Nobody said athletes were not qualified to succeed at Harvard. Athletes have a very high graduation rate – similar to non-athletes. The claim I disagreed with and quoted in the original post was that score/grade standards were not lower for typical athletes.
3 Likes
Funny, but it would be a modern medical miracle for Wilt to be walking around Disneyland. 
(I know you were just using him for your analogy)
I tried to look this up. There is not a lot of robust, publicly available data concerning the standardized test results of recruited Ivy League athletes.
-
According to USNWR, the reported Harvard SAT mean score for matriculating students in 2021 was 1520. 25th percentile was 1460. 75th percentile was 1580.
-
According to the Harvard Crimson in 2021 in their annual survey of the freshman class, they reported the following results: average overall SAT score was 1494, average SAT score for non-athletes was 1501, average SAT score for athletes was 1397. This data suffers from the biases inherent in any convenience sampling instrument. However, If 1397 represents the real average of recruited athletes, it is significantly below the 25th percentile score of 1460 and probably more than 1 standard deviation below the mean.
If applicable, this doesn’t mean that the recruited athletes attending Harvard aren’t qualified or don’t deserve to attend the school. It only means that their standardized test results are lower than the rest of the the school. My guess is that the disparity in scores has become more apparent because the 25th percentile score for Harvard has steadily increased during the past decade from 1390 to 1460, while the scores of its recruited athletes probably haven’t increased as much. Interestingly, during the same 10 year period, the 75th percentile score at Harvard has remained relatively constant from 1570-1600.
https://www.usnews.com/best-colleges/harvard-university-2155/applying
1 Like
Ignoring test optional limitations, 1 SD below the mean corresponds to ~16th percentile in a normal distribution, so 1397 seems reasonable for -1SD, if 1460 is 25th percentile.
1 Like
Exactly. Not to mention we have no idea what proportion of athletes responded to The Crimson’s survey, or what proportion applied with test scores, or of the ones who participated in the survey, which respondents gave accurate answers. Doesn’t make sense to even give these ‘findings’ the time of day.
Lastly, a point that continues to be ignored is that many, maybe most, athletic recruits take the ACT/SAT in freshman/soph years of HS, while non-athletes often don’t start testing until junior year. Just that timing difference of 1-2 years would lead to differences in test scores by group, so you are comparing apples and oranges…in a metric that tells us nothing about one’s intelligence, or future success potential.
2 Likes
This.
2 Likes
My brother, an Ivy & Stanford grad, ceased recruiting for his company at the Ivies. He says the students are exceptional test takers, but they don’t present as a good fit for his team based environment.
The calculated standard deviation is 89. So 1 standard deviation below the mean (1520) would be 1431. 1397 is easily lower than 1 standard deviation below the mean. However, Harvard’s SAT results may not follow a normal distribution. 75th percentile is already 1580 and 1600 is the maximum value.
Well, USNWR is wrong. It’s probably close, but still wrong. Harvard does not report percentiles on a 1600 scale; they report percentiles for EBRW and for math. It’s such a rookie mistake to simply add together.
2 Likes
You are correct. Many recruited athletes take their tests early. For our sport, sophomore and junior year is more accurate. However, there are many (at least 50%) who don’t take the test at all until after they are committed. Some take the test multiple times in order to ensure they meet the school’s minimum. We knew of 1 family that planned to take both the SAT and ACT every month it was offered (after they committed) until the application deadline. They don’t try to score any higher than what is required. There is no reason to.
I posted this information out of curiousity as well as historical interest. The Ivy League became test optional for recruited athletes for 2021 and 2022. They haven’t announced their official decision for 2023, but it seems likely they will remain test optional for the foreseeable future. However, the decision to become test optional occurred on 8/13/2020. By that time, I think most recruited athletes for the HS 2021 class had already taken a standardized test because the EA applications were due on November 1. The 2021 recruits that we knew had all taken at least 1 test a long time before the announcement. As a result, I don’t think we will see future information on recruited athlete’s test scores for a long time, if ever. Remaining test optional will effectively remove this unwanted source of potential differentiation between recruited athletes and regular applicants.
1 Like
Yes, you’re right. I prefaced my remarks with the statement that there isn’t robust, publicly available data. I tried to work with what I could find. USNWR even if inaccurate is the public’s go to source for college admission information and their information is widely copied by other sites. What can you do?
I would also add that AI averages by team can be very different as long as the AI for all recruited athletes in a year are within 1 standard deviation of the campus mean. Football for instance is allowed 30 recruits, of which 20+ can be greater than 1 standard deviation from the mean. It’s also generally assumed that some other sports are also given lower team AI targets, like men’s basketball, ice hockey and lacrosse. Allowing certain teams to have lower AI averages (below 1 standard deviation) means other sports have to make up for this deficiency. However, based on the Harvard litigation data, it seems quite evident that as an overall body, Harvard makes full use of the allowed lower standards for recruited athletes.
1 Like
Not build statistical arguments on bad data, for starters. Then one is exacerbating the issue with GIGO.
Yes agreed, but we have no evidence that AI is still be utilized in the new test optional world. There are still coaches who do require a score from recruits, but that’s inconsistent in a given school/conference. IME of course.
That’s a pretty rough comment. This is a forum, not a scientific publication. Even then, many research articles published in peer-reviewed journals contain methodological and scientific flaws because the world isn’t perfect and some flawed studies are still worth publishing because the issues are noteworthy. I did acknowledge the limitations in advance. Shouldn’t your criticism be aimed more at the Harvard Crimson and USNWR for publishing their material?
1 Like
Just curious. Do you know if coaches are asking all of their recruits to submit scores or only a subset where they a think a good score will help the recruit get through the pre-read?
You asked a question. I gave an answer. Had USNWR asked the question, I would have given the same answer.
I will also point out that my use of an impersonal pronoun is so you don’t erroneously think I’m criticizing you
That’s a straw man. The issue is hardly noteworthy. Having been debated on dozens of threads, it’s not even novel. But I’ll move on for the moment.
The NCAA said it would require test scores for eligibility in 2023. That could change, of course, and Harvard wouldn’t have to consider those scores for admission or report them, but if the NCAA requires them for eligibility to play, the Ivy league would also have to require them.
Practices vary and continue to develop. Some coaches require scores (talking about TO schools), some don’t require them, some do tell students they would be a stronger candidate with X score. I have some senior recruits still testing to qualify for larger merit awards (obvs that’s not at the Ivies or other non-merit providers).
1 Like
That’s interesting, I didn’t know that. I feel that there is some hypocrisy expecting a recruited athlete to take a standardized test if the rest of the college applicants are not required to do so.
1 Like