How Do U.S., Indian Developers Really Match Up?

Based only on my experience with outsourced work, this:

"That the gap has narrowed, Desai said, is attributable simply to Indian programmers having gained more experience"

becomes

"That the gap has narrowed, Desai said, is attributable simply to Indian programmers rampantly cheating"

but I could just be incredibly bitter from bad experiences.

edit: I can't decide if this comment is coming across as racist or not, but I just can't put any faith in the Indian education system anymore. Seeing a masters degree in computer science on a resume has never failed to disappoint.

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.

These "tests" are hardly a measure of competence. A good quantitative score to measure competence should be measured over a longer period of time.

There are good programmers in India. But the signal to noise ratio is very high. IMO, this is attributed to an over abundance of engineering colleges with poor faculty quality.