Nobody can–or should–be defined by a number. Nonetheless, numbers matter quite a bit in the college admissions process. But should they?
If the numbers up for discussion are high school grades, just about everyone agrees on the importance of evaluating past academic performance as a predictor of future academic success. Despite dramatic variance in grading practices from state to state, school to school, and even teacher to teacher, applicants usually embrace being defined by these numbers, at least when they’re good.
But when the numbers in question come from standardized tests, the conversation changes. So many critics reject the value of scores from admissions tests like the SAT or ACT that these educational institutions require defending from time to time. The College Board asserted the continuing predictive validity of the newest iteration of the SAT not long ago. Now, a champion steps forward to explain why test scores remain vital in college admissions. Wayne Camara, Senior Vice President of Research at ACT, has an extensive background in industrial organization, admissions, and both educational and psychological assessment:
The biggest misperception I see is the argument that high school grades are the best indicator of college success and, therefore, we don’t really need standardized admissions tests. This notion is misinformed. That’s a polite way of saying it is nonsense.
Let’s start with the fallacy in the argument. High school grades are not, in fact, the best indicator of college success. Neither are test scores alone. In fact, the best predictor of success in college coursework is the combination of the two—grades and test scores together. Hundreds of independent studies have shown this to be true…
High school grades have their limitations. They not only reflect the idiosyncrasies of individual teachers’ grading standards and differences in course rigor, but they also contain inflation. More than 55 percent of college-bound students report having high school grades above 3.25, and 25 percent of U.S. high schools report an average GPA of 3.75 or higher for their graduating class.
We often accept the hypothesis that grades are fair and unbiased indicators of future success without much scrutiny, but grade inflation has steadily increased in the past few decades, and it has increased more rapidly for white and Asian-American students coming from more highly educated families.
Educators acknowledge that there are differences in the quality of schools and the rigor of curriculum. Test scores are one measure that helps colleges navigate and mitigate those differences, allowing them to compare the preparation of students coming from different backgrounds and different experiences. Without test scores, colleges must rely on their own subjective impressions of different groups of students and the quality of different high schools. We know that subjective impressions and decisions have biases which are often implicit and never as accurate as empirical data.
Admissions tests provide a common metric that allows colleges to evaluate students who attend different high schools, live in different states, complete different courses, and receive different grades from different teachers. High school GPA simply cannot do that…
There is no single measure that can definitively predict future behavior by itself, and all measures have limitations. The best decisions are made when multiple sources of data are considered. There is no reason to ignore test scores, just as there is no reason to ignore previous accomplishments, high school grades, or personal factors that have influenced a student’s development and aspirations.
Our ultimate goal should be to help students land where they have the best possibility of success, and there is no question that admissions test scores help accomplish this goal.
Mr. Camara’s full article adds further evidence and persuasive logic to a very cogent explanation of the value of admissions tests in the application process. Colleges from the highest tiers to the lowest value the ACT for its proven validity, reliability, and fairness. If you’re going to be defined by a number, at least this one can be trusted.