You forgot one of the most important:
- SAT/ACT scores have minimal if any statistical power in predicting college-going success.
(There are others, but that’s a really big one to have left off the list.)
7 Likes
That’s not a valid objection in the same way as the other two because many of the other admission critera aren’t justified on the basis of their predictive power. Some of us also disagree with these studies for their definitions of college “success” and their lack of rigor.
In terms of predictive power, test scores are better in aggregate, even if not individually in all cases. For instance, a group of students with 1500+ SAT scores will certainly outperform another group averaging less than 1000 SAT scores.
What measure of “success” do you prefer to use? Wouldn’t defining such a measure of “success” be necessary to determine what application aspects have more or less predictive power of “success”?
1 Like
I don’t think college “success” can be easily defined. “Success” may mean different things to different people. For some, graduation by itself may be a “success”. Many others would find such “success” leaves much to be desired. There’s no simple recipe for “success”.
1 Like
My kids do well on standardized test (both are National Merit Finalists) which is as much correlated with the ability to stay focused for 3 hrs and be alert and awake at 8 am as your knowledge of pre-calculus math and your reading comprehension. However, I support TO. Why would a kid who has taken 6 APs with ‘5’ by end of junior year, be required to jump yet another hurdle in taking the SAT? S/he has already proven that s/he is capable to do college-level learning. Kids that do not have these data points (AP results) may want to provide SAT results.
I think people may not understand (not the knowledgeable people on this forum, but in general) that ‘holistic’ evaluation at the top schools is also based on objective criteria and hard stats. They want the kid that qualified for the AIME or USAMO, the platinum medalist on USACO, the scientist that won those national science competitions, the pianist that was selected to perform in Carnegie Hall, the MVP soccer star. They want proof that your kid is better than the other kids based on an objective measure. They are not going to take a kid that is nice to the janitor or has written an amazing essay of self-discovery. What is unclear to me is how to compare the accomplishments of a prima ballerina to a winner of scholastic gold portfolio medal to somebody with published research on the same scale. You can’t. And that is why these acceptances are a lottery
7 Likes
So am I really understanding you correctly here? That you would like to see SAT/ACT scores continue to be used because they measurably correlate with something that is defined inconsistently and can’t be measured?
Because if that’s what you actually mean, I really do hope that you can see why some of us don’t find that terribly compelling. (And if it isn’t, please unpack your meaning a bit more.)
You’ve probably missed what I’d stated a number of times before, but I’ll summarize my positions here for you:
-
Test scores aren’t necessary for a few schools (e.g. MIT/Caltech) because they have other data to validate applicants’ academic abilities. Test scores are also not needed for colleges at the other end of the spectrum. However, for the majority of colleges, test scores serve a useful purpose as the only uniform and objective measure to assure that applicants meet some minimum standards. It’s a very good and helpful aggregate measure, regardless how well it correlates for any individual applicant (although it does pretty well in the majority of cases, IMO).
-
Test-optional policies may sound like a good compromise between those who have good scores to submit and those who don’t. Unfortunately, you can’t have a cake and eat it too. Because colleges don’t want to be accused of discriminating against non-submitters, the perception that tests don’t matter will take hold, leading to fewer and fewer applicants to submit scores. The economics of testing will then seal its fate, making the tests disappear altogether and depriving those colleges a tool they need.
2 Likes
Respectively:
- You seem to be assuming that test scores are useful for colleges. That has not been demonstrated, and the information that @Data10 has posted* very strongly suggests that your assumption is wrong.
- You seem to be assuming that test scores are useful for colleges. That has not been demonstrated, and the information that @Data10 has posted* very strongly suggests that your assumption is wrong.
Neat how it all boils down to the same thing, innit?
* Which, remember, isn’t limited to colleges at the “top” or “bottom” of the scale (however defined).
1 Like
That would be relevant if colleges admitted based on test scores alone, but that’ s not how admission works at the “top colleges” mentioned in the subject header and the original post. It’s not a question of whether the average 1500+ scorer will outperform the average 1000 SAT scorer. Its a question of how the kids that highly selective colleges admit under a test optional policy will do… kids that have top grades in a highly rigorous curriculum + glowing LORs + amazing essays + stellar ECs/awards related to planned field of study + … kids whose excel throughout their application, but their SAT/ACT may be lower than typical among applicants who excel in transcript, rigor, LORs, essays, ECs/awards… . Or more simply, does the test score significantly add beyond the many other criteria that is used to select applicants?
This isn’t a purely theoretical discussion. ~1000 colleges had gone test optional prior to COVID, including some highly selective ones, such as University of Chicago and Bowdoin. There is plenty of information about what happened to colleges upon going test optional and how successful test optional admits were in comparison to test submitter admits at those colleges. For example, the Bates 25 years of test optional study was posted earlier in thread. The test optional Bates students had nearly identical graduation rate and GPAs as test submitters Bates students. I know you don’t like colleges using these metrics and think they should instead focus on some other metric that you still have not defined, but the test optional policies seem to be successful in the criteria that is important to the colleges. Of the ~1000 colleges that have gone test optional prior to COVID, I’m not aware of any that decided this was a bad idea upon seeing the results and instead switched to being test required again.
Chicago is probably the closest pre-COVID test optional college to the ones mentioned in the original post. Chicago went test optional a few years before COVID. How has the school changed since then? We certainly don’t see Chicago watering down their classes or rigorous core curriculum. Many of the Chicago forum posters, including current students, seem to believe Chicago has a more rigorous core curriculum than any of the “top colleges” mentioned in the original post. We don’t see them becoming more formulaic in admissions. I am not aware of Chicago students failing by any metric more than previous years when they were test required. Instead it seems to be business as usual.
4 Likes
Let me try to summarize your positions that I know of on this issue:
-
All testings, not just standardized testing, are useless.
-
All grades, regardless how they’re earned, are the same.
Did I miss anything? We’re so far apart on this issue that we just have to agree to disagree.
3 Likes
I guess you haven’t read the posts in just the last couple of days on CC that some LoRs were provided by the students and their teachers just signed their names, or that a professional essay “reviewer” asks for $6k to “review” a single essay.
Chicago went test optional for the same reason that it did before to increase the number of applications. We don’t know how many students were accepted without test scores, what the rate of acceptance was for that TO group, what the profile of that group is, what the other characteristics of that group, and how that group did. We don’t have any of those information.
Even without those information, I bet most applicants were admitted with their test scores. Those admitted TO are the minority. It’d be a whole different world when no standardized test is available to anyone.
A little overstated. It would be more accurate to say that my positions are:
- Standardized testing does not add meaningful value to admissions decisions.
- Grades are generally (read: in the vast majority of cases) comparable between high schools.
Those are rather different, but I can see how you got what you did. On (1), I did say that exams have been found to be less useful as an assessment tool than other forms of assessment, though I never said they are useless; I did say that there’s a general move away from them in postsecondary education, but that’s a very different claim. On (2) I will admit that I didn’t exclude the extremes of the distribution of high schools in my initial phrasing, though in my defense I thought that would have been obvious given that we’re talking about statistical measures here, and weird things simply happen at the tails of a distribution curve.
So assuming that those were what you honestly thought my positions are on those issues and not simply taking an opportunity to paint me as more extremist than I am, I do hope that this has clarified things suitably. Do feel free to ask for any further explanation you might need.
So if someone makes a post on CC about hearing that someone else wrote their own LOR or hearing that someone else paid $6k for someone to review their essay, that means SAT/ACT should be required? If instead someone makes a post on CC about hearing that someone else got extra time on SAT/ACT without an accurate diagnosis, does that mean colleges should switch the interview from optional to required?
The relevant comparison is not whether there is an isolated post on CC or whether a particular measure is imperfect. The relevant comparison is instead does the test score significantly add beyond the combination of many other criteria that is used to select test optional admits at highly selective colleges?
We don’t have that information for Chicago but we do for other schools, and that information has been posted in this thread multiple times for multiple schools, including Bates which was mentioned elsewhere in the post you quoted. I see no point to repeat that information again, as you’ve obviously seen it.
Of course. Most applicants to non-LAC “top schools” are admitted with test scores even today, when a significant portion of applicants were unable to take the test. For example, Penn reported that 62% of ED applicants submitted scores, and 76% of ED admits submitted scores. It will probably be a lower percentage for RD, but I expect the majority of both Penn applicants and Penn admits will submit scores this year.
A large portion of applicants submitting scores at highly selective non-LAC colleges is expected, particularly ones that are newly test optional. The vast majority of applicants at highly selective private colleges have high test scores. And many of the new applicants who apply primarily because the test barrier has been removed do not excel in the many other non-score criteria that is considered, so their admit rate is often lower than test submitters. For Penn ED, I expect there is also a far lower portion of ALDC hooked kids among test optional applicants, particularly legacies, which influences the relative admit rate.
2 Likes
These are some CC posters’ first hand experiences. I think they have more credibilty than some of the same studies you often cite.
The point is that most other admissions criteria are not only subjective, but also subject to manipulations. Standardized testing is one of the very few remaining objective measures.
1 Like
In other words, anecdotes over data.
However, just because it is objective does not mean that it is of high value (predictive or otherwise). Indeed, different standardized tests may have different values. Unfortunately, it seems that the SAT and ACT are becoming more dominant within US standardized testing even though some universities have found them to be of lower predictive value of things like college GPA than other standardized tests like SAT subject tests.
3 Likes
Data can be manipulated, unfortunately. Someone’s first hand experience can’t.
Someone’s first hand experience can be an unusual outlier, rather than the norm.
4 Likes
It doesn’t have to be the norm. If it were the norm, we’d be in deep trouble.
Getting extra time is not a manipulation? What about taking an expensive prep class? Or attending a high school that teaches to the test in special classes? Or getting a copy of the test beforehand (In the past the College Board has scrapped SAT tests across entire countries due to cheating)?
I haven’t seen the thread , but the point is not whether they are first hand experiences or 2nd/3rd hand hearing about what someone else did. The point is that they are anecdotal, uncommon, and do not naturally lead to your conclusion. For example, suppose you personally knew Felcity Huffman and knew with 100% certainty that she paid someone $15k to take her daughter’s SAT/ACT and get her daughter a high score. Does that mean the SAT/ACT should be scrapped? Instead you need to consider things like how common does this occur, and how much the measure adds beyond the rest of the available criteria, in spite of its imperfections. Even if you conclude the measure is non-ideal, it doesn’t follow that a different and largely unrelated measure should be required.
3 Likes
I don’t know what your positions are on any of them. I assume you’re in favor of at least some of them. BTW, I’m against all of them.