In my opinion, and as noted by my colleague, that will definitely be the case. If all the characteristics of the population were evenly distributed across Canada, there would not be many problems. A 10% sample would then be sufficient. But this is not the case. Some features cluster in sub-populations difficult to identify. We are talking about immigrants, Aboriginals, poverty and perhaps all those combined. We are perhaps even talking about sub-populations that have not yet been identified and will never be, but that would need the support of the Canadian government.
It is not really the size of the sample that is a problem; it is the nature of the population. To get back to your question, I would say it is absolutely certain that, scientifically, we are going to invalidate the results that could make the survey scientifically valid.