Are SAT/ACT scores good predictors of student success?
- Follow the lead of Dartmouth College and Harvard University, Yale University recently announced it will no longer require incoming applicants to submit scores for the optional timed essay portion of the SAT and ACT. Meanwhile, admissions directors continue to debate just how useful the test is at all to predicting student success, and whether that predictive ability is worth compromising access, given the noted cultural biases in these and other standardized tests.
- At last week's New York Times Higher Ed Leaders Forum Conference, NYC Lab School for Collaborative Studies director of college counseling Kay Rothman and Rice University vice president for enrollment Yvonne Romero da Silva disagreed on whether the tests had any value. Rothman expressed her belief that they're only indicative of first-year success while da Silva said at the institutions she's worked for, University of Pennsylvania and Rice, they've been predictors of success past the first year. But da Silva emphasized the importance of recognizing "one-size doesn’t fit all." "I think that the important way to approaching test scores is to ... see how [students] perform."
- For admissions offices that are strapped for time, test scores can provide a place to start to identify whether or not a student will be successful, but da Silva also said there has to be a commitment to reviewing each application and reading the students' stories as well. "At the end of the day, that’s what it’s about: Whether or not the student will be successful within my curriculum, and keeping in mind that you don’t want to shut the door," she said.
If Rothman had her way, institutions would go test optional, and they would also stop looking "at the rigor of the curriculum based on APs, because what it does is it makes young students take courses they don’t care about."
But both women agree stronger advising is needed at the K-12 and higher ed levels to help students evaluate what colleges and universities they can be successful at before they apply. And, Rothman said, admissions directors at college fairs should stop encouraging students whose test scores and GPAs fall way below their institutions' averages to apply anyway, a perpetuation of the "recruit to deny phenomenon" that allows institutions to boost rankings, but that is not in the best interest of students. da Silva agreed, saying "I think there is a moral and ethical element to how you recruit and to whom you recruit."
The idea of boosting rankings, which still dominates a number of institutional priorities and practices, and doing the most an institution can to increase college access and promote student success can stand in direct conflict with each other. And though rankings generate a lot of money for the entities that release them and publicity for the institutions which score high, they do not actually paint a good picture of which institutions are doing a good job of increasing access, graduating happy and well-prepared students or closing equity and attainment gaps to keep pace with a rapidly-changing national population.
In fact, pursuit of high rankings scores inherently means blocking out students who need the access to higher education to achieve social mobility the most, because the rankings and federal scoring systems create a disincentive to admit lower-income students and underrepresented minorities.
Institution leaders are faced with the challenge of reconciling their missions with regional needs and the desire for prestige. There are 115 Research 1 institutions in the country, 107 Research 2 and 112 Research 3 institutions. The U.S. Department of Education tracks more than 7,000 institutions, though it is commonly accepted that around 5,000 of these are colleges or universities — meaning 334 of roughly 5,000 institutions, or 6.7%, are set up to compete in the rankings.
In the words of Red Deer College President Joel Ward, "Everybody wants to be Harvard, but there’s only one Harvard — the rest of us need to do our jobs."
Follow Autumn A. Arnett on Twitter