“the empirical evidence of parental witness” The problem is that this is witness in recollection and the diagnosis of autism is a slow one, and is also often associated with other problems. Autism isn’t a medical condition but a behaviour and it has many possible causes. Parents will often fix on one event and decide that this is an identifiable cause of all that happened afterwards. There have been illnesses and deaths caused by faulty vaccine batches but there is no evidence- diagnostic, except Wakefield’s, shown to be unreliable, or statistical- that this has happened with M.M.R.

One of the problems of medicine- mentioned by Goldacre- is the unreliability of human observation, including parents’, and human inability to understand science or scientific discoveries or their implications. For example, he has dealt with the problem of intuition in some of his problems, yet doctors will ascribe- to themselves at least- medically educated intuition which makes themselves better at diagnosis than others. With non-specific and ill-defined areas of medicine such as psychiatry the effects are disastrous. As Lewis Wolpert pointed out in The Unnatural Nature of Science science is counter-intuitive in the way it works an often in its revelations- an obvious and disastrous example is the fact- shown again and again- that rewarding good behaviour is much more effective than punishing bad behaviour as a means of getting the behaviour we want from people, that people will agree in the abstract and will revert to punishment in the family or society without noticing its ineffectuality.