- EphBlog - http://ephblog.com -

Does Amherst Lie About Its Admissions Data?

Williams and Swarthmore (and most other liberal arts college) have a meaningful number of students with sub-600 SAT scores. For Williams:


Amherst (pdf) does not.


I think that Amherst is “lying” about its admissions data. Recall this discussion and this one. There is simply no way for an elite liberal arts college to have competitive sports teams (especially in male helmet sports) and meaningful racial diversity (especially African-American) without around 5% of the student body having either math or verbal SAT scores (or both) below 600.

The trick:

Amherst consistently reports SAT plus ACT totalling 100% of their enrollees (or the 99% that results from adding rounded numbers).

Both Swarthmore and Williams are consistently reporting a total SAT plus ACT in the 110% to 120% range.

Clearly what is going on here is that Swarthmore and Williams are reporting both the SAT and ACT for students submitting both (as they are supposed to according to the instructions).

Amherst is not. Amherst is pulling a “Middlebury” and only reporting whichever score (SAT or ACT) they used for admissions purposes, presumably whichever is higher. (I know that they receive both scores for the dual test takers). It is incomprehsible that not one single enrolled Amherst freshman took both the SAT and the ACT when 20% of both Williams and Swarthmore’s freshman classes the last two years took both.

This a 2008 comment was from HWC, whose contributions I still miss. Looks like Amherst is still cheating. In the latest Common Data Sets, we see for Williams:


For Amherst:


Do you see the trick? About the same percentage of students at Williams and Amherst report ACT scores. That makes sense! Williams and Amherst draw their students from the same populations. But Amherst claims that only 52% (instead of 68% at Williams) report SAT scores. That is the lie. Amherst almost surely gets SAT scores from about the same percentage of its students. It just chooses to ignore those scores from those students whose SAT scores are worse than their ACT scores, pretending that it did not “use” those scores in making its admissions decisions. If that is what they are doing (and it almost certainly is), then Amherst is guilty of fraud. How else to explain their divergence from places like Swarthmore:


And Pomona:


And Wellesley:


There is a great story here for the Record, or for The Amherst Student . . .

Perhaps our friends at Dartblog can help us out. For Dartmouth:


I think that Dartmouth is “pulling an Amherst.” Way more than 51% of the first year students enrolled at Dartmouth took the SAT and reported their scores to Dartmouth when they applied. Dartmouth just “forgets” the scores for those with better ACT than SAT scores when it reports its data. How else could the SAT percentage be 51% at Dartmouth but 85% (!) at Harvard, 74% at Yale, and 67% at Brown. (I think that SAT percentages at Harvard/Yale are inflated due to their prestige. Brown’s percentage is in-line with elite liberal arts colleges. Is there an innocent explanation for Dartmouth’s low percentage? I doubt it.)

Comments Disabled (Open | Close)

Comments Disabled To "Does Amherst Lie About Its Admissions Data?"

#1 Comment By anonymous On May 4, 2017 @ 9:31 am

Perhaps these “fact-based” institutions are using “alternative facts”…

#2 Comment By KSM On May 4, 2017 @ 9:32 am

While I agree with you that the cause of this discrepancy most likely can be traced back to a policy at Amherst of suppressing the weaker score in cases where a student has submitted both SAT and ACT scores, I don’t agree that this is tantamount to “fraud,” especially in the loosey-goosey world of college admissions. To paraphrase Bill Clinton, it all depends on what the definition of “submit” is. I believe Amherst’s position would be that they automatically discard the lower of the two scores (per the concordance tables) on behalf of applicants, so the weaker score was never actually “submitted.” Other schools would be free to do the same, and probably should.

There are other ways that schools are self-serving in their interpretation of the CDS instructions. For example, if a student submits SAT scores from three separate dates, does the school include all three scores? No, they superscore it, and just report the best subtest scores across all three dates, while suppressing all the lower scores. I believe ALL schools do that. And then there is the way ACT scores are reported. Some schools superscore, others do not.

#3 Comment By hmm On May 10, 2017 @ 4:40 pm

Might be interesting to note that Amherst reported the highest test score ranges of its peer schools this year.

The terminology in the CDS is a bit vague: “Include information for ALL enrolled, degree-seeking, first-time, first-year (freshman) students who submitted test scores.” If Amherst did what KSM is suggesting, it’s not really doing anything wrong, as every enrolled student is accounted for in some fashion. The 3% overlap may be with students whose SAT and ACT are identical. But it is a deliberate choice they’re making that none of their peers are.

Dartmouth is rather strange. They require either the SAT or ACT as far as I can tell. Why don’t their totals add up to 100%?

#4 Comment By hmm On May 10, 2017 @ 4:48 pm

Actually, I gave their admissions profile a glance and nothing seems off: https://www.amherst.edu/media/view/669797

The amount of students Amherst admits with scores below 600 is in the handful. Even if Amherst is hiding low scorers, do the math for accepted/applied for each category and you’ll get Amherst’s admit rate: ~16%.

Amherst gets so many applicants with high test scores that they can easily have a class with 0% scoring below a 700/32 if they really wanted. They may be unlike Williams, Swarthmore, and Pomona in placing a hard line on those scoring 600 in a section unless they’re a famous legacy or a once-in-a-lifetime candidate. Only around ~10 students are getting in at Amherst with those scores in any section of their testing.

#5 Comment By hmm On May 10, 2017 @ 4:52 pm

Aha! Found the discrepancy.

4035 with SAT
3565 with ACT

7600 scores total

8406 applicants

That’s nearly 1000 applicants unaccounted for! Something is definitely fishy.