4 ms·
> I had to pass essentially an IQ test to get placed into those classes. I don't see why race has anything to do with it. Simply put, we don't know how to make
by lmkg 4y ago
> I had to pass essentially an IQ test to get placed into those classes. I don't see why race has anything to do with it.
Simply put, we don't know how to make objective IQ tests yet.
The questions in IQ tests tend to bake in a bunch of cultural assumptions, often in ways you don't expect. So the test measures a combination of IQ, and affinity with a specific cultural background.
First, let me be clear about this: The issue is friction. Even at the same difficulty, a question that's more accessible is less stressful and less mentally taxing. Especially when you're dealing with kids. This matters more at the cut-off than the extremes of the bell-curve, but has statistical impacts.
I just searched for "Fourth Grade IQ Test" and clicked a random link. Four of the ten questions are about sharing cookies. Baking cookies and sharing cookies are typical middle-class-white activity, but I don't know how that shakes out in other cultures. Maybe they share other foods? Maybe sharing food is offensive because it implies the other kids' parents can't provide for themselves? Maybe the kids aren't in charge of food division? The underlying math is, of course, objective. But the math is being framed through a perspective which is more or less familiar to different backgrounds.
- dpbriggs 4y agoAny further reading recommendations? It's unclear why the cultural example of baking and sharing food would cause significant differences in test results but that's picking at an example. Were the six other questions also stereotypical of white middle class families? I imagine they were more abstract (lacking cultural context?). Would the solution here be to remove cultural context, or have culturally tailored exams? It would be interesting to take IQ tests from other countries and compare results. I imagine other factors such as home life, economic status, and cultural emphasis on education would be much, much stronger predictors on IQ test performance.
- lmkg 4y agoA few straight arithmetic questions, one about the alphabet, a few about parts of speech. These don't immediately jump out to me as cultured, although anything involving letters or grammar will be sensitive to the linguistic background of the students. Certainly home life, economic background, etc have an impact on educational outcomes. But I'm making a narrower claim here. A student with a poor home life is likely to test poorly regardless, although the difference might be larger in some tests than others. But a student with the same educational aptitude may test better or worse depending on if they're being tested on topics that are familiar or unfamiliar to them. Removing cultural context is certainly a solution. I'm not sure it's the best solution. Excessively-abstract problems are only one aspect of intelligence. Anything involving reading comprehension is going to have some linguistic or cultural elements, but I'm not sure it's valid to just remove reading comprehension altogether. If you want to see a standardized test from a different culture, here's one that's made the rounds a few times. This is for eight-graders in Kentucky in 1912. The sections on Geography & Physiology certainly stand out, but even the math section has a notably different emphasis than modern education. https://www.bullittcountyhistory.com/bchistory/schoolexam1912.html https://www.bullittcountyhistory.com/bchistory/schoolexam191...
- tick_tock_tick 4y agoIQ tests are incredibly predictive nowadays.
- nabla9 4y agoThe Cattell Culture Fair III (CFIT-III) is really good. If we are talking IQ scores within the same western country, we know now to make culture fair fluid intelligence test (it's not the same as future achievement score)