Tony Clement shows very worrying incompetence

In his interview earlier today with Evan Solomon on Power and Politics, Tony Clement showed two worrying signs of incompetence when attempting to establish the credibility of his new voluntary National Household Survey.

Contrasting the government's decision to scrap the mandatory long form census with the recommendation of StatsCan experts to keep the status quo, Clement offered what he seemed to think was an explanation showing that the new voluntary survey would be a statistical equivalent to the former mandatory long form census.

"We knew, Clement said, that the response rate would be lower, that's why we doubled the sample size. So the evidence is there that we're making it clear that of course you're going to have a lower response rate from a voluntary form versus a mandatory form. That's why you double the sample size and that's why even if you go down to 50% you have a fairly large sample to draw from."

The first issue in this comment is that Clement does not know his own plan. The former mandatory long form census was sent to 20% of households. The new voluntary form is proposed to be sent to 30% of households. Last time I checked 20 times two did not equal 30. For a minister that whose portfolio encompasses responsibility for Statistics Canada, this is worrying.

Secondly, Tony Clement seems to be implying that the by increasing the sample size, you're doing away with the main problem of a voluntary survey. While it is true that this might mean a similar final sample size, due to lower rates of participation on a voluntary survey, it does not mean that the two options would be comparable. The quality of the two different samples is not the same just because they are the same size.

With a voluntary survey, the sample is no longer as random. In fact, it becomes somewhat self-selecting. Certain groups will be less willing, due to factors such as busyness, to participate in a voluntary survey. So the sample that Clement's new survey will yield will not be as random or varied. Therefore the results will be much less of an indicator of the life of an average Canadian. Increasing the sample size of a voluntary survey does nothing to prevent this.

Tony Clement and the rest of the government have shown contempt for government based on evidence and information. The decision to scrap the mandatory long form census further showed that. Now they are either proving their incompetence, or believing that Canadians cannot see through such specious arguments.

Recommend this Post at Progressive Bloggers If you liked this post, please vote for my blog at Canadian Blogosphere