You are viewing a single comment's thread.

view the rest of the comments →


[–] iswallow 10 points -8 points (+2|-10) ago 

Part of the race gap can be explained by test being biased, i.e. measurement invariance. This means that two people with the same level of ability, should have the same chance of answering a question correctly. For some questions people with the same ability however do not have the same chance of answering correctly (example could be a math question in a football setting for which males might have a higher chance than females with the same math ability for answering the question correctly).

If your test has measurement invariance, it's fair to correct for it. The aim is to make sure that people with the same ability have the same scores. A problem however is that current corrections seem to be overcompensating just to get equal outcomes, instead of purely correcting for measurement invariance.


[–] HipFlask 1 points 9 points (+10|-1) ago 

Source your statements, they are false. Intelligence tests used to be blamed for asking culturally loaded questions (e.g. you had to know what a "regatta" is). It was argued that a black kid from the inner city would be at an disadvantage. Then they studied that and tested it. Turn out: Nope. Wether a test item was culturally loaded had no impact on the ability of blacks to solve it. They were just less intelligent. This gap in intelligence (aboput one standard deviation) is what is compensated in SAT scores. My source is "the bell curve". More specific papers are sopurced in that book.


[–] iswallow 1 points -1 points (+0|-1) ago 

You want a source on measurement invariance? Pick up any test design book, or any item response theory or modern test theory book. You can test for each item in a test wether it is measurement invariant. That is you can test for each item whether two individuals with the same ability that the item tries to measure, wether two individuals from different groups have the same chance of answering an item correct.


[–] otto_mobile 0 points 5 points (+5|-0) ago 

if you use a new word like "measurement invariance", you can sound smart while repeating the exact same criticisms of IQ and related tests that have been made for the past hundred years of IQ and related tests existing.


[–] eagleshigh 1 points 1 points (+2|-1) ago 

Negros DO score higher on IQ tests than Whites!