Statistics Canada has always tried to search for a compromise with respect to the size of samples. You know just as well as I do that the size of the sample is directly associated to the total cost of the survey. The objectives of the survey are taken into consideration, but the number of surveys carried out by Statistics Canada is huge. As stipulated in section 41 of the Official Languages Act, Statistics Canada did everything to ensure that when it comes to key surveys, I mean surveys which represent important stakes and a marked interest for official language minorities, the department should attempt to oversample.
Obviously, there is a cost, which is usually quite high, associated to this. We carry out surveys and sampling, we try to adhere to proportion of subpopulations in each province. Yet, external clients, other than the federal government, are usually the ones who fund surveys carried out by Statistics Canada. Our responsibility is to tell them that a given subject may be of interest for official language minorities, for example and ask them if they could possibly carry out an oversampling.
From time to time we are able to find people who are willing to fund the oversampling. That is what we were able to do for the study on adult literacy. We were able to carry out oversampling of francophones in four provinces in addition to the anglophones living in Quebec. We carried out a study on the vitality of official language minorities, specifically because it is often very difficult to make accurate representation of each francophone community within the various provinces. The surveys on health provide a very clear example.
With respect to the survey on health care in communities, a sample of 133 000 respondents allowed us to make estimates concerning official language minorities for each province, which was certainly not the case when we talk about immigrants. We know that the percentage of immigrants is very low. We would also have to carry out oversampling in the case of immigrants.
In fact, a longitudinal study on Canadian immigrants had initially involved 12,000 people. We are constantly having to raise awareness among fundraisers and explain to them of a possible need for oversampling in a survey dealing with a subject of national interest. Obviously, that brings in additional costs.
The rationale for the post-census survey was precisely to compensate for the situation. On the health funds, I believe that we are going to obtain some very interesting results. In fact, I was saying earlier, the census is underused insofar as issues dealing with immigration. Regardless, there is reason to plan for surveys that would take into consideration stronger representation of French-speaking immigrants within the sampling.