Only one in every 20 road accidents is caused by the driver breaking the speed limit, the Government has admitted.
Critics said the research, based on official police reports, destroyed the case for speed cameras.
It came amid a separate row over the accuracy of the Government's figures for the number of people being injured or killed on the roads, which showed a three per cent fall last year.
Critics say the figures, compiled by police, do not tally with the higher toll of hospital admissions.
Ministers have justified the explosion in speed cameras - which raised £120m last year by trapping two million motorists - by insisting that 'speed kills'.
But the first published study into the causes behind road accidents yesterday found exceeding the speed limit was a factor in only five per cent of cases.
Accidents were far more likely to be caused by drivers failing to pay attention and making mistakes, or pedestrians not looking properly.
The Department for Transport attempted to inflate the figure for speeding, claiming it was a factor in 15 per cent of accidents.
But this was only achieved by adding on the ten per cent of accidents in which drivers were considered by police to be driving too fast for the conditions - but, crucially, were not breaking the speed limit.
The Safe Speed campaign group said it was now clear the Government's 'entire road safety policy has been based on dodgy data'. It has seen the number of cameras rocket to 3,300 fixed sites and 3,400 mobile devices.
Founder Paul Smith said: 'Safe Speed has been pointing out for years that the concentration on speeding was a deadly mistake. 'Speed cameras must be scrapped.
'They have focused everyone on the wrong safety factor and have proved to be a dangerous distraction. Countless opportunities for live-saving policies have been missed - because of speed cameras.'
The 'contributory factors to road accidents report', the first of its kind, showed a total of 147,509 accidents last year.
Of these five per cent, or 7,314, had breaking the speed limit as a factor. This rose to 12 per cent, or 325, for accidents which ended in a fatality.
The police deciding the driver was going too fast for the conditions was a factor in a further 10 per cent of all accidents, or 15,436, and in 14 per cent of those ending in a death, or 357.
More significant were people failing to look properly (32 per cent), drivers of pedestrians failing to judge the other person's speed properly and road conditions or layout (15 per cent).
Separate figures showed the number of people killed in accidents involving drink-driving fell by three per cent last year to 560.
A total of 3,201 people were killed overall on the roads, down one per cent. And the number of people seriously injured fell 7 per cent to 28,954.
Total casualties - deaths, serious injuries and slight injuries - fell three per cent to 271,017.
Cyclists killed or seriously injured rose by 2 per cent overall to 2,360, and the number of fatalities increased 10 per cent - from 134 to 148.
The figures appeared to put the Government on course to meet its target of reducing, by 2010, by 40 per cent the number of people killed or seriously injured in road accidents compared with the average for 1994-98.
It is also aiming for other targets on the number of children killed or seriously injured - down 11 per cent last year to 3,742 - and a cut in slight injuries.
The DfT said the number of people killed or seriously injured in accidents reported to the police was already close to or below all the 1994-98 averages.
But the independent Statistics Commission said that if it based its casualty figures on hospital statistics, rather than police ones, the Dft would 'not be on track' to meet its targets.
Figures from hospitals suggest the true rate of death and serious injury could be half as many again as the number claimed by the Government.
Shadow transport secretary Chris Grayling said: 'The apparent contradiction between Government figures on road deaths and injuries and the story coming out of our hospitals is extremely worrying.
'The Government should start taking proper action to deal with the problem, and stop sheltering behind statistics about which there are very real concerns.'
Yesterday's figures also revealed 597 children were killed on journeys to and from school last year - the equivalent of 11 every week.
Of these, 471 were on foot, 56 on bicycles and 34 were in cars. Six out of ten were boys. The child pedestrian death rate is much higher than that of many other EU countries.