I would like to reiterate what was said by Mr. Murdoch and Mr. Beaud, namely that the problem is not so much the size of the sample as it is its representativeness. Whether 20%, 30% or 10%, it may work, but what is really important is that completion be mandatory for the sample group. We want to have a good representative sample.
What is at stake here is not really the percentage, the total number or the size of the sample, but the guarantee of the sample being representative. In addition, as we mentioned already—and it is a fact that is misunderstood—many surveys, projects and polls are carried out first, and then their data are corrected using the Canadian census data.
If we undermine the validity of the Canadian census, we indirectly invalidate much of the other data available.