OECD study shows UK has biggest science gender gap of any major developed country.

Self-confidence makes a bigger difference to people's likelihood of success in Britain than in almost any other country, a major international study suggests.

Analysis by the Organisation for Economic Co-operation and Development (OECD) shows that teenagers with a stronger belief in their own abilities score more highly in science and maths tests than those who doubt themselves.

Although the pattern can be seen across the developed world, the effect is noticeably more marked in the UK and New Zealand than elsewhere.

The study also shows that British schoolgirls are lagging further behind their male classmates in science than those in any other major developed country.

The study, based on results from international academic tests found that while girls work harder, are more academically focussed and outperform boys overall in the classroom across the world, they are lagging behind in maths and science in many countries.

In the UK boys achieved almost 13 points more on average in the science tests than girls - the equivalent of around three months extra education.

Only Colombia, Lichtenstein and Luxembourg had a bigger male-female advantage in science.

British boys also outperformed girls by a similar amount in maths, although the gender gap was even greater in a handful of other comparable countries including Germany, Ireland and Spain.

Significantly, the study also measured the effect of self-confidence on pupils' test scores and found that boys were more confident about their abilities in maths and science tasks than girls.

Although it was shown to boost results significantly in most countries, the effect was much more pronounced in the UK and New Zealand.

"What emerges from these analyses is particularly worrying," the report warns.

"Even many high-achieving girls have low levels of confidence in their ability to solve science and mathematics problems and express high levels of anxiety towards mathematics.

"Results indicate that even among boys and girls who are equally capable in mathematics and science, girls tend to report lower levels of subject-specific self-efficacy and self-concept."

The findings come from a major new analysis of tests scores from the Programme for International Student Assessment (Pisa), which compares standards achieved in maths, science and reading tests by more than 500,000 pupils aged 15 in more than 60 countries.

Overall it showed that boys are more likely to underachieve than girls, making up 60 per cent of those who effectively failed the tests, and lag behind them in reading skills across the board.

Boys spend an average of an hour less per week doing homework and are also less likely to read for pleasure and more likely to spend their free time playing video games, it notes.

But when it came to maths and science problems - particularly tasks in which they had to apply their skills and "think like scientists" in more abstract situations, boys led the way.

The study also noted evidence that boys appeared to be better prepared for the world of work and job hunting than girls despite being less academically successful on average.

Yet in a handful of Asian countries girls performed in a par with boys in maths, casting into doubt the assumption that the disparity is simply down to innate differences between the sexes.

The biggest male-female science gap was in Columbia where boys scored 17.6 points more in the tests than girls on average, just ahead of Lichtenstein and Luxembourg.

British boys scored 12.7 points more than girls, more than twice the difference in Shanghai and more than three times that in Ireland, the Netherlands and Canada.

British boys also enjoyed a 12.5 point advantage in the maths tests, meaning that the UK's maths gender gap is greater than in two thirds of the countries in the study.

Teenagers were also asked a series of questions to measure their "self-efficacy", meaning the extent to which they believe in their own ability to solve specific maths and science problems. Each was then given a self-efficacy score which was compared with their actual test results.

On average for every one point their confidence score increased, their science test results rose by 37.2 marks.

"In general, girls have less self-confidence than boys in their ability to solve mathematics or science problems," the report notes.

"Girls - even high-achieving girls - are also more likely to express strong feelings of anxiety towards mathematics.

"However, when comparing boys and girls who reported similar levels of self-confidence in mathematics and of anxiety towards mathematics, the gender gap in performance disappears."

Caroline Jordan, president-elect of the Girls' Schools Association and headmistress of Headington School in Oxford, said: "It might be that there are differences between the way girls and boys are wired but we are talking about a study that looks across the world so it is no good saying girls are girls and boys are boys.

"We have to ask what are we doing culturally that makes a negative difference in the UK.

"The findings on confidence are interesting, girls for some reason do not feel as confident about science, girls are often very good at making abstract links so what is it particularly about science?"

Mrs Jordan, who teaches physics, said one apparent advantage of single-sex schooling is that girls are less likely to consider some topics such as engineering a male domain.

"It could be down to the fact that we are training our children from a young age about the choices that they should make," she explained.

"I don't think that girls are necessarily less confident, it is more about science than generally.