“A new study suggests affirmative action policies in college admissions might be a good idea.” …
Per article- SAT test prep has ruined its reliability. My opinion - Its no exaggeration to say that some students spend years and thousands of hours of study trying to enhance their score. If successful, it makes them look far better on paper than their actual skills will likely demonstrate in the classroom. Those students who do not prep will offer a true statistical projection of ability.
BTW the thread should be labeled "For Blacks students. SAT scores are more than likely an accurate prediction of ability’.
So, the point is is that the SAT is better at predicting academic success in black students than in white students?
Wow, what a misleading title.
What it should be, according to the article:
“SAT scores more likely to predict college performance for black students than for non-blacks”
And to be honest, points are one thing, but what percentage increase was seen? If, as posted elsewhere on CC, blacks tend to have lower GPAs, then it is even more remarkable.
What I’d like to see:
- GPAs of blacks and whites and Latinos (as usual, everyone in one category no multiracial folks exist
) for each range of SAT scores - Proof of socioeconomic class being immaterial for all races, including poverty and single parent or both parents working homes
- Also look at rural vs. urban, and relative HS quality (top 10 percent in one public school <> top 10 percent in another)
And: “The researchers argue that these test score policies might end up excluding able black students.” What? You just said if there are kids with higher test scores, they DO do better? Why would you lower standards, you want kids with lower GPAs who can’t get jobs? It doesn’t seem like a valid conclusion from the data and results they present.
I am reading it like this, assuming 200 points is the standard deviation for this example:
White student A gets a 1500 SAT score and white student B gets a 1700 SAT score. White student B does 0.347 points higher in GPA than white student A.
Black student 1 gets a 1500 SAT score and black student 2 gets a 1700 SAT score. Black student 2 does 0.499 points higher in GPA than black student 1.
So, assuming 3.0 for each of the first students:
White student A: 3.0
White student B: 3.347
Black student 1: 3.0
Black student 2: 3.499
Is that significant? Would an employer hire a 3.5 student over a 3.3 student, or is the question if an employer would hire a 3.5 student over a 3.0 student?
I’m not comfortable with the race thing, but without a doubt the SAT is better at predicting ability for those who do not overly prepare for it.
^ Yes, although they infer that a conclusion could be that the SAT is better at predicting college success for students who cannot afford test prep sessions. For students who do participate in SAT test prep, the SAT is not that meaningful of a predictor. They focus on predictors of college success for blacks vs whites, but their story could apply to lower-income vs middle-income students instead.
@rhandco - In regards to the article? The real question is - do students who over prep for the SAT have a natural ability commensurate with their score?
^ No, the idea is that for students who prep for the SAT (and presumably get a much higher score due to the preparation itself) the resulting score is not a good predictor of college success (given other information such as class rank already taken into account).
What is “over prep”? Some urban schools have access to free prep courses. For example, although I don’t live in a city, there is a free college prep program at the local Y that includes SAT prep, but only for URM and/or poor students (free/reduced lunch).
You can argue that GPA is easier to inflate than standardized test scores because of the difference in teachers and school policies. But this study appears to look only at test scores. What they needed to do is look at the top 10 percent compared to the 2nd decile, the second 10 percent (11th through 20th percent) students, and see if being in the top 10 percent mattered more than test scores. Because I would be that there are many students in the second 10 percent who did better on SATs than some in the top 10 percent.
And I’m in a town which has a population of rich kids, and I know not one who spent “years” and/or “thousands of hours” on SAT prep. The vast majority who took test prep courses took one 40-hour course and bought a review book and spent maybe another 40 hours on review. That’s not anywhere near thousands of hours.
But I do know some kids have daily tutors who help them significantly with homework and projects. There is a reason that they use both quarter grades and mid-term and final exam grades for getting into advanced classes.
GPA’s can certainly be inflated, which actually adds to the problem, so really the only solution is to continue the holistic approach.
http://ftp.iza.org/dp8733.pdf is the actual working paper.