The issue we had--and we had a discussion on this at the committee--was that we had two numbers about the rate of visible minority hires coming into the public service. One number was 8.9%, which I thought was really a problem, and the other number was 17%, which is much, much better and gives a much better indication of what we were trying to do.
The reason we had these two numbers for the same time period is that they were collected differently. Our preoccupation was whether that was a function of the collection. To try to understand this, we asked what was going on there. Seventeen per cent is collected as part of our automated application system, which is a computerized system. People are forced to answer every screen. The other number, the self-identification number, is a number that is collected when somebody is hired, given a form, and asked to fill it in and send it back.
That whole process--that self-identification process in the departments we looked at and how that was managed--is really not managed very consistently or very well, so we know for sure that we have a problem with the collection of the data. That is not a very reliable number until we do a better job of collecting the information.