In the discussion of gender and MIT grad rates, part of the puzzle is gender distribution in majors. The another part is graduation rates and GPA per major. I found the first part, but had no luck with the second: http://tech.mit.edu/V128/N47/women/tables.pdf
@theloniusmonk
GRIT was selected by educators for lack of a better word. I don’t know that they met to imply a military boot camp or navy seals swimming in ice cold water. A better word in the case of my referenced work would be “error term” or “unexplained variable.” The point being that the unexplained portion of the variance analysis started at 50% in the first year and increased to 75% in the second year. Buy the third year the variance in test scores showed no significant relationship with the junior year college classroom GPA.
:-t
@StevenToCollege
If the predictive ability of standardized test scores are NOT explaining actual variances in GPA performance in college, they are not, for college selection purposes, “objective.” Differentiating selection of one student over another based on an 80 point lower score is bad science because the evidence of scientifically significant results is not there. Students, teachers and guidance counselors understand that secondary school GPAs have a subjective element to them. It is worth noting that, as poor as they may be, they actually do a better job of predicting college classroom success than do the standardized test scores.
If you have 100 randomly selected math SAT scores at 550 and another 100 randomly selected test scores at 750, there is a significant difference in their performance in calculus class. The range of standardized test scores applying to these highly selective schools are grouped only at the higher end.
“If the predictive ability of standardized test scores are NOT explaining actual variances in GPA performance in college, they are not, for college selection purposes, objective.”
Well, the standardized tests do not need to explain the variances in GPA performance in college to be objective. They are objective as long as they are uniform.
“Differentiating selection of one student over another based on an 80 point lower score is bad science”
80 (1560 vs 1480) points difference on the new SAT, that translates to 14-15 questions difference. It’s actually a lot considering it’s about 10% difference on the scale of 100.
As an example, looking at the approximate November curve, while there are many possibilities, the difference could be 8 questions, 4 for math and 2 each for reading and writing:
Student A, 1560 (800 math -0, 760 EBRW -5 R 37, -1 W 39) -6
Student B, 1480 (760 math -4, 720 EBRW -7 R 35, -3 W 37) -14
Note that the 99th percentile begins at 1480.
Perhaps, as David Coleman might suggest, Student A practiced on Khan Academy and got himself a 100+ pt improvement, while Student B did not.
Looked through the OSU study and collected some points on gender distribution per major in the college of engineering. Skipping over the “pre-majors” and only looking at students who declared their major, we get the following:
Total declared: 4519 students, 846 female students, 3673 male students in college of engineering
51.1% of male students are declared in ECE, CSE, MSE, or Civil E.
32.9% of female students are declared in ECE, CSE, MSE, or Civil E.
20.6% of male students are declared in Biomed, CHBE, and ISE
37.5% of female students are declared in Biomed, CHBE, and ISE
These percentages reflect how much these majors affect the “average GPA” and grad rates across the college per gender. So, if there are any differences in the grad rates or GPA per major groupings noted above (i.e. the “difficulty” of the major), it becomes hard to declare that female students are “doing better” than their male counterparts. They may be, but the data here doesn’t prove it.
The MIT report mentioned that women had a higher GPA, after controlling for major. I don’t find this relationship a surprise since women have a higher GPA and graduation rate across the full national distribution, in both HS and college.
I haven’t seen stats showing how test scores differ for male and female students at MIT. However, I have seen stats showing that female applicants to MIT have a much higher admit rate than males, among similar high stat ranges, so I wouldn’t be surprised if women are treated as an underrepresented group and have slightly lower average test scores than men. The vast majority have high test scores at MIT, particularly on math, so there is a limited range for both women and men. IPEDS lists a 25 to 75th percentile range of 760 to 800. MIT has said that their admission process is no different for applicants with scores throughout this range, as well as somewhat below, as quoted:
“The MIT report mentioned that women had a higher GPA, after controlling for major. I don’t find this relationship a surprise since women have a higher GPA and graduation rate across the full national distribution, in both HS and college”
There is no data showing this, just a vague reference to the registrar. Browsing through the registrar docs do not reveal this info at all. True or not, the study does not support this statement in meaningful fashion. And higher gpas in college across all majors does not take the comparative rigor of different majors into account.
This in no way is stating that I don’t believe women can’t reach such success. However the supporting evidence is not a solid as advertised and I do find that asking everyone to take it as a given is a little lacking as well.
While they don’t list the specific GPAs, they give various details about the process, such as " multivariate regressions controlling for race, major, expected graduation year, and living group", “being based an analysis of first year and final year GPAs for all SB graduates who entered MIT as freshmen between 2004 and 2011”, and listing the percentage of women in different majors. Sure it’s possible the authors are lying or don’t know how to do a regression control for major, but I certainly wouldn’t assume this.
Women have a higher average GPA in HS, as well as college. Some historical stats for HS, in which major is not relevant, are at https://www.nationsreportcard.gov/hsts_2009/gender_gpa.aspx . Note that women average ~0.2 overall GPA higher than males, with a higher GPA in all core subjects during all surveyed years. The NAEP stats also indicates women average a more rigorous HS curriculum than males while achieving the higher average GPA.
MIT claims to norm by major. UCSD is simply “engineering” and F are highly overrepresented in Bioengineering. OSU has the breakdown you show. Obviously without having more detailed breakdown it’s impossible to know specific effect of major, but there would have to be a very significant and consistent difference in grading in majors within the same school to account for UCSD and OSU. And even then, it would unlikely to be enough to reflect the 30 difference.
@hebegebe you seem not to understand that it doesn’t matter if it’s a “marketing ploy” - if more colleges, particularly “Barons 1” or even “Barrons 2” colleges select a signficant # of students who are test optional, they have to believe they will “survive and thrive” (or they don’t give a ship about rankings - which is pretty unlikely) Grad rates, rep etc. all get factored into different rankings. Bates, for instance, consistently lands fairly high for LACs (around 44 ave):
USNews: 23
Niche: 51
Forbes: 35
WSJ: 66
Do you really think Bates is risking a top 50 or top 100 ranking for a marketing gimmick. What is Bates annual frosh enrollment? 500? You think they need to go “test optional” to fill that class. Ok.
That’s exactly the question/point. Do they “need” to? The point of this thread is that many (800 or so) don’t believe they do “need” to, and given that SAT/ACT is a large, expensive and time consuming business, students would be much better off without it if it does not meaningfully contribute to the admissions process.
There are a number of schools that feel they don’t “need” SAT/ACT scores. There is some (limited) evidence, provided by SAT themselves that SAT + GPA = “better” prediction of FYGPA (and that SAT + GPA + SATII = “best” prediction of FYGPA) but there is also evidence (USCS, for instance) that GPA is the “best” predictor of FYGPA.
The question we’re batting around here is do colleges really “need” SAT/ACT. My guess is “no.” Aside from extremely big colleges with limited adcom resources (UCs, etc.) who likely use standarized test scores + GPA to do a cutoff of applicants, most competitive colleges have adcoms that could probably look at a student’s app and tell you their SAT range.
At the end of the day, are SATs doing more than paying some guy (or guys) in Jersey 500k a year? And, even if colleges “need” SAT scores, why not at least let students self-report (like some colleges do with grades) so students are at least saved the expense of multiple submissions and rush scores. Once accepted, official scores can be sent to one school to confirm.
I know SAT/ACT score doesn’t have a high correlation to GPA in college, but I have a feeling that it does on how you do after school.
99.9% SAT score was for me a great way (maybe the only way) to get into an Ivy with 3.0 GPA in high school. I was always a lazy student academically (still am), but when I realized as a Junior that I had no chance to get into as good college as my friends with 4.0 GPA, my sense of pride took over and I studied consistently for the first time in my life 1.5 hour every day during summer vacation and aced SAT. Therefore, I am all for keeping SAT in college admission.
I even applied to Harvard and got denied. But hey, my SAT score was higher than the average SAT score at Harvard so I thought why not. That was the best 100 hours of study I did because it got me a free college education where I also managed to get 2.9 GPA. I like to learn but not study that much so engineering was out for me. I needed a major which required no discipline to graduate.
I never used the word “gimmick”. Your use of it suggests that you underestimate the importance of marketing to a college in terms of attracting the best student base that it can. Repeating what I wrote on the first page of this topic:
This is Marketing 101: Differentiate yourself from your competitors.
There are roughly 200 colleges that are selective. Among them, there is a distinct subset where test-optional makes the most sense, and these are colleges with small enrollments where the admissions department can review each application individually. Bates is an excellent example. Since the state flagships don’t meet this criteria, perhaps there are 75 in all among the top 200.
Beyond that, selectivity effectively disappears. So when you talk about 800 colleges being test-optional, only about 75 of them are using test-optional to actually improve the class relative to what they could otherwise get. The rest of them simply want to remove barriers so that they can fill their classrooms with warm bodies, any warm bodies, that help pay the bills. You are giving them far too much credit to assume they are going test-optional because it helps them create a “better” set of students.
I tried taking notes during lectures but when the lectures were actually interesting, I was too engrossed with what I was hearing to bother with taking notes, and when lectures were boring, well, I was too bored and was thinking about how to make a certain cute girl like me or play basketball later to bother taking notes. I married a woman who tried very hard academically and got nearly perfect grades but my kid says I am smarter. Lol. Therefore I tend to believe that SAT score does correlate to being sort of smart (or not dumb) even though it doesn’t correlate well with GPA which takes prolonged discipline and study efforts.
Looking back, I am amazed that I got 2.9 GPA in college without having never studied longer than one hour at a time and not taking notes. As a result, I feel I learned in college but didn’t get good GPA.
@hebegebe Right on point. I know many students with good GPA and bad test scores who did do very well at school but no way they would have gotten into top 20 schools. But if they had, they would have done well there also. All the respect for them. But you know why they do well? Because there are lazy students like me who don’t mind getting bad GPAs.
Nonetheless, as your post clearly agrees - these colleges, at least the 75 “selective” colleges clearly think this “marketing” will get them a better level or as good a level of student.
So again: why SAT? If you can simply get as good or better a class by getting rid of it?
Yet SAT themselves claim only ~15% of all SAT takers are “low GPA/high SAT” students. (And they have a vested interest in this scew being as big as possible.
So the question is, for the other 85% who have to pay, study, sit (1.6+ millon unique students per year, I think) is it worth it for that 15%.
Or, perhaps, could it be optional? Could colleges simply say: Hey, if you have a 3.85 of rigorous grades, don’t bother, we don’t need it. If you think your GPA does not reflect your ability, you can send standardized tests?
Or, perhpas, colleges are able to look at core subjects, level of grades, trend in grades, LOR, ECs, and determine: Well, here’s a 4 year 3 sport athlete with 4 consisent 4 hrs. a week of genuine volunteer work… we can assume some of this lower GPA is due to a full and rigorous schedule.
How is it we don’t trust colleges to do that. Again SAT studies show nly ~15% are low GPA/ High SAT, why should the other 85% (at least 1.3 - 1.4 million kids per year) spend the time, money and stress to take a test that will simply confirm what their 4 years of consistent effort and work already shows?
And at the very least, why don’t colleges allow students to self report so there are less economic barriers to applications. (I know why, and it has to do with mortgage payments in Far Hills NJ…)
"I know SAT/ACT score doesn’t have a high correlation to GPA in college, but I have a feeling that it does on how you do after school.
I even applied to Harvard and got denied. But hey, my SAT score was higher than the average SAT score at Harvard so I thought why not. That was the best 100 hours of study I did because it got me a free college education where I also managed to get 2.9 GPA. I like to learn but not study that much so engineering was out for me. I needed a major which required no discipline to graduate. "
No disrespect, but statistically a 2.9 GPA is not a high level of success in college. It is lower than most every average college GPA. So it appears your HSGPA was more predictive than your 99.9% SAT. If your 99.9% GPA were predictive you should have a 3.9 or so. Or at least an upper 25%. In fact, you provide very strong, if anecdotal, evidence for my argument. Your GPA indicated your likelihood of college success more than your SAT. Much more, apparently.
This is exactly the point: SAT does not even claim the SAT is predictive beyond FYGPA. While you may have that feeling, I don’t think there is much in the way of studies to confirm your feelings on this.
