I have heard the argument that intelligence as measured by IQ tests is a Western construct that does not translate to other cultures. The claim is that what counts as smart in one culture is completely different in another, so tests built on Western concepts of reasoning and problem solving are inherently biased toward Western populations. Is there any merit to this argument or does cognitive ability research actually hold up cross culturally?
There is a weaker and a stronger version of this claim and they need to be separated. The weaker version is that specific test content can be culturally loaded and that format familiarity varies across populations. That is true and well documented, and it is why test developers put significant effort into reducing surface level cultural content from reasoning tasks. Abstract matrix puzzles and figural reasoning items exist precisely to minimize the advantage of Western educational exposure. The stronger version, that intelligence itself is a Western construct with no cross cultural validity, is not supported by the data. The positive manifold, meaning the tendency for cognitive abilities to correlate positively with each other, shows up consistently across every population that has been studied including non Western and non industrialized ones. G emerges from factor analyses of cognitive data regardless of where the data was collected. Processing speed, working memory, and fluid reasoning predict outcomes cross culturally. The architecture of human cognition appears to be species typical, not culturally specific, even if the expression and valuation of particular abilities varies across contexts.
The positive manifold appearing consistently across unrelated cultures is probably the strongest evidence against the Western construct argument. You do not get the same factor structure everywhere by accident if the whole framework is just a cultural artifact.