[Lancey Howard]

I wouldn't interpret it that way at all. First, performing well on a standardized test isn't a reliable indicator of intelligence. All it means is that you did well on the test.

The problem with your response is that the graph doesn't show the results of just one person. When large numbers of people of the same ethnic/racial backgrounds produce a graph like the one shown, the results are more likely to produce reliable data.

[HitmanLV]

I didn't suggest the data wasn't reliable, I just said some conclusions could be more dubious than others, and offered some other conclusions.

It's reasonable to conclude from that graph, for example, that asians and whites spend more time, money, and resources to preparing for the exams rather than hispanics or blacks.

Every test just measures what is being tested. Any test with a consistent methodology can be prepped for. If a test has a notable bias, that's a good thing: train to handle the bias. If you can isolate it, that goes a long way towards cracking it.

It helps when you think of these things in 3-D, not 2-D. Or heaven forbid, 1-D.