April 12, 2012

SAT scores & race: Everything you ever wanted to know

The Unsilenced Science has the world's most comprehensive blog post on SAT scores and race. Check it out.

13 comments:

Chuck said...

So I wrote Chris Brand about the lack of all but a small Black-White IQ gap in the UK; his response, like Greg Cochran's, was to dismiss my numerous data points and then to argue that the tests were not g-loaded enough, even though I pointed to research which established otherwise. Sounds like hereditarian rationalization to me.

rob said...

I've always wondered why the SAT doesn't have any sections that load on visuospatial ability. It's important for engineering, organic chem. Including VS questions would most likely help boys, hurt Jews, help Asians, and hurt blacks. I think VS tests are harder to study for, but I could be totes wrong.

As for non-respondents to race: non-Whites would have to be really stupid to forego the racial preferences, and I think I remember AA being a pretty big deal in the news c. 2000, so marginal whites were more aware that schools wanted to discriminate against them. I can buy that they'd try to make discrimination harder for schools.

Anonymous said...

Wow, a graph that finally picks logical colors to represent each group.

One of the graphs seems to make no sense. It shows IQ peaking for SAT scores around 1300 and then falling. How does that work?

Bob Arctor said...

Immigrants are are a biased, not random, sample. Look at South Asian scores in US/Canada versus India and Pakistan for an example of this.

Anonymous said...

Chuck,

What tests in the UK are you referring to? My impression was that the gaps are significant for the GSCE's, but I could be wrong.

Chuck said...

"What tests in the UK are you referring to? My impression was that the gaps are significant for the GSCE's, but I could be wrong"

You can refer to my extended discussion of this over at Occidentalist. I could be wrong on this issue. But the goofy explanations that I'm getting are not, in the least, reassuring.

nooffensebut said...

@ rob

I cannot buy the racial-preferences explanation because the students also did not answer unrelated questions. Plus, I would not expect such a group to just disappear. If anything, polls show that affirmative action support has declined since Obama’s election.

@ Anonymous

I agree that this graph does not make sense. I am sure that Frey and Detterman would defend their equations by saying that the equations should not be applied to the full SAT range because the equations resulted from individuals within a specific range of scores.

Anonymous said...

"I've always wondered why the SAT doesn't have any sections that load on visuospatial ability. It's important for engineering, organic chem. Including VS questions would most likely help boys, hurt Jews,"

I think you just answered your own question.

Look, they added a redundant writing section so as to help women and those with strong verbal skills. The new SAT double counts verbal/BS ability.

Anonymous said...

"So I wrote Chris Brand about the lack of all but a small Black-White IQ gap in the UK; his response, like Greg Cochran's, was to dismiss my numerous data points and then to argue that the tests were not g-loaded enough, even though I pointed to research which established otherwise. Sounds like hereditarian rationalization to me."

Hmm, how about the blacks in the UK? Aren't they mostly from Barbados? How does Barbados do on tests? Also, how black are they? Are they mostly mixed race with the white contribution from males in Barbados? Has anyone looked at that? I mean only one black was admitted to Oxford and only six at Cambridge. So, if they are really that close, why such a huge gap in college admissions?

http://www.guardian.co.uk/education/2010/dec/06/oxford-colleges-no-black-students

fing ding said...

I've analyzed NAEP data by race and STAR (California) data by race and poverty and created PDFs with the results:

2011 NAEP Math, Science and Reading (whites outperfrm black in every subject in every state):
http://depositfiles.com/files/t6k1t79nd
or
http://pdfcast.org/pdf/naep-test-results-by-race


2011 STAR results by race and economic status (whites outperform blacks in all 21 subjects in each economic category -whites who are economically disadvantaged even outperform black who are not ecomically disadvanted in 20 out of 21 subjects):

http://depositfiles.com/files/8jpgerl85
or
http://pdfcast.org/pdf/2011-star-test-results-by-race

Chuck said...

"Sounds like hereditarian rationalization to me"

If anyone's interested, I would be willing to pay up to $200 for a basic analysis of the UK PISA (2006/2009) data. I just want plausible value scores by ethnicity by nativity. I would do it but I don't currently have the time or energy. The data can be downloaded online. You will need a software package such as SPSS to analyze it. It took me about 10 hours to figure out how to analyze the US data so I am figuring $20 and hour x 10. If you are interested drop a comment over at Occidentalist --Chuck

Anonymous said...

The Native American IQ and scoring gaps are very interesting. Mainly because the gap does not appear to be that large at all.

Anonymous said...

The Native American IQ and scoring gaps are very interesting. Mainly because the gap does not appear to be that large at all."

You only have to have a small fraction of Native A. ancestry to claim it for AA purposes. Plenty of those people in the Native A. demo are 87.5% white. Others have a bit more Native A. DNA which brings the average down some. Hispanics generally have more native A. DNA and also have lower scores. Truly white folks like myself who have some Spanish ancestors from Mexico don't self identify as "hispanic".