Don't get me started on lefty liberal constructs like "social", "interpersonal" or the newly fangled "emotional intelligence".
For me intelligence is raw cognitive power. Things like interpersonal intelligence involve not being an a-hole and have nothing to do with intelligence IMHO.
That is unfortunately a very common and very arbitrary take on a protean notion of intelligence. The supposed generality of human intelligence is not a law of the universe, but a convenient postulate of the talentless plethora. Most experts in fields like science in math are not trapped within the limited rules of that field (and not by chance that almost all practical contributions in science/math, come from experts within a field - not plumbers with a sense of rationalistic, intellectual grandiosity. High expertise in one (open) system, doesn't mean that one isn't free to create elementary associations with information outside of their field. If a field of study includes facts (elements) A, B, and C, there is nothing that prevents an expert from associating element B with element D (on the grounds of some subtle pattern or anomaly), from some other field. The brain does not compartmentalize information sets, it has very broad neuronal networks that constantly associate patterns. Of course, there are always those 'intellectual dicks' who are purely interested in aimless 'fact finding' within their respective fields - these are people confuse their ability to learn with intelligence, and are extremely narrow thinkers. It's very likely not just IQ, but limitations of the neural inter-complexity and computational speed (intuition) of individuals is what allows some to actually apply information they've learned in ways to solve problems and exhibit higher learning through forming higher n order abstractions. Individual performance can and should only be evaluated at any specific given moment and time and are the product of developmental factors. One can only abstract from a given system like 41, 25, 49, n , if they are given that system to analyze - effectively nullifying any supposed quality of 'intelligence'. You can't give someone half of an idea, and then give them credit for the whole. In the real world, we don't know what elements are in are set, or what, when or exactly where we have to think hard. And ideally, 'smart' people should be defined as those who come up with new ideas from their own experimental models (sets)....yeah we don't live an ideal world.
Ultimately I'd expect a far closer correlation of chess players with IQ than e.g. scientists; because chess lines up much better with IQ's testing methodology, and all IQ measures is how well you score on IQ tests.
From what I've read, there's a positive correlation between chess skills and IQ, and it's very weak. Chess skills mostly depend on the time and effort you put into practicing and learning chess. IQ is a very very poor substitute.
I wasn't saying I expected being good at IQ tests makes you better at chess, but more the opposite: getting good a chess makes you better at IQ tests.
Chess itself includes some very specific skills, but the spatial reasoning tests/puzzles they do in IQ are more general so to my estimation are more likely to improve if you improve your chess ability (especially if you play a lot of varients).
I'm not familiar with the studies you're referencing, did the people do IQ tests then learn chess or the other way around?