White people tend to be nervous of raising the subject of race and education, but are often voluble on the issue if a black person brings it up. So when Trevor Phillips, chair man of Britain's Commission for Racial Equality, said that there was a particular problem with black boys' performance at school, and that it might be a good idea to educate them apart from other pupils, there was a torrent of comment. Some of it commended his proposal, and some criticized it, but none of it questioned its premise. Everybody accepts that black boys are a problem. On the face of it, it looks as though Mr. Phillips is right. Only 27% of Afro-Caribbean boys get five A-C grades at GCSE, the exams taken by 16-year-olds, compared with 47% of boys as a Whole and 44% of Afro-Caribbean girls. Since, in some subjects, candidates who score less than 50% get Cs, those who don't reach this threshold have picked up pretty little at school. Mr. Phillips's suggestion that black boys should be taught separately implies that ethnicity and gender explain their underachievement. Certainly, maleness seems to be a disadvantage at school. That's true for all ethnic groups: 57% of girls as a whole get five A-Cs, compared with 47% of boys. But it's not so clear that blackness is at the root of the problem. Among children as a whole, Afro-Caribbeans do indeed perform. badly. But Afro Caribbeans tend to be poor. So to get a better idea of whether race, rather than poverty, is the problem, one must control for economic status. The only way to do that, given the limits of British educational statistics, is to separate out the exam results of children who get free school meals: only the poor get free grub. Poor children's results tell a rather different story. Afro-Caribbeans still do remark ably badly, but whites are at the bottom of the pile. All ethnic minority groups do better than them. Even Bangladeshis, a pretty deprived lot, do twice as well as the natives in their exams Indians do better still. And absolute numbers of underperforming whites dwarf those of underperforming Afro-Caribbeans: last year, 131,393 of white boys failed to hit the government's benchmark, compared with 3,151 Afro-Caribbean boys. These figures suggest that, at school at least, black people's problem is not so much race as poverty. And they undermine the idea of teaching black boys separately, for if poor whites are doing worse than poor blacks, there's not much argument for singling out blacks for special measures: whites need help just as badly. According to the text, the public response to Mr. Philips' claim is