Letter: Reconviction: scientific methodology vs empirical experience

From Mr Len Salem

Sir: The only valid test for the usefulness of Professor Copas's formula is this: take, say, 100 case histories not included in the original survey sample and apply his method to predict the likelihood of re-offending. Give the same case histories to a selection of experienced probationers and ask them to do the same. Compare the predictions with what, in fact, happened and you will know which is the more reliable guide. It may well be that this elementary test has already been performed; if not, it should have been.

Yours sincerely,

Len Salem

London, W5