Mr. John McKeon:

The first criterion is that a person has to be long-term unemployed. Then there are differences as to whether people have been unemployed for one to two years or two to three years. We try to get a representative sample of durations. While it is random, it is a stratified random sample. We refer people to the various employment services based on the capacity of those services. We try to get a caseworker average ratio of approximately 125:1. We know the number of staff in the local employment service and we are trying to maintain this. We do not want to over-refer people so we end up with a caseworker in the local employment service trying to manage 1,000 people.