Research has always been susceptible to fraud, but regulations are now much tighter than they were, says David Wootton