Popular schools are "skimming the cream" of applicants to boost their league-table rankings, according to research.
Over-subscribed voluntary-aided and foundation schools, which control their own admissions, are able to select pupils who will do well in exams.
In a study of 2,862 secondary schools, Anne West and Dabney Ingram, of the London School of Economics' centre for educational research, found that 96 per cent used admissions criteria relating to siblings. Distance was used by 86 per cent, socialmedical need by 73 per cent and catchment area by 61 per cent.
However, schools used a wide range of criteria to decide who to admit if there were more applicants than places. Some of these have "the potential to be used to enable schools to select students who are likely to be easier to teach", said the researchers.
These criteria include aptitude or ability in a subject, used by 3 per cent of schools overall. Others used student interview (2 per cent) or parental interview (2 per cent).
However, voluntary-aided and foundation schools were almost 30 times more likely to admit gifted children when faced with difficult decisions over admissions - 9 per cent compared with 0.3 per cent of community schools.
The study also found evidence that "some schools used interviews to establish that the attitudes and values of the students parents were in accord with those of the school, so allowing certain students to be 'selected in' and certain to be 'selected out'."
Foundation schools - the former grant-maintained sector - were also less likely to consider a child's medical or social needs - 52 per cent, compared with 80 per cent of community secondaries.
Only 15 per cent of voluntary-aided and foundation schools took account of special needs, compared with 48 per cent of those run by local authorities.
Researchers also found some schools were using criteria which could contravene legislation such as the 1976 Race Relations Act. While only 5 per cent of community schools chose the children of employees, the figure for voluntary aided or foundation schools was 20 per cent.
The report said that in some schools "admissions criteria appear to be used in such a way that students admitted to the school are higher attaining and thus likely to perform well in public exams, so enhancing the school's league-table position."
The study concluded that all admissions should be run by a "body with no vested interest in the outcome such as the local education authority". This would ensure that one group of students did not benefit over another.