4 ms·

It is well known and universally acknowledged that SAT scores are highly predictive of educational outcomes. See eg. https://collegereadiness.collegeboard.org/e

by xyzzyz 7y ago

It is well known and universally acknowledged that SAT scores are highly predictive of educational outcomes. See eg. https://collegereadiness.collegeboard.org/educators/higher-ed/test-validity-design/validity-studies https://collegereadiness.collegeboard.org/educators/higher-e... . You can use other alternative measures to predict educational success, but it’s important to get it right, otherwise you’ll end up with low quality student body that isn’t able to learn much and will drop out at high rates. Currently, high school GPA is also good at predicting success, but it’s highly variable among schools. Should SAT testing be abolished, parents who now pay for prepping will instead put their children in schools where they are guaranteed high GPAs. With all of its flaws, standardized testing is still the best of all available alternatives for the purpose it’s used for. You need to acknowledge that no matter what approach you use, it will be gamed.Thus, if you want to ensure fairness, you need to choose one that’s hardest to game. Clearly, high school GPA, essays and volunteering is much easier to game than SAT score.

thatfrenchguy 7y ago
> See eg. https://collegereadiness.collegeboard.org/educators/higher-e.. https://collegereadiness.collegeboard.org/educators/higher-e.... Quoting the company that sells you those tests to tell you those tests are predictive is quite fun.
- xyzzyz 7y ago
  Their result match everyone else’s, so I see no reason to not cite them.
  - timmytokyo 7y ago
    And apparently you see no reason to cite "everyone else's", even when that would have strengthened the persuasiveness of your argument. Hm.
    - xyzzyz 7y ago
      https://psycnet.apa.org/record/1976-08398-001 https://psycnet.apa.org/record/1976-08398-001 https://www.tandfonline.com/doi/abs/10.1080/00221546.1998.11775147?journalCode=uhej20 https://www.tandfonline.com/doi/abs/10.1080/00221546.1998.11... (this one summarizes something like 20 other studies in a nice table) https://www.tandfonline.com/doi/abs/10.1207/S15326977EA0801_01 https://www.tandfonline.com/doi/abs/10.1207/S15326977EA0801_... https://journals.sagepub.com/doi/abs/10.1177/07399863930153004 https://journals.sagepub.com/doi/abs/10.1177/073998639301530... https://journals.sagepub.com/doi/abs/10.1177/001316447903900220 https://journals.sagepub.com/doi/abs/10.1177/001316447903900... https://psycnet.apa.org/record/1968-09427-001 https://psycnet.apa.org/record/1968-09427-001 You want more? I could give you probably 50 more studies. The literature here is vast and consistent.
fiftyfifty 7y ago
That doesn't jive with the submitted article in the Washington Post: "Research has consistently shown that ACT and SAT scores are strongly linked to family income, mother’s education level and race. The College Board and ACT Inc., which owns the ACT, say their tests are predictive of college success, but (as with many education issues) there is also research showing otherwise." There are studies that have shown rigor of high school course work is the strongest predictor of college success: https://www.chronicle.com/article/Study-Says-Rigor-of/8973 https://www.chronicle.com/article/Study-Says-Rigor-of/8973 I think we can all agree that allowing two private companies to be the gate keepers of higher education is a terrible idea.
- danieltillett 7y ago
  This is because all of these factors are heavily g loaded. I do agree with your last sentence.
- haihaibye 7y ago
  I thought the best measure was to combine school grades and tests. But grades are so subjective - who has time to evaluate the rigor of a school curriculum, the teacher teaching it and a student learning it each year? Hence standardized tests. > Research has consistently shown that ACT and SAT scores are strongly linked to family income, mother’s education level and race. So is IQ, school and career success - and anything that predicts them, which is supposedly what we want. But innate differences are crimethink, so best shift from measure with predictive ability to those that give the desired results. > I think we can all agree that allowing two private companies to be the gate keepers of higher education is a terrible idea Yes, in my country this is done by the government.
- xyzzyz 7y ago
  This WaPo quote is a good example of a common practice in modern journalism, which is to write something which literally is true, but leaves the reader with the wrong impression of the reality. Yes, the rich people tend to do better on SAT. So? Why would anyone expect all groups to do equally well on SAT? There is no problem with that. It would only be a problem if SAT had different predictive validity between the rich and the poor, that is, poor people with a given SAT score had different outcome in predicted variable than the rich people. By and large it is not true. > The College Board and ACT Inc., which owns the ACT, say their tests are predictive of college success, but (as with many education issues) there is also research showing otherwise." Of course there is research showing otherwise. There always is. The real question here, which WaPo is eliding, is what’s the preponderance of evidence, and it overwhelmingly is on the side of high predictive validity of SAT. Don’t believe me, see for yourself: go to Google Scholar and search for “sat predictive validity”. > There are studies that have shown rigor of high school course work is the strongest predictor of college success: https://www.chronicle.com/article/Study-Says-Rigor-of/8973 https://www.chronicle.com/article/Study-Says-Rigor-of/8973 It is marginally better than SAT, and it’s worse than GPA + SAT. I predict if we dropped SAT, it’s predictive validity would fall significantly, because it’s much easier to game GPA than to game SAT. > I think we can all agree that allowing two private companies to be the gate keepers of higher education is a terrible idea. Where I went to school, a government agency designs the test, of which all students take the exact same version on exact same date. Universities by and large exclusively consider the score in admissions. I think it’s superior system, more fair and egalitarian.

by xyzzyz 7y ago

thatfrenchguy 7y ago

> See eg. https://collegereadiness.collegeboard.org/educators/higher-e.. https://collegereadiness.collegeboard.org/educators/higher-e.... Quoting the company that sells you those tests to tell you those tests are predictive is quite fun.

xyzzyz 7y ago
Their result match everyone else’s, so I see no reason to not cite them.
- timmytokyo 7y ago
  And apparently you see no reason to cite "everyone else's", even when that would have strengthened the persuasiveness of your argument. Hm.
  - xyzzyz 7y ago
    https://psycnet.apa.org/record/1976-08398-001 https://psycnet.apa.org/record/1976-08398-001 https://www.tandfonline.com/doi/abs/10.1080/00221546.1998.11775147?journalCode=uhej20 https://www.tandfonline.com/doi/abs/10.1080/00221546.1998.11... (this one summarizes something like 20 other studies in a nice table) https://www.tandfonline.com/doi/abs/10.1207/S15326977EA0801_01 https://www.tandfonline.com/doi/abs/10.1207/S15326977EA0801_... https://journals.sagepub.com/doi/abs/10.1177/07399863930153004 https://journals.sagepub.com/doi/abs/10.1177/073998639301530... https://journals.sagepub.com/doi/abs/10.1177/001316447903900220 https://journals.sagepub.com/doi/abs/10.1177/001316447903900... https://psycnet.apa.org/record/1968-09427-001 https://psycnet.apa.org/record/1968-09427-001 You want more? I could give you probably 50 more studies. The literature here is vast and consistent.

fiftyfifty 7y ago

That doesn't jive with the submitted article in the Washington Post: "Research has consistently shown that ACT and SAT scores are strongly linked to family income, mother’s education level and race. The College Board and ACT Inc., which owns the ACT, say their tests are predictive of college success, but (as with many education issues) there is also research showing otherwise." There are studies that have shown rigor of high school course work is the strongest predictor of college success: https://www.chronicle.com/article/Study-Says-Rigor-of/8973 https://www.chronicle.com/article/Study-Says-Rigor-of/8973 I think we can all agree that allowing two private companies to be the gate keepers of higher education is a terrible idea.

danieltillett 7y ago
This is because all of these factors are heavily g loaded. I do agree with your last sentence.
haihaibye 7y ago
I thought the best measure was to combine school grades and tests. But grades are so subjective - who has time to evaluate the rigor of a school curriculum, the teacher teaching it and a student learning it each year? Hence standardized tests. > Research has consistently shown that ACT and SAT scores are strongly linked to family income, mother’s education level and race. So is IQ, school and career success - and anything that predicts them, which is supposedly what we want. But innate differences are crimethink, so best shift from measure with predictive ability to those that give the desired results. > I think we can all agree that allowing two private companies to be the gate keepers of higher education is a terrible idea Yes, in my country this is done by the government.
xyzzyz 7y ago
This WaPo quote is a good example of a common practice in modern journalism, which is to write something which literally is true, but leaves the reader with the wrong impression of the reality. Yes, the rich people tend to do better on SAT. So? Why would anyone expect all groups to do equally well on SAT? There is no problem with that. It would only be a problem if SAT had different predictive validity between the rich and the poor, that is, poor people with a given SAT score had different outcome in predicted variable than the rich people. By and large it is not true. > The College Board and ACT Inc., which owns the ACT, say their tests are predictive of college success, but (as with many education issues) there is also research showing otherwise." Of course there is research showing otherwise. There always is. The real question here, which WaPo is eliding, is what’s the preponderance of evidence, and it overwhelmingly is on the side of high predictive validity of SAT. Don’t believe me, see for yourself: go to Google Scholar and search for “sat predictive validity”. > There are studies that have shown rigor of high school course work is the strongest predictor of college success: https://www.chronicle.com/article/Study-Says-Rigor-of/8973 https://www.chronicle.com/article/Study-Says-Rigor-of/8973 It is marginally better than SAT, and it’s worse than GPA + SAT. I predict if we dropped SAT, it’s predictive validity would fall significantly, because it’s much easier to game GPA than to game SAT. > I think we can all agree that allowing two private companies to be the gate keepers of higher education is a terrible idea. Where I went to school, a government agency designs the test, of which all students take the exact same version on exact same date. Universities by and large exclusively consider the score in admissions. I think it’s superior system, more fair and egalitarian.