Students’ high-school grade point averages are five times stronger than their ACT scores at predicting college graduation, according to a new study from the University of Chicago Consortium on School Research.
UChicago Consortium researchers found that the predictive power of GPAs is consistent across high schools—something that did not hold true for test scores. At many high schools, they discovered no connection between students’ ACT scores and eventual college graduation. The authors were also surprised to find that, at some high schools, students with the highest ACT scores were less likely to succeed in college.
“Across the high schools studied, students with high school GPAs under 1.5 had around a 20% chance of graduating from college. For students with GPAs of 3.75 or higher, those chances rose to around 80%.”
That 20% of students with HS GPA < 1.5 graduate from college suggests that there is a non trivial number of “late bloomers”, and that one’s HS academic performance does not necessarily mean that one will continue to do similarly later.
I’ll file this one under “duhhhh”. This study was already done with SAT scores, I don’t know why a different outcome would be expected.
“ These findings suggest that college admissions may be overemphasizing test scores,” seems to be a huge, unsupported guess. I see nothing in the description of the study that they examined how all colleges weigh GPA vs test scores and how they quantitatively evaluate “overemphasize”.
“ Extensive time spent preparing for standardized tests will have much less payoff for postsecondary success than effort put into coursework, as reflected in students’ grades,”. “Extensive time” preparing for tests in no way compares to the 4,000+ hours a student spends on coursework over high school, so I don’t know why this needs to be pointed out. Is anyone really stopping schoolwork and getting poor grades to spend time on SAT prep (if so, shame on them).
Isn’t it pretty common knowledge that HS grades and course rigor (“ school effects may be the result of more rigorous academic programs”) are the very top criteria used by a vast, vast majority of colleges? So schools’ top consideration are those that are most predictive. Makes sense.
It shows that a less-weighted component of an application is a less predictive tool. I would wager that this is true of all application components that aren’t as important as grades - essays, EC, etc., etc.
The statement that “some” or “many” schools show odd relationships sounds like cherry picking. If it was a majority, I’m sure it would be pointed out. When I have time I’ll try to hunt down the study and see how many there were.
If the inference is that a school shouldn’t use any criteria that aren’t the most predictive, then I can assume those that support this idea think a transcript and school profile are the only things that should be on a college application.
One explanation is the institutions that will admit students with sub 1.5 GPAs have easy grading standards and try their best to graduate anyone who demonstrates at least minimal competency.
https://journals.sagepub.com/doi/10.3102/0013189X20902110 is the actual paper.
1.5 HS GPA would only realistically get into an open admission college like a community college. But it seems unlikely that a new HS graduate with a 1.5 HS GPA would make even a 2.0 GPA to avoid being dismissed from college. The more likely scenario is probably going to a non college path first (work or military service) and later returning to school as a non traditional student with much more motivation and maturity than s/he had in high school.
The study only included traditional students for modeling graduation rates.
It’s too bad the UC Academic Senate STTF report wasn’t available to the authors.
I wonder what 4 year colleges the 13.9% of students in this sample who had under a 2.0 HSGPA went to, or the 30.7% of students who had below an 18 on the ACT? (There is also have a typo on Table 2…assuming the ACT column labeled ‘14-16’ should be ‘14-15’)
There are so many questions I have with their methodologies…including considering all courses in unweighted GPAs, not just core. I expect there would be huge variability across schools (even in a school system that is rated on many measures as ‘low performing’) in this measure…perhaps enough to affect the outcomes. My point is comparing students who have 16-20 core courses, with those who might have far fewer, could skew results.
I wrote a possible reason why the two studies had opposite results in the other thread:
http://talk.qa.collegeconfidential.com/discussion/comment/22590079/#Comment_22590079
"1.5 HS GPA would only realistically get into an open admission college like a community college. "
Right, but all the kids in the study got into a 4-year college including those with below 2.0 GPAs and in some cases below a 17 ACT. They didn’t include kids who went to a c/c or military. It’s interesting data, but you have to know the colleges and majors the kids went to before you made any policy decisions from it though.
It already has, it looks like. Knowing about this may have been behind the decision to make all standardized testing optional.
That said, it’s a really interesting study, and a really complex one. All of the data comes from the Chicago Public Schools, which means (a) there is a large variation among the schools, which is a big part of what was being studied, but (b) there may well be less variation among public schools in the same district than there would be if the whole range of schools in different places with different governance and different curriculums were considered. In other words, notwithstanding the extreme variety of public schools in Chicago, it would not be that surprising to find that a “B” grade has more consistent meaning across those schools vs. across all schools of all types nationally. And that finding – relative consistency of GPA meaning across schools – is central to the study. There are other interesting aspects to look at, too.
Also, I suspect this study – which focuses on college enrollment and 6-year college graduation rate – tells us almost nothing about admissions and success at UChicago or any other hyperselective, elite college. I can’t look into that now, though.
I wonder whether underlying the close correlation between good grades and success in college is another correlation - between the quality of the high school in which those good grades are obtained and the level of difficulty of the college to which those good grade-earners tend to apply and flourish in. An academically demanding school like Chicago would not assume that good marks earned in a poor school would translate to success. More would be required. I expect good test scores would be a helpful clue in such a case - along with everything else that makes for a good application, including sterling character, impressive exploits, and an inspiring life story replete with plaudits, accolades, and patents. And don’t forget the quirkiness and the killer prose!
There’s already a thread about this: http://talk.qa.collegeconfidential.com/college-admissions/2174485-uchicago-consortium-study-finds-high-school-gpas-outweigh-acts-for-college-readiness.html#latest
Will this impact the way they choose who gets in?
https://news.uchicago.edu/story/test-scores-dont-stack-gpas-predicting-college-success
This looks like one of those studies where they are just trying to prove their point to propagate a cause. GPA is very student, teacher, course selection, and school dependent. I’m not supporting ACT but GPA and ACT together give a better idea than on their own.
- Tend to agree. The perspective and specific requirements of a highly-selective institution such as UChicago may limit the applicability of this more generalized set of conclusions, particularly because most of the GPA and ACT ranges would be considered on the lower end of the distribution for incoming admits. UChicago might instead be more interested in the UC Faculty Senate study, since that one specifically addressed the role of standardized testing in admissions policy for a highly selective research institution. The Faculty Senate study also examined a more detailed set of "success" measures in addition to college graduation rates, such as freshman GPA and retention rates and overall college GPA (which would factor in to graduate school decisions).
One nit about this study: Where are the R-squared values? I’d think that they would want to include measures of predictive validity for GPA and ACT, especially because they use terminology like “predictive power” in the press release! The study highlights GPA as packing the greater punch in an admission decision; however, if the R-square is lower than for ACT, it’s simply a less reliable predictor, even if the odds appear more favorable. For instance, the presence of grade inflation might both increase the odds AND lower predictive validity.
Actually that is exactly what the study says. Both are predictors, but GPA is a better one.
6-year graduation rates isn’t the only important outcome measure. Unless they’re all going to for profit diploma mills, I’m curious where the sub 2.0 students are getting admitted. UCB mentioned CCs, but according to the study only students enrolled in 4-year colleges were included.
That’s right - the study skipped gappers and others who delayed the start of college, and only looked at and followed enrollments at 4 year colleges/uni’s.
“Actually that is exactly what the study says. Both are predictors, but GPA is a better one.”
- I do wonder if that's what the study is actually saying. In a regression analysis, "better predictor" refers to the statistical validity of the variable, not whether it leads to a certain estimate of the dependent variable (in this case, odds of finishing college in 6 years). If ACT is better than GPA at nailing the odds, then ACT is a "better predictor" than GPA even if the expected odds are lower. Simply because there is more confidence surrounding the odds estimate.
It’s important to point out that the CPS data, including the 11th grade required ACT, is a treasure trove for a social scientist; however, it’s not surprising to see that predictive abilities would be consistent across high schools in the same district. Whether that conclusion can be generalized will depend on similar analyses using data from other districts, even those with different characteristics (ie suburban or rural).