I'm skeptical that your premises are correct. But even assuming they are, for argument's sake, how does a characteristic about the extreme tails of the distribution of "ability" (arguably >3sd) translate into a meaningful comment about a gender split in chess participation closer the mean, say +1-3sd, where most players fall? You can't meaningfully infer the expected performance of the gender distribution from the tails.
Ok, which premise do you doubt?
Do you doubt that there is a difference in spatial ability between boys and girls? My understanding is that spatial ability is a reason why the strongest boys still do better in math than the strongest girls, even though girls are generally doing better in school than boys these days.
I do, for the reasons ultramarina has elucidated. It's the classic problem of inferring ability from outcomes without adequately cotrolling for confounding factors. The studies that attempt to tease out gender differences in ability don't control for socialization or prevailing gender norms, for example. To see such a profound change in the gender ratio over time suggests I'm correct that the causal factors haven't been properly modelled yet in the literature.
So do I believe studies have found a gender difference in spatial skills? Sure. Do I think the studies have actually measured what they purport to measure, in a ceteris paribus way, with a level of statistically significant precision? Absolutely not. I would be very hesitant to make such sweeping generalizations on such a shaky empirical basis.
Do you doubt that spatial ability is important? My son wouldn't use the term spatial ability, but he would tell you that his ability to look 8-9 moves into a position helps him win a lot of games.
It's reasonable to think spatial ability is important to chess performance. However, your argument logically relies on spatial skills being the prime determinant of ability. I don't think that assumption is reasonable. It sounds like you have one piece of data and are trying to fit the argument to the data.
For the final part, consider the reason that Malcom Gladwell gave as to why a disproportionate number of professional soccer players are born in January, February and March in his book Outliers. His explanation was that because these were the oldest kids in their class, they were more slightly more talented than the others. They received more coaching over time, which continually increased their skill levels over their slightly less capable players. The less skilled players dropped out as competition increased, leaving a disproportionate number of players born in the first three months of the year.
It could be rational that clubs will disproportionately apportion resources to the highest ability players.
I have to question the underlying goal of chess clubs, though. Is it to nurture talent at the very top, or to maximize performance across all members? I would think some combination of the two has the most value for both the highest ability players and society.
For the reasons I discussed above, I think chess clubs' reaction function is off. That is, I think they've mis-calibrated their target (boys, at the exclusion of girls) and are optimizing over the wrong set.
There is a benefit to society at large, I think, to having children accustomed to thinking strategically. Why game theoretic thinking isn't taught in school is beyond me. But then, I have a graduate degree in game theory.
