culture

Abolish the SAT?

Genesis
Response
Penultimate
Finale
January 18th, 2022
Let me start in the middle first. The SAT is in fact a strong predictor of college completion; more than that, it's a strong predictor of many future life events. Study after study has shown that the SAT predicts freshman college performance very well. In fact, research published in 2010 found that giving 12-year-olds SAT tests provided strong predictive information on whether they would go on to get a PhD, get tenure, even whether they will hold a patent. The SAT is a highly predictively valid test. So why do some data sets suggest otherwise? Because we have to adjust for restriction of range mathematically.
Why do we have to adjust for range restriction? Think about it. Studies that compare SAT results to college grades can only collect data for people who went to college. But there are also people who take the SAT and don't go to college. Removing them from the data pool before analysis reduces the strength of the correlation. If we do range restriction adjustments, we find the test is a strong predictor.
Yes, GPA also does a good job of predicting college performance. But remember that it's possible for two indicators to show strong relationships without perfect or very strong overlap. And in fact there are students who flourish on the SATs while struggling with high school grades - including poor and minority students. We should want students with a diversity of strengths in our institutions. Managing a high GPA over four years typically entails stability at home many students don't enjoy, especially poorer students. The SAT allows them to demonstrate their talent in a short time frame, without the pressures of daily attendance and long-term work.
Here's the problem with saying that the SAT shows race and income biases: so does GPA, the metric nominated to replace the test! The Nation's Report Card, which contains data from a variety of academic sources, shows that Black and Hispanic students perform worse in GPA than white and Asian. And academic research going back decades shows income effects on GPA as well. To de-emphasize SAT in order to emphasize GPA more does nothing to reduce the impact of race and socioeconomic class on college admissions!
Indeed, race and income effects are present in all educational data - yes the SAT, but also GPA, the NAEP, state standardized tests, graduation rates, and all manner of academic research data. That's because the problem does not lie in the SAT but in our society. Racial and class stratification is what we should expect, given the burdens that poor students and students of color face. We should certainly make every effort to address those burdens; I support race-and-class-based affirmative action for this reason. But we can't get rid of them by banning the tests that reveal them, which is like getting rid of a fever by throwing away your thermometer.
0 Comments