In Monday’s New York Times, some new research is cited to argue that performance on Standardized Tests  (SAT and ACT) predicts a high GPA for students from “Disadvantaged High  Schools.”

As Christina Paxson, the president of Brown University, recently wrote, “Standardized test scores are a much better predictor of academic success than high school grades.”

 

Source: Friedman, Sacerdote and Tine | Data from the entering classes of 2017 to 2022, excludin

The research conducted by the College Board itself indicates that ” the predictive ability (or r squared) of the SAT I is just .22, meaning the test explains only 22% of the variation in freshman grades. With a correlation of .54, high school grades alone do a better job, explaining almost 30% of the variance in first-year college performance.” https://fairtest.org/sat-i-faulty-instrument-predicting-college-success/ .

The recent studies of college success and the SAT are flawed in that they don’t consider the impact on:

1.  Students who waive the SAT’s to apply to certain schools that don’t require them
2. Students who drop out after the initial year in college, lacking disciple, despite their potential.
3. Students who do not apply to four-year schools after a mediocre performance on the SAT’s

In other words, the recent research evaluates only those students who perform fairly well on the SAT’s and have persistence and confidence to remain in college, despite the challenges of the first year, a turning point for most college dropouts.  Students who persist in college must have discipline, as well as high test scores.

Naturally poor, non-white, and test-averse students are well-presented in these absent populations in the recent studies heralded by the New York Times and other SAT advocates in recent weeks. The research showing the high predictive reliability of test scores  on student success over four years in college does not represent the students who are not good test-takers, but could succeed in college with the opportunity and good follow-through mentoring.

The legitimate problem of evaluating college applicants without test scores has been exacerbated by the Supreme Court’s ruling against affirmative action, a policy that gave wider options for students who do not test well.  What the studies do not show is how multiple indicators can be combined to get a broader picture of student potential. Colleges have developed methods of profiling students without tests, including high school grades (not valid by themselves), evaluation of grading within individual high schools to detect grade inflation, student essays, student performances in extra-curricula activities  (especially in art, music, artistic dance), and other exhibits students send to distinguish themselves. As indicators are combined, reliability is increased.

Another problem is the use of AI to write essays of application. Colleges have recently been challenged to evaluate the authenticity of  student writing at the college level. Probably the best current way to screen essays would be a signed statement, such as “This is my own work, unless otherwise indicated,” a statement college freshmen had to sign in first-year composition at my undergraduate school. Discovered writing fraud should disqualify the student.

While SAT results seem to give objective evidence of student potential, studies of the entire population of applicants would prove less reliability in their predictive validity of the SAT’s. Those who have succeeded on standardized tests are more confident in their predictive potential. After all it predicted their success. Those who have been less successful in the environment of timed and single-answer testing feel their potential will not be recognized in those tests.  Indeed they are less reliable than the rest of the application that represents student achievement and potential.