believe that states are interested in increasing college attendance rates; the party in charge right now rose to power by winning bigly with people who didn’t graduate from college. But even if you take a charitable view of our representatives, capacity at selective state flagships is unlikely to increase, making this a zero sum game.
Assuming for the sake of argument that states want more students to go to college, they can set up systems to get the student names and reasonable estimates of academic performance to the colleges, who will do this at no cost, or even negative costs, to the states. But don’t try to accomplish this by making your students take a test that is largely useless, and then allow the testing agencies to profit from the sale of the names.
There are willing partners in college attainment out there, and they are, of course, the ones who do the educating. Increasing attainment doesn’t require testing for more, and certainly not for all.