Who are taking these tests? Unless it is an unbiased sample of the population, the results may not have any relation to those of the whole population. If, for e

Who are taking these tests? Unless it is an unbiased sample of the population, the results may not have any relation to those of the whole population. If, for example, people looking for a job, as opposed to those that already have jobs, are more likely to take the test, that would arguably lead to a depression of scores compared to the entire population. If working conditions in India are such that more developers take the tests, that would then skew the Indian scores higher than the American ones. Or maybe the other way around. Such a study would be very difficult to make unbiased, and without quantification of the biases the results are really meaningless.