re: #325 Talking Point Detective

Well - for one, grades correlate better. And yes, combining measures gives you better results. But the entire notion of trying to find a short-hand - which schools need to do because of numbers - for the complicated and detailed processes it would take to really evaluate applicants comprehensively, in the end reflects biases and prejudices within the system and perpetuates inequities that exist in the status quo.

If you really want to understand how a student will do in college, you need to understand at a deeper level how they will do when faced with adversity, how much they are really interested in learning, how they feel about themselves as learners, what motivates them - and not use some statistical measure that is mistakenly viewed as “objective.” Why do you think such a huge percentage of students drop out of school in their first year?

No, grades do not correlate better — at least, not in the studies I’ve been reading. Instead they’re about as well-correlated as the SAT. Worse, consider the grade inflation that occurs when the grades are made. (see Georgia and grade inflation in high schools in response to HOPE scholarship adjustments.)

As to the rest of your points - sure. I agree. But how do you do this in any effective fashion? Interviews? Portfolios? The former are excellent provided the interviewers have time and training. Time is already in short supply, while training is also going to have to overcome the issues of equitability and bias. Portfolios? You’re returning to whether the high schools adequately train and prepare students to prepare these. They’re inevitably going to encounter the ‘teach to the test’ problem, and preliminary evaluations are that they’re going to be no more valid or reliable than the SAT or GPA.

Look. Properly, the SAT and the GPA are screens, not predictors. Values below certain thresholds are highly unlikely to succeed in college. Values above are not guarantees, but by using them the colleges can (if they wish) use the more time-intensive methods such as interviews on what passed the first screen.

Unfortunately, colleges are businesses — even state colleges, sadly. Regardless of likelihood of success, each body is more money in the coffer. As long as that fact applies, anything that does better than random is good enough.