Faith schools which operate their own admissions get better results when they have competition, research suggests.
Competition could raise standards - and widen social divisions
An institute at the London School of Economics has published a study of school choice, using a sample of 200,000 pupils in England.
Faith schools - outside local authority control - were compared with community schools to measure the impact of parental choice.
Researchers say when a faith school has no competition, its results are lower.
Where a voluntary aided school had competition, that is, where parents were actively choosing it over others, they said the extra "value added" for pupils in such schools was equivalent to 16 to 19 weeks of progress.
However, the researchers said if there was no competition - that is no other schools nearby - the value added dropped by the same amount.
The research also attacks any assumptions about inner-city educational disadvantage - with findings that densely-populated, urban areas with many competing schools have higher standards.
The research from the LSE's Centre for Economic Performance, carried out by Stephen Gibbons, Stephen Machin and Olmo Silva, examines the impact of parental choice on schools, using a sample of 2,400 schools in south-east England.
Brighton: Parents were divided over changes to school admissions
It found that where schools were in "competition" - either because there were many schools close together, or where a faith school was operating outside the local authority system - it could have a positive impact on standards.
Allowing schools to run their own admissions systems was recently criticised by an IPPR think-tank report, which argued that it was used as a form of selection.
But the LSE research says that where schools run their admissions process - and there are other competing schools where pupils could be sent - "there is evidence that competition is positively related to performance".
The research is published as the ATL teachers' union attacks faith schools for a lack of accountability - and calls for them not to "discriminate" in their admissions process.
Conservative leader David Cameron and Shadow Chancellor George Osborne have recently said they intend to send their children to faith schools - a choice also taken by Prime Minister Tony Blair.
The study is an attempt to cast light on the consequences of parental choice for schools - and the researchers say that, despite many strong opinions, there has been little clear evidence.
Inner-city education was found to be more effective
The study says that trying to find evidence of the impact of choice was difficult because the system in England counter-balances choice and competition.
If a school is very popular with parents and becomes oversubscribed, children are still allocated by the local authority to less popular schools, "preventing competitive incentives from operating".
However, the example of schools which control their own admissions, outside of local authorities, which are usually faith schools, provided the researchers with evidence for how parental choice could affect standards.
These schools were actively chosen by parents, rather than places being allocated by local authorities.
Within this sector, the researchers found a positive impact, with non-selective primary schools showing increased "value added" measurements - in terms of how much pupils improve between the ages of seven and 11.
The study also found that, counter to any perceptions about inner-city schools providing a "poor education", densely-populated, urban areas were more successful in terms of value-added.
Among such secondary schools, the researchers suggested that stronger pupil performance was a reflection of "greater inter-school competition".
"Pupils perform slightly better when they are at school in places that are highly urbanised and, particularly importantly, where there are many other neighbouring schools."
The implications of school choice remain unclear, say researchers
However the research also emphasises the ambiguity surrounding the consequences of allowing or promoting school choice.
In the primary schools where schools ran their own admissions the "diversity of abilities" within the school became slightly more narrow. This might make it more attractive to parents looking for a good school, but it raises questions about "segregation" by ability and social class.
The researchers highlight the questions that are raised by parental expectations of school choice.
Breaking the link between where people live and where they go to school is more socially inclusive in terms of allowing families in deprived areas to have access to popular schools outside their area.
But if such an open market were to be introduced, such an extension of choice could give even greater advantages to the most mobile and ambitious middle-class families.
The research also shows that despite the attention drawn to admissions disputes, such as the introduction of a lottery system in Brighton, that for many families the practical choices are much more limited.
Very few children have more than three schools which they could realistically attend, say researchers - and one in four only have one school within a reasonable travel distance.