Industry Committee on July 27th, 2010

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

Thank you, Mr. Chair.

I want to thank our witnesses for coming today, a beautiful summer's day, and for giving it up to be here. Obviously this is a very important issue.

I'd like to start with Mr. McLeish on a question regarding the Statistical Society of Canada. From what I understand, you have members who work in various areas of statistics gathering and reporting. Who exactly are the members of your society? Maybe we'll start off with that. Can you give us just a brief overview?

I realize that we don't want to take up too much time, but this way we'll establish where you're coming from.

[Professor Don McLeish, President, Statistical Society of Canada]

We have about a thousand-and-some members, including a large number of academic statisticians from universities across the country, students, practising statisticians--a number from Statistics Canada, of course, and provincial statistical agencies--and statisticians who work in biostatistics and for private firms.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

Very good. You have a lot of credibility and a lot of people there.

This seems to have just shown up out of nowhere. Were any of your members, or were you as a society, consulted before any announcements were made?

[Prof. Don McLeish]

No, we were not consulted--but then, I would have been a bit surprised if we were.

I think what we might have expected is that statisticians, in one or other of the various related agencies, would be fairly thoroughly consulted on this issue.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

Are you aware of any of your members who were thoroughly consulted?

[Prof. Don McLeish]

No.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

No. So it's out there, and....

The long-form census is something we've heard a lot of information about over the last little while, but especially today. What do you believe would be the effect of making this a voluntary item as opposed to enforcing that people actually fill it out?

[Prof. Don McLeish]

Well, in general, the presence of bias in voluntary surveys is well known and well documented. I think any statistics society anywhere would support that, including the American Statistical Association and the Royal Statistical Society. The real question is how much bias is present in each of the responses. It can be quite pernicious in some cases; it can be less severe in others. So the correct answer to the question is that it's a complete unknown.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

So there's no way of knowing what we'll end up with down the road.

[Prof. Don McLeish]

No, not without a carefully conducted pilot, which could be done provided the 2011 census went ahead in its standard, traditional mandatory form. A pilot could be conducted that would at least try to investigate the degree of bias associated with a voluntary survey. But we know that bias will be present.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

A plan should have been though out, then. What you're saying is that a plan should have been well thought out as to what the implementation would be for voluntary. Would you have a parallel survey? Is that the way you would run it, with the voluntary and the mandatory going through similar tracks so that you could compare the outcome of the two?

[Prof. Don McLeish]

Well, I'm not referring to what should have been done.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

What would you normally do with your...?

[Prof. Don McLeish]

You have a copy, I think, of a statement we made. We're suggesting, among other things, that the mandatory long form be reinstated for 2011 and that Statistics Canada, with the advice of the National Statistics Council, be directed to undertake studies to show the impact of a voluntary survey over a mandatory one. That's essentially what I'm saying. I'm looking forward, not back.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

Very good.

A lot of the decisions that are made, whether they're on business, science, or social programs, rely on the accuracy of the long-form census. When making decisions, whether they're provinces making policy, municipalities, or employers, there's a bias there. How does this affect them? Is it an unknown? What exactly would it mean to someone who's trying to come up with a decision when they're given information that really doesn't reflect what happened in the past? Where do they go with this? What do they do? Or do they just throw it aside and...?

[Prof. Don McLeish]

That's entirely up to them. Clearly if there's an undercount in a certain ethnic, social, or economic group and we don't know the degree of that undercount, there isn't a great deal they can do except perhaps run their own attempt at estimating those quantities.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

That's where I'm going with this, or trying to figure out. We have groups that are represented, as it is now, and by making it mandatory everyone has to answer the questionnaire so everybody is represented. What groups will be more represented and what groups will be less represented on something like this?

[Prof. Don McLeish]

That's part of the difficulty. We really don't know. It's a combination of factors, obviously. I have a daughter with three children under the age of 2, all in diapers. There are days when she can't brush her teeth. She has a neighbour with one daughter who is nine years old. Which of those two, given the voluntary survey, is more likely to be inclined to fill it out? I think there will be a clear bias in favour of people who have time, who are motivated, who see some benefit to the data coming from this, and who are perhaps civic-minded and law-abiding citizens.

Unfortunately, governments at all levels make policy decisions not just for the law-abiding citizens who drive under the speed limit and don't litter but for all of society, including those who exceed the speed limit.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

Would it be fair to say that those who need it most will likely be underrepresented?

[Prof. Don McLeish]

I don't want to speculate too much on who's undercounted. We do know that people with language difficulties and the poor will probably be undercounted. But the only way of knowing for sure is to do a properly designed study comparing the two.

[Anthony Rota, Liberal, Nipissing—Timiskaming, ON]

So from what I'm getting, then, or from what I'm understanding, it wouldn't be unfair to assume that the decision to get rid of the long-form census was rather rash and not well thought out, and doesn't lead to good policy-making in Canada.

[Prof. Don McLeish]

No, I'm not going to say that. What I'm going to say, on behalf of statisticians in Canada, the U.S., the United Kingdom, Australia, New Zealand, and every other place where a census is essentially mandatory, is that I would like the government to go forward with the compromise position that the National Statistics Council has outlined. It's very much in parallel to what the Statistical Society has provided you with today. It provides for reinstatement of the 2011 census and, simultaneously, a study of the impact of a voluntary survey for any bias that might--and I think very likely will--creep into the responses on certain questions.

[The Chair, Conservative, Michael Chong]

Thank you, Mr. McLeish, and thank you, Mr. Rota.

Monsieur Bouchard.

[Robert Bouchard, Bloc, Chicoutimi—Le Fjord, QC]

Thank you, Mr. Chair.

I would also like to thank the witnesses for being here with us this afternoon.

My first question is for Mr. Simard.

Mr. Simard, you are a professor and researcher. I believe you use the data from Statistics Canada. So you know how important it is to have reliable and accurate data.

As we know, the minister's decision is to replace the long questionnaire, which is currently mandatory, with a voluntary questionnaire.

Why are you opposed to the minister's decision?

[Martin Simard, Research Professor, Department of Human Resources, Université du Québec à Chicoutimi]

I would first like to mention that I am a geographer and urban planner in addition to being a professor at the Université du Québec in Chicoutimi. The census data are very useful for the applied research we do on the environment, transportation, residential construction, urban planning, and so on. This research is often conducted jointly with local municipalities, which are also major users of census data.

Personally, I am against the changes or the reform the government wants to implement right now. So I join with the Canadian Association of Geographers and the Canadian Institute of Planners, of which I am a member, and which have clearly expressed their opposition to these changes. We are opposed to this for a technical reason, meaning that it could seem like a rather minor change, but this minor change, when you know the methodology of surveys and the scientific methodology, may have serious consequences on the validity and reliability of the data from Statistics Canada.

It is a shame since Statistics Canada is a well-respected institution in Quebec, and in the rest of Canada, a source of pride for many. So the data may be deemed as less reliable, especially for academic research. That may hinder our research, making it less accurate than the research done in other countries. That may also affect private companies that conduct market studies to choose locations for restaurants or businesses. Major problems will also arise in the development of public policies, especially locally and regionally where data may be even more inaccurate.

So, there are a number of negative consequences. Finally, we ask ourselves why we should change something that works well. If we were on The Red Green Show, we would say: “If it ain't broken, don't fix it.”