Considers the relative effectiveness of using experimental training programmes. Uses the US experience of programme evaluation to make two main points. First, many of the key questions concerning the effects of training cannot be answered using simple estimators, even in the presence of...