New pet peeve: test optional at top schools

Parents Forum
907

There are 2 factors at play here:

  • the emphasis on AP (or IB) means students followed a nationally defined curriculum with similar content, types of questions, and grzding rubrics. While one can nitpick that teachers don’t all grade the same, getting an A in an AP class is relatively comparable from school to school and adcoms know what to expect(unlike, say, the course content covered in “regular” physics, in “honors anatomy &physiology”, or “English 11”).

  • Generally speaking, colleges also look for personal attributes. Students who took challenging classes and got As show these attributes regardless of school. A math student who is only one among 8 to reach precalculus at her rural school with most students stopping at Algebra 2, will not have covered content a student in calc BC covered, but has displayed math ability within the context of her school. Now, if that Calc BC student is 2 years advanced in a school where 10% take MVC senior year, does that show more or less aptitude than the student whose school doesn’t offer calculus at all?
    To make things even more complicated: Keep in mind even elite schools will have 3, 4, or 5 levels of math that freshmen can take (including "math for citizenship " classes that suppose neither calc nor precalc.).

So, while some high schools may be more rigorous than others, the AP curriculum shows rigor regardless of HS per se, and student aptitudes can be derived from more than 12 APs v. 6 APs.
(That last point seems very confusing to many HS posters.)

3 Likes
908

I look to look at the actual study for such articles. The referenced study for college admission is at https://www.schoolcounselor.org/getmedia/11e887c3-bf04-4345-90c0-432f7dd8d69a/Gender-Ethnic-Bias.pdf .

The study indicates the following analyzed factors were significantly correlated with college admission, listed from highest correlation to lowest correlation in isolation. A positive correlation indicates increased chance of admission (in isolation). A negative correlation indicates decreased chance of admission (in isolation).

Highest Correlation with College Admission

  1. Difficulty of HS Program: 0.45
  2. HS Grades: 0.44
  3. SAT Score: 0.37
  4. Word Count in LOR: 0.18
  5. % Achievement Words in LOR: 0.12
  6. % Standout Adjectives in LOR: 0.03 (not statistically significant)
  7. % Agenic Adjectives in LOR: -0.01 (not statistically significant)
  8. % Ability Adjectives in LOR: -0.03
  9. % Grindstone Adjectives in LOR: -0.08
  10. % Communal Adjectives in LOR: -0.09

It suggests that LORs may have increased correlation with admission if they have a high word count and many achievement words… but few communal, grindstone, or ability adjectives. The combined multiple variable regression model that controlled for stats found that only word count had a slight positive coefficient with admission. All the adjectives in the LOR had a slight negative coefficient with admission, when controlling for stats. The combined regression model that controlled for GPA, difficulty of HS program, and scores found that ethnicity of applicant had no correlation with the word count (0.00 coefficient). Being female had a very slight positive coefficient of 0.01. More significant was the 0.04 (0.02) coefficient for grindstone words, which was negatively correlated with admission. Recommenders were slightly more likely to use grindstone words when describing female applicants than male applicants, after controlling for stats. Some specific numbers are below for correlation in isolation (not controlled with similar stats).

Correlation with Word Count

  1. URM Applicant: -0.07
  2. Male Applicant. -0.03
  3. Male Recommender: -0.03

Correlation with Grindstone Words

  1. URM Applicant: -0.01
  2. Male Applicant: -0.09
  3. Male Recommender: -0.06

I also included HS grades and SAT scores for a reference comparison.

Correlation with HS Grades

  1. URM Applicant: -0.22
  2. Male Applicant: -0.04
  3. Male Recommender: -0.08

Correlation with SAT Score

  1. URM Applicant: -0.33
  2. Male Applicant: +0.02
  3. Male Recommender: -0.13

In summary it appears that reviewers seem to write slightly more about female applicants than male applicants, which was positively associated with admission. However, recommenders also used slightly more grindstone words (“tireless”, “committed”, …) when describing females, which was negatively associated with admission. With the exception of grindstone words, the degree of difference between genders on LORs was similar to the degree of difference between genders on GPA and scores in this sample.

It don’t think this is good evidence of that we should get rid of LORs, nor does the author of the study. The author does recommend that counselors and teachers be aware of possible biases when writing LORs. However, I do think t that LORs should not be used in isolation, and instead they should serve as an addition to other components of the application.

3 Likes
909

As someone who historically read LoRs for grad school admissions, I will say that the only thing we looked for were things we took as warnings—yellow or red flags, basically. We pretty much ignored what we saw as the inevitable hyperbole of the American LoR genre.

Don’t know if undergrad admissions works that way, too, but that could explain the low correlations with things like “achievement words”.

1 Like
910

In theory… but there are some high schools or districts where the most common AP score for students who earned A grades in the AP course is 1. Obviously, students in those high schools are being short-changed on the content of the AP course (and perhaps the prerequisites to that AP course).

3 Likes
911

Neither college GPA nor graduation rate are good measures of college “successes” or good outcomes. These studies aren’t really worth much IMO. College courses vary greatly in rigor and grades are not only inflated accross the board but they’re much more so in certain types of courses, especially the ones lacking in rigor. Most colleges these days have plenty of such courses to “help” their students achieve respectable GPA and graduate.

912

I’m not sure what colleges you’ve worked for and the field you’re in (behavioral science if I recall?) but LoRs by respected professors in the same field are of utmost importance in admission to a PhD program.

2 Likes
913

What would you consider to be a good measure?

1 Like
914

AP courses are perfect examples to counter claims that grades are sufficient. None of colleges gives credit based on grades. They all use AP exam scores to validate the grades. Some people on this thread still claim (one did just a few posts ago) that an A is an A, regardless where it came from.

1 Like
915

If it were easy to define, someone would have done it.

916

When it comes to college admissions, yes an A is an A: scores don’t count (much if at all) for admission, what matters is getting an A in the AP class. The assumption is that if the student took the most rigorous curriculum offered and got an A, they cannot be penalized for their school system’s failings.
(In Mississippi, taking AP’s is uncommon and 1 or 2 is the most common AP score, so students from Mississippi wouldn’t be penalized for their state’s poor public education system. Top students from Mississippi remain in-demand, because there aren’t that many of them.)

BTW AP exams are exams, not standardized tests. They require knowledge of specific content and skill with various types of exercises.
Their interest from a college adcom standpoint is that they offer a similar curriculum to a group of (self identified or faculty selected) good students, so that regardless of teaching quality an A will indicate success in a “known quantity”.
Credit is subsidiary to US admissions and since at top colleges AP’s are the “default” assumed preparation for freshman courses, they do not carry any credit.

917

AP exams are standardized, like final exams on specified content (versus SAT and ACT that are kind of a mix of content and some attempt at finding “aptitude”, not really optimal for either purpose). I.e. they resemble the SAT and ACT in the standardization aspect, but they differ in the goal of the test in terms of what is tested.

Besides Caltech and Harvey Mudd, what “top” colleges give nothing (credit units, subject credit, or advanced placement) for any AP scores?

1 Like
918

Back when we were doing the admissions rounds, more then one adcom told us they look for As because they knew students who got As would do what it took to get them. The bar may be different at various high schools, but those student had the trait they were looking for.

Regarding APs they said they didn’t need to see tons of scores. Students who would do well would continue to do well. Three APs were as good to them as nine if the student opted to take other things to broaden their education instead (aka non-AP subjects).

At our school lower AP scores are quite common. It’s not the A student in the course getting the 1 though. It’s more common for them to get a 3. Those who can get 5s often have very nice choices of colleges to select from.

2 Likes
919

A chunk of what I had typed didn’t get posted (my computer turned black right when I hit submit). I mentioned scores being used for placement, for instance. But that’s only after admission - I was pointing out how, for admission, adcoms want to see the A, the score itself is tied to credit which may or may not be claimed, may or may not result in credit or placement.

The point, though, was that credit is subsidiary for ADMISSIONS.
(For US Adcoms, taking APs and getting an A is what they want to see. Ap scores aren’t required for admission and aren’t required to “validate” the admission, ie., if you don’t submit them you’re not rescinded. UK “courses” have a different view).

AP exams aren’t exactly standardized in the way the ACT and the SAT are (e;g.,they deliberately require to be prepared before they take them and that preparation is a year long class v. ACT/SAT which aren’t supposed to be prepared but are nevertheless) but that’d be off topic I think.

920

Pretty sure that everyone has given up on the idea that SAT and ACT are not supposed to be prepared for. The College Board has some basic SAT practice stuff and links from its web site: SAT Practice and Preparation – SAT Suite | College Board

The difference is not the standardization, but the target of the tests. AP tests (and the discontinued SAT subject tests) are supposed to be like standardized final exams for course material learned in courses taken in high school. SAT and ACT are supposed to be “general” tests.

1 Like
921

Please re-read what I actually wrote. I wrote nothing that goes against what you replied with.

The existence of LoRs is important. The content? Not nearly as much, absent (subtly phrased) warnings. (That’s assuming it’s from an American or Canadian professor, of course. Letters from British faculty, those are a different genre entirely.)

And my field is linguistics, though I also have had experience in this regard in rhetoric/composition-oriented programs.

922

The reason AP exams aren’t used in US college admissions (unlike in UK) isn’t because they aren’t useful. According to Fitzsimmons of Harvard, APs are the best indicators of students “successes” at Harvard. They aren’t used in admissions because AP courses aren’t offered uniformly and qualities of such courses vary significantly even when they’re offered. They are viewed in the context of the high school, which means any As are relative, not absolute, when not accompanied by the corresponding AP scores.

923

I don’t think preparation is what distinguishes AP exams from SAT/ACT. They’re all prepable to some extent, but preparation alone generally doesn’t get you a perfect score.

1 Like
924

So tests should be required because test optional admits will be far inferior to test submitter admits in a non-GPA/graduation rate measure of “successes” that we cannot define, but know it must be there? . I suspect you are overrestimating how much SAT/ACT score adds to predicting this undefined measure beyond the many other criteria that is available within the rest of the application at “top colleges.”

1 Like
925

I don’t think current versions of SAT/ACT are useful for every student or every college. However, for a vast majority of colleges and students, they serve a purpose. They help “standardize” the vastly uneven grading at high schools across the country. There is nothing better that anyone has discovered yet that can serve that purpose as well.

926

If trying to match grading to standardized achievement measures, then the discontinued SAT subject tests (for regular high school level courses) and AP tests (for advanced level courses) would be more useful than the SAT and ACT.

1 Like