John Read appears to feel we were less than generous in our paper to the hypothesised relationship between child sexual abuse and schizophrenia. One of the many frustrations which beset researchers is that having chosen an analytical method and set the level of significance, you just have to live with your results, equivocations around trends notwithstanding. You certainly cannot, as John Read does in his letter, state about our results that ‘males who suffer child sexual abuse were 1.3 times, and abused females 1.5 times, more likely to have been subsequently treated for schizophrenic disorders’, when those relative risks were non-significant. I can assure your readers that had we been able to squeeze out a significant association between schizophrenia and child abuse from our data we would have done so. After all, dramatic and unexpected results tend to acquire that coveted accolade of citation more frequently than do the mundane and predictable. As we hopefully made clear, our study did not exclude an association between schizophrenia and child abuse – how could it – but simply failed to support such an association.

Dr Read refers to the associations found in a number of studies between endorsing symptoms which can occur in psychotic disorders and prior sexual abuse. I would suggest this is not quite the same thing as associations with schizophrenic illness. It should also be emphasised that correlations do not necessarily reflect causal relationships even if you chose to describe them as ‘powerful relationships’.