High school grades, not test scores are best predictors of college success
A study in 2001, published by the Brookings Institution, has found that 50-80% of test score improvements were temporary and were "caused by fluctuations that had nothing to do with long-term changes in learning..." meaning that the standardized tests were unreliable in measuring a student’s performance.
William C. Hiss, the former dean of Admissions from Bates College, has also stated that their study has found that there was not any significant difference between those who has submitted in their standardized test scores and those who didn’t.
So if a student’s performance on a standardized test just have temporary improvements