@evergreen5 . You’re right about PrepScholar apparently using an old CollegeBoard conversion table. It matches exactly. The true correspondence will have to come from the score distribution of students taking the test.
Some more confusing data from Amherst this time
According to the Amherst web site -here are the stats for the freshmen class
Their SAT composite (the “old test”) is 2232—also a record. And their redesigned SAT (first introduced with this class) is 1469.
But according to the College Board concordance table an old SAT score of 2230 is equal to a new SAT score of 1530.
So the avg for the new score seems very low if the concordance tables are accurate
From what I have seen so far all the colleges reported new SATs somehow lower than the expected ones (based on the concordance tables and especially on the upper end). There are older threads that discuss why that have happened but there is no agreement. I would assume it will take some years for things to clear up.
@wisteria100 As the first class submitting the new test, I think it’s possible some students submitted the old version score instead. So one doesn’t necessarily equal the other - if a kid took the new one at all, and didn’t like the score, they’d send the higher old version score, no?
The new SAT [given in 2016] was harder at the tippity top end of the scale (small mistakes hit hard on points). The new SAT is close or slightly easier for kids in the middle range of scores and easier on the low end.
I think that they scaled the first concordance tables to the low/med range to hid the disparity. It may be that the have scaled the SAT this year (after seeing all the results) to be more consistent - and that may be another reason for new concordance - the new concordance is for the newly scaled tests given this year vs last year (Class of 2017).
That could be true that the later new SAT test are more difficult than the first few. I have seen a big drop in the mid 50 of new SAT this year versus last year at least in one school. Nevertheless, the sample size with new SAT score was very small last year that the mid 50 may be not an accurate reflection of the class.
I’ve been reading this thread and I think it’s super interesting but…can someone tell me why any of this matters anymore? Aren’t most kids who are applying with new SATs now? And won’t most schools have updated info on their new admits from last year using the new SAT?
It matters because we still using a concordance from CB that isn’t based on actual test results. D18 is applying to schools who still haven’t posted a common data set for class of 2017 and it’s hard to tell whether to send her ACT or SAT scores. Because it was the first year, the CDSs may be skewed anyway - many students avoided the SAT altogether.
To me it’s interesting because of the increased importance of early decision applications and the downright dishonesty being displayed by so many schools. High school seniors have to decide where to apply early. Yet, the information provided makes it excruciatingly difficult to decipher. More often than not, schools are providing predictions based on the concordance tables even after they have the actual data. There is no question the new SAT scores are much lower than predicted, at least at the highest scores. One school that is not conflating enrolled and admitted and new and concorded scores is Princeton. Its scores, which are instructive, are as follows:
Princeton
Statistics for Enrolled Class of 2021
(as of July 15, 2017)
Class Overview:
50.4 percent of students in the class are female and 49.6 percent are male.
13 percent of students in the class are children of Princeton alumni.
Middle 50 percent SAT Scores (the new SAT):
Math: 700-780
Evidenced-based Reading and Writing: 680-760
Middle 50 percent ACT Scores:
Composite Score: 31-35
https://admission.princeton.edu/how-apply/admission-statistics
That is dramatically lower than the concordance predictions.
It also matters for students building their college lists, and for those trying to decide whether to apply Early Action or Early Decision, and if so where. To estimate one’s chances of admission, it helps to know where your SAT scores lie relative to each school’s typical range (and expectations, since admissions officers are also biased by the concordance table). It’s not clear if colleges will set their expectations using last year’s batch of new SAT scores (which suggests a lower range), or by using the CB concordance table to translate their old SAT scores to new SAT scores (which suggests a higher range). Depending on what these expectations are and what the actual SAT scores will be this year, students might over- or underestimate their chances. Even a percentile table of new SAT scores actually achieved to date would be more useful than the existing concordance table, which was largely based on a single test administration in May 2016. As we get more data, this problem will eventually resolve itself, but it’s causing a problem for students in this year’s application pool.
The official conversion app/site at CB is still not updated yet. Interesting, I entered a new SAT score to get the old SAT score, and then I entered the calculated old SAT score to convert back to new SAT score. It went down by 30 points.
So…is everyone finding that, when schools DO update their SAT data on the CDS, those SAT scores are lower than what would expected using the CB’s concordance tables?
@homerdog The CDS may be a problem for this year’s seniors if it only reports one set of SAT scores, with Old concorded to New via CB’s tables or vice versa. Unless the CDS contains only New scores and no concorded Old, I think it’s going to be hard to interpret.
Individual school website reports that include middle 50 ranges for both New and Old scores separately are very helpful. However, there are arguments in this long thread that students submitting new vs old scores were different, with the better students supposedly taking the Old SAT or ACT. I’m not sure I buy that, but nonetheless, that would muddy the data even more.
Really glad I don’t have a current senior…
Another area where the accuracy of the College Board’s concordance tables comes into play is with automatic scholarships. Many state schools that offer automatic merit are still using the concordance table scores (specifically, table 7) to set SAT score levels that correspond to ACT.
Similarly, a number of schools automatically admit into honors programs and a few, such as Pitt, automatically admit into med, law and MBA programs based on SAT. Last year’s automatic for honors is 1450. That is starting to look like a high number on the new SAT.
There are other possible explanations of why SAT scores for admitted students are lower than expected based on concordance.
For example:
- Universities are passing over students with higher SAT scores for students with lower SAT scores.
OR - The students applying with the new SAT are weaker than the students who used to apply with the old SAT. Maybe the whole student population is getting weaker over time. Maybe some of the stronger students are applying with the ACT so there are not as many very high SAT scores.
Traditionally, concordance was established by administering both tests to the same group of students. This statistical sample allows us to say that on average, a student who scores X on test A scores Y on Test B.This is how the old SAT - ACT concordance was done. Is there evidence that the CB concordance tables fail this test?
Of course, for a student applying to college, it does not matter whether the concordance tables are correct. What matters are the average SAT scores of the admitted students. So it is better to focus on the SAT scores of admitted students than on the concordance tables, assuming the data is available.
Just catching up on this thread. All I can say is that it is such a shame that College Board didn’t have the test roll-out better planned. I have to believe that colleges are aware the new SAT is not an accurate representation of a student’s scoring ability. Very upsetting for the senior class of 2017-2018.
The Vandy ED scores don’t seem to have changed enough between the old SAT and the new SAT.
“Very upsetting for the senior class of 2017-2018.”
It will cause more kids to focus more of their time on the ACT and to either send the ACT to schools or send both tests and let the schools decide how best to decipher the scores.
“The Vandy ED scores don’t seem to have changed enough between the old SAT and the new SAT.”
Vandy is more focused on test scores compared to grades compared to some of its peer schools. In addition, Vandy, given its location, is predominantly an ACT school. Last year, the percentage of kids submitting SAT scores declined from 41% to 37.6%, so perhaps the kids who did not do as well on the new SAT either did not submit them or were rejected.
There were no kids more adversely affected by the SAT fiasco than the class of 2017. The class of 2018 will have he benefit of colleges already having their own true basis of comparison based on actual scores they received of new SAT Class of 2017.