« Keeping the Concord Review Afloat | Main | Weighing Reading First's Impact »

SAT Scores and GPA Predictability


A major restructuring of the SAT expanded the test to four hours, in part because of the new writing section, which was intended to paint a truer picture of students' readiness for college work.

A new study released by the College Board today has found that the revised SAT is a little more effective at predicting how well students will do in their college courses than the previous version, and that the writing-section results are the best predictor of later college performance. But the best alternative to a crystal ball—no test-prep needed to figure this one out—is a combination of SAT scores and high school grades.

So, just to make sure you were paying attention: A high SAT score and good high school grades are to a respectable college GPA as ??? (Oh, right. I was never very good at analogies. How about you?)



If this is an important subject, shouldn't you be reporting about what a poor correlation between SAT scores + High school grades have had with 1st year GPA.

An "adjusted" R value of 0.61 (for the combination) is not a good correlation.

Shouldn't the poor correlation and the lack of predictablitity of the SAT + High school grades mean more than the very small improvements made by the new SAT?

With the number of influences that first year college students typically experience, both positive and negative, it may be difficult to predict first year performance. Could the SAT and high school grades predict long term performance?

Comments are now closed for this post.


Most Viewed on Education Week



Recent Comments