Scientific studies based on average results do have some validity, but what these studies are telling us is not as much as many people think.
Again, though, these studies looked far beyond the "average results" to extremely specialized and difficult math exams and competitions. I'm not understanding the criticism here. If someone has a more substantive comment, I'm interested in discussing it. FTR, although I am a writer, not a scientist, this is my field. I read hundreds of social and cognitive science studies a year for my job, so I'm pretty familiar with assessing the quality of a study.
This was more a general statement about areas of science where complex systems are involved. Not specifically a criticism. It is very difficult to explain why these types of studying can be lacking.
This particular study is not as narrowly focused as some studies I have seen. However, I would need to analyze the test questions on these various classes of tests to see how much variation there is in the complexity. How people view the complexity of testing is definitely a topic of debate.
I would say this particular study is definitely telling us further study is required. The information is interesting, but on it's own, it gives no indication as to why the results are what they are. At least the information tells us that the idea of males being better in math as found in other narrow studies may not be right.
Back in the 70s and 80s, we were constantly hearing how almost everything was found to cause cancer. The experiments were done using 2 control groups of rats. One group had none of the substance and the other group was given an extreme dose of the substance. Almost every substance tested was shown to cause cancer.
Later when much better science was used to test for the effects of cancer, many control groups were tested with each having gradually increasing doses of the substance. A graph where the incidences of cancer gradually increased as the dose increased were considered to be evidence the substance caused cancer.
In cases where no cancer appears to be caused by the substance until an extremely high dose is applied are considered not to be evidence the substance causes cancer. The graphs in these cases will suddenly spike at the high dosage levels. What this early poor science had discovered was that almost any substance can cause cancer at overdose levels. It was the interpretation of the findings which was flawed.
If I was to do a study to determine if one gender was better in math than another, I would first have to ensure the control groups were similar in almost every way other than gender. This study compares results between different cultures and this adds too many variables into the equation. You would also have to ensure as many factors as possible were the same between the male and females in the study. This is virtually impossible to guarantee.
A study specific to math ability would just have too many variables involved. Now if we can find the pattern of cognitive abilities across a wide variety of skills, this can remove a degree of error resulting from nurturing variations due to gender. This requires an understanding of how various skills are related to work, but I believe this is the minimum it would take to have anywhere near convincing evidence.
What I have been trying to do is come up with what I call the periodic table of cognitive ability. In chemistry, they had once tried to do chemistry using earth, wind, fire and water. The periodic table is what moved chemistry forward. I feel the same about analyzing human cognitive abilities using high level subjects like math and science. This is feel will not get us anywhere in understanding human cognition.
If you ask me who is better at math, I will tell you my answer is I don't know. If the question is which gender would statistically be faster at solving a Sudoku puzzle, I would take a guess it would be males. This is a specific enough skill, I have a fairly good idea of what neural mapping would suit this skill. Males on average have much more tightly wired neural mapping with greater localization of neural connectivity. For a very focused simple skill, this is probably ideal. For a more advanced skill, I wouldn't hazard to guess.
I'm hoping the cancer study info helps with understanding how the shape of the graph comes into play.
Don't take the cognitive stuff too seriously. Any statement I make regarding the brain are at best slightly educated guesses. I just don't want to have to say I think or speculate constantly.
