Tapsearch Com EditorIQ, race and prediction
September 28th 2007 01:13
I'm a little skeptical of this J. Philippe Rushton VDARE piece.
One of the common points made about standardized testing is that tests predict life outcomes for blacks as well as they do for whites. A white person and a black person with the same SAT or IQ will do about the same in school and life, and if anything the black person will do a little worse -- if IQ tests were biased against blacks, they would underpredict black achievement, and blacks would do better than their scores predicted.
But this is what the VDARE article says:
The Bell Curve documented that IQ scores predict equally well for all groups. For example, Blacks with IQs of 114 have an equal (or better) chance of graduating from college than Whites and Latinos with the same IQs—68%, 50%, and 49%, respectively, and also of getting top jobs (likely a result of affirmative action programs).
Actually, those numbers indicate that IQ tests are biased against blacks. A higher ratio of blacks than whites with a given IQ will graduate college -- that means IQ scores are underpredicting black performance. Did Rushton get his numbers messed up?
Of course, affirmative action does as good a job explaining this as underprediction does, but the numbers don't have the meaning Rushton ascribes to them at all.
One of the common points made about standardized testing is that tests predict life outcomes for blacks as well as they do for whites. A white person and a black person with the same SAT or IQ will do about the same in school and life, and if anything the black person will do a little worse -- if IQ tests were biased against blacks, they would underpredict black achievement, and blacks would do better than their scores predicted.
But this is what the VDARE article says:
The Bell Curve documented that IQ scores predict equally well for all groups. For example, Blacks with IQs of 114 have an equal (or better) chance of graduating from college than Whites and Latinos with the same IQs—68%, 50%, and 49%, respectively, and also of getting top jobs (likely a result of affirmative action programs).
Actually, those numbers indicate that IQ tests are biased against blacks. A higher ratio of blacks than whites with a given IQ will graduate college -- that means IQ scores are underpredicting black performance. Did Rushton get his numbers messed up?
Of course, affirmative action does as good a job explaining this as underprediction does, but the numbers don't have the meaning Rushton ascribes to them at all.
| 71 |
| Vote |
Add To: 
