Tens of thousands of young people leave school with no or very few qualifications in England. This paper asks: what is the ethnic dimension of the low achievement problem? We focus on six aspects, using the National Pupil Database: whether the relation between ethnicity and incidence of low achievement is symmetrical with the relation between ethnicity and achievement; whether economic disadvantage is an equally large risk factor for low achievement among pupils of all ethnicities; whether the relationship of ethnicity with low achievement varies by peer group; whether the relationship between gender and chances of low achievement differs by ethnic group; whether the relation between ethnicity and low achievement changes with age; and lastly, whether there is sorting by ethnicity into schools of better or worse quality. We find that: (1) the relationship between low achievement and achievement is "not" symmetrical and, therefore, that it is valuable to measure low achievement directly rather than as the negative of achievement; (2) economic disadvantage has a stronger effect on White students than those of ethnic minorities; (3) the relationship between ethnicity and low achievement does vary with age; (4) ethnic peer-group effects are not significant; (5) the gender gap in the probability of low achievement is significantly larger for most ethnic groups compared with the White group, at least by one measure of low achievement; and (6) attending poorer-quality schools is part of the explanation of differential ethnic minority performance. We offer explanations for our findings and for the presence and/or persistence of inter-ethnic differences in the chances of low achievement. (Contains 15 tables, 1 figure and 3 notes.)