You also have to assume, though, that a lot of those that did not submit had another significant hook. Because of this, we really relied more on scoir (which shows all scores at our school, is my DD's actual pool, and makes the outliers really clear). We also decided that using the 25th-75th% as a surrogate for 50th% and up for test required schools was a good enough rough estimate.

Look at the school's 2020-21 Common Data Set, the last admission season prior to test optional policies.

I had ChatGPT do it and thought the results were reasonable, as it talks through the assumptions it makes. And then when some colleges went back to test required, the scores dropped more or less as ChatGPT predicted.

You can roughly do it on your own. If you have the average SAT score among those who submitted (not sure if the CDS gives this), then you have to assume a score for those who didn't submit. I'm guessing somewhere south of the reported 25th percentile, and perhaps well south of it. But then calculate the weighted average of that score and the average score using the percent who submitted as weights.

Anonymous wrote:I think a lot of people just pull the last test-mandatory CDS and use that.

This is the answer. CDS gives quartile numbers and percentage who submit scores. Won’t get an exact answer but you can get a good ball park estimate. Those the schools want to shape a class, athletes, URMs, etc., won’t have to submit scores to be considered. Basically good test scores validate high GPAs or in rare cases can add clarity to a poor but improving GPA.

I think a lot of people just pull the last test-mandatory CDS and use that.

and dang it, I meant Test OPTIONAL

I'm looking at schools trying to figure out if DS's stats are at median or low or what given the option not to offer test information. I see the percentages that submit and imagine it's a formula like "if 1/3 submit, that's actually the top 1/3 of the class, so assume a slope between 75% and 25% and slide the values further down the slope to get the real quartiles." (I know that's a little vague.)

Does anyone have any reasonable insight here?