The statistics are likely to embarrass ministers at a time when the Government is planning to extend selection in England and Wales.
In Northern Ireland primary schools where fewer than 11 per cent of pupils are entitled to free meals, 52 per cent gained grade A last year, putting them in the top quarter of the ability range. Only 16 per cent of pupils got top grades in primaries where more than half were entitled to free dinners.
Moreover, the department says that the gap between the most and least deprived schools has widened since the switch from verbal reasoning to tests of attainment in the new curriculum three years ago.
However, it admitted this week that it had no plans to change the system though it will be under strong pressure to do so because of the impact on Catholic children, who are more affected by poverty than Protestants. In the four final years of verbal reasoning tests, roughly equal proportions of Catholic and Protestant 11-year-olds got grade A. Since 199394, when tests in English, mathematics and science were introduced, Protestants have opened up a lead of 7-8 percentage points, gaining 1,713 more top grades than Catholics in the past three years, reversing the pattern in the previous three years.
Despite this, Catholic schools are performing well in the selection tests when like is compared with like. In schools where 41-50 per cent of children are eligible for free meals, one in three Catholics gained the top grade, compared with only one Protestant in five. One conclusion is that deprived Protestant children are doing very badly. In the poorest Protestant schools, only 11 per cent are getting grade A compared with 52 per cent in affluent schools.
Dr Tony Gallagher, from Queen's University's school of education, said he and others had predicted at the time that the new tests would hit working-class pupils and thus Catholics. "Verbal reasoning tests have a social class effect, though they are based on the belief that you can measure what pupils are capable of. But all research shows that attainment tests have an even greater social class impact, and these figures provide concrete evidence," he said.
Although the department has said it does not intend to change the format of the tests, it will fine-tune them. "Schools have generally welcomed the attainment tests because they reduce the distortion of the primary curriculum, " said a spokesman for the department.
"We will try to improve the lot of these poorer schools through targeting social need and raising school standards. We have also given every Primary 1 class a teaching assistant, and we shifted about Pounds 6 million from secondary to primary schools in the transitional years of local management of schools."
Teacher unions were unimpressed by the response. "It always was an uneven playing field and now it is even more uneven," said David Allen, general secretary of the Ulster Teachers' Union.
"We have been tinkering with selection since 1948. It is now time for the Government to take the bull by the horns. There is no fair mechanism by which children can be transferred in a selective system at 11. If the Government cannot go the whole way to abolishing selection, at least it should defer it until 14."
The Irish National Teachers' Organisation repeated its demand that selection should be scrapped and the director of the Council for Catholic Maintained Schools said the tests should be looked at again to make them more suitable for socially deprived children.