Intelligence tests of various kinds have been conducted in many places, on many different kinds of people, under many different kinds of circumstances. Certain tendencies keep showing up in the scores between members of different races. All races have a wide distribution with more people around the average than at either extreme. But they don't have the same average, and the extremes aren't equally common. Essentially, if you make a curve graph showing any given race's distribution of test scores, the entire curve is shifted to the left or right between different races. Most of any graph still overlaps with most of any other graph (meaning there are plenty of people above and below their own races' averages who are also above or below another race's average or above/below most of some other race's graph), but the differences are still substantial.

So he's not talking about any one particular test; he's talking about a vast pattern that turns up in years of testing in large numbers. Generally, black people's averages on various tests in various places run about a standard deviation (the equivalent of 15 IQ points) lower than white people's on the same test. Sometimes it's less, and sometimes it's more, but it's always there. White people's average also often runs a few points lower than eastern Asian people's average. Sometimes that difference doesn't show up at all, and sometimes it's as high as the equivalent 6 IQ points, but 2-4 is more common. It also has the pattern of showing up more strongly in mathematical reasoning than in linguistic ability. (African races other than black and caucasian haven't been tested enough for such comparisons, and if any such data are available on Australian aborigines and their relatives, I haven't seen them.)

This is not a disputed question among psychometricians; it's a fact. The issue is how to explain it. These test score differences are not only ubiquitous but also very stubborn. The problem has been that every single offered explanation has failed. They can be checked for, and they've just not panned out.

For example, you could say it obviously must be because white Americans are rich and black Americans aren't, but the test score difference is still there even if you're only comparing white and black people of equal wealth and income. You could say it's due to educational differences, but the test score differences don't go away when you're comparing students at the same school. You could say it's due to test bias, but there are ways to confirm or refute test bias and its importance to an outcome, and that explanation doesn't work either. You could say it's due to a psychological effect from oppression or history of slavery, but the test score differences still show up in other countries with completely different histories, including countries where black people are the "ruling class". (In fact, it's bigger in some such countries, around a standard deviation and a half.) And some of the attempted explanations in terms of environmental disadvantages in present American society also would have applied to eastern Asians in the USA, so why are the results the opposite? (And why would the differences between people from eastern and western Eurasia also be about the same in all Occidental and Oriental countries, never more than a few points bigger or smaller anywhere?) Referring to African countries in particular, someone here said something about people in such a culture not being trained for written tests like we are in the USA, but the racial test score differences don't just show up between countries; they also show up among the scores of people from the same country who would be equally culturally unprepared.

So whatever the explanation is, it's either biological, or an environmental issue that nobody has yet guessed, or both, because the environmental explanations that have been tried so far just aren't it, no matter how many times they get repeated. And yes, that goes for the ones right here in this thread.

Also, in addition to the fact that it's been known for centuries that the races' skulls aren't shaped quite the same, it's now known from genetics that a few certain alleles for brain development & function have been strongly selected for, spreading throughout populations much too quickly for genetic drift to account for, which means the alleles must have real-world effects that natural selection could act on... and one such allele is found primarily in black people, while the others are found primarily or exclusively in Eurasians. So the existence of at least a few racial differences in brains is also, like the stubbornness of the test scores, not disputable; only the exact nature of those differences is unknown. (The alleles' functions are known in a general sense of "something about how the brain grows and works", but not with much specificity.)

There are also some reasons to doubt a biological explantion for the test scores, though, or at least to doubt that that's the ONLY reason, even if one can't imagine what environmental ones are still left to examine. For example, Latinos in the USA don't do as well as non-Latino whites, even after adjusting to eliminate the possibility of language being the explanation. But racially, American Latinos generally ARE either white or a mix of white and Amerind, the latter of which is just a splinter group from eastern Asia and doesn't have substantially lower scores overall than Latinos do, so a biological explanation can't fit these various facts at the same time. Also, white European Jews score higher than white European gentiles, even though other Jews elsewhere don't.

On Watson's idea that low intelligence is generally predictive of a bad future for Africa's various social problems, though... it doesn't really matter whether he's talking about just the black ones or everyone else in Africa too, or whether the low intelligence in those countries results from one cause or another. Its effect on African countries' ability to fix their problems is the same regardless of the low intelligence's cause and the distribution. The only real question for what he's talking about is how much of an effect a society's overall intelligence distribution has on its well-being and ability to overcome dire challenges.