Occasionally a new piece of research demolishes a myth with one fell blow. It does not happen often (social research tends to run along familiar tracks), but once in a while an iconoclastic study changes ideas.

No one reading Self-Esteem - The Costs and Causes of Low Self-Worth by Professor Nicholas Emler of the LSE, should feel quite at ease again using a modern piece of psychobabble that has infused the language of sociology, criminology and education without real scrutiny until now. The accepted view has been that self-esteem - or the lack of it - lies at the root of almost every disorder from delinquency and drug abuse to violence and child abuse. One standard text after another takes this as a given fact without any scientific evidence, repeated as gospel from right to left, from Melanie Phillips to Oprah Winfrey.

More than 2,000 books currently in print offer self-help prescriptions for raising self-esteem. A vast array of expensive social programmes in Europe and the US designed to solve drug dependency or delinquency are based on attempts to raise self-esteem. Some have tried to raise the self-esteem of whole schools or even an entire citizenry, describing self-esteem as a "social vaccine" against anti-social behaviour. Low self-esteem is the zeitgeist social disease. It has many useful attributes: it elevates self-love and sanctifies self-satisfaction. It justifies the introspection of the therapy addict. It excuses bad behaviour, turning perpetrator into victim. For teachers, it makes dealing with bullying, arrogant and disruptive pupils almost impossible, if beneath the insufferable exterior there is supposed to be a whimpering, self-loathing child in need of affirmation and praise.

Professor Emler turns all this on its head. Scrutinising all the available research on both sides of the Atlantic, he finds no evidence that low self-esteem causes anti-social behaviour. Quite the reverse. Those who think highly of themselves are the ones most prone to violence and most likely to take risks, believing themselves invulnerable. They are more likely to commit crimes, drive dangerously, risk their health with drugs and alcohol. Exceptionally low self-esteem is indeed damaging - but only to the victim, not to anyone else.

Those with low self-esteem are more likely to commit suicide, to be depressed, to become victims of bullying, domestic violence, loneliness and social ostracism. There ought to be a collective sigh of relief among many professionals on reading this eye-opening work. It is one of those moments when the blindingly obvious suddenly emerges from a fog of unquestioned nonsense. Teachers, social workers and probation officers do not have to massage the already inflated egos of bullies with unwarranted praise. Asserting his own superiority over his classmates, over-confident of abilities he does not have, it will do no harm to try to bring him down a peg.

Emler looks at the relation between self-esteem and academic success. Does competition in school cause damaging failure? Most surprisingly he concludes that academic success or failure has very little impact on pupils' self-esteem. High self-esteem pupils will explain away failure to suit their previous high opinions of themselves: they make excuses that they were unlucky, suffered some bias or that they didn't try. Odder still, those with low self-esteem will not be buoyed up by academic success either. Sadly, they will regard it as a fluke and continue with their previous low estimation of their abilities. He concludes that it is exceedingly difficult to shift people's pre-existing view of themselves, even with tangible success.

Nor is self- esteem any predictor of how well or badly someone will do academically. Even if confidence boosting worked (which he doubts) it would have no effect on exam results. So where does self-esteem come from? Looking at studies of twins, Emler concludes that genetic predisposition has the single strongest effect. Less surprisingly, after that it is parental attitudes. If they love, reinforce, praise and respect a young child, the effect lasts for life. Physical and above all sexual abuse of children is devastatingly and permanently damaging to self-esteem. Beyond these early influences, everything else that might be done to increase/ decrease self-esteem has virtually no effect. (This is bad news for the therapy business.)

An interesting example: it was assumed that to belong to an outcast ethnic minority would harm self-esteem, but Emler finds it has no effect. People draw self-esteem from the good opinions of their own group and reject abuse from outsiders as the fault of others, not their own. Men have slightly more self-esteem than women. Low self-esteem in young women does increase the risk of teenage pregnancy, while low self-esteem in boys increases the risk of unemployment later in life. Anxiety about appearance does undermine women's self-esteem.

But Emler's more curious finding is that there is very little correlation between how people think they look and how they actually look: their perceptions about their appearance are shaped by their level of self-esteem. Altogether Emler finds people have profoundly unrealistic views of how others see them, both negative and positive. How we think we are perceived is shaped by self-esteem. His conclusion is that all the myriad programmes designed to cure anti-social behaviour by raising self-esteem are wasting their time. Better by far to concentrate on the particular drug or crime problem and not on an imagined self-esteem deficit: self-esteem enhancing programmes he describes as "snake-oil remedies".

This research deserves to cause a stir. It was always a kindly liberal notion that inside the anti-social bully was a timorous, tender soul waiting to be released. Emler is not suggesting that the violent are not damaged or might not be cured, but he has conclusively dismissed the intellectually woolly concept that lack of "self-esteem" is the root of all evil. · Self-Esteem by Professor Nicholas Emler (Joseph Rowntree Foundation). p.toynbee@guardian.co.uk