So each province to hold its own census would be OK then. Sorted.

no need to.

How then are the provinces to plan the appropriate levels of health care and education related resources?

You rightly point out that they are provincial responsibilities, but that doesnt really address the use of the demographic information gathered in a federal census to effectively plan for changes in the age and cultural or ethnic make up of the population.

look, the "long census" has been changed to be voluntary rather than mandatory. If there are people that feel strongly about it (like you) then they can simply answer the census and a good statistical sample will be realized. What is the big deal?

To answer the second part of your question, demographics are best observed at the local level. Most of the info needed with regards to age, etc, are on the short census and income tax returns. "Cultural make-up" should have no business what so ever in determining government policies. What type of policy's do you see and support the federal government implementing based on "cultural make-up"?

Ideally it would be none. But we don't live in an ideal world. Say you find out that 25% of your population is chinese with poor english - do you say screw them or do you adjust your polices to take this into account. The dogmatic approach is not necessarily the most effective - but that's by the by really. Without knowing policy makers don't even have the choice.

The big deal is that the chief statistician at Stats Canada felt that making it voluntary and changing the sample size corrupts the integrity of the data sufficiently that he was driven to resign over it, something career civil servants seldom do.

Maybe in part his concerns were tied to the idea floated in the first post here that if its voluntary then interested parties can mobilize themselves to try and influence the outcome of the survey?

The actual policies and decisions made dont interest me as much as the fact that any decision making process is only ever going to be as good as the data its based on. He appears to believe strongly that this data will not be as solid as it has been in the past. Wont stop politicians making bad decisions of course, but anything that helps them get it right (or at least the civil servants spending my money to implement policies) has to be worth looking at.

Im not qualified to second guess the judgment of the late departed chief statistician over what changes made to the long form census will make the data less valid, to be honest that's not something very many people would be qualified to do.

You have got to come up with better examples than that. Here is the short census. http://www.statcan.gc.ca/imdb-bmdi/i..._Q1_V3-eng.pdf

He will no doubt retire comfortably after 30 years as a civil servant. This is from the original article posted by the OP.

"Previously, 20 per cent of Canadians households received the mandatory long-form census every five years, providing information on issues such as ethnicity, education, employment and disability. That questionnaire will now be replaced with the voluntary National Household Survey and sent to one-third of households."

Here is the long form census. http://www.statcan.gc.ca/imdb-bmdi/i..._Q2_V3-eng.pdf

The real debate is about whether it should be mandatory for those 20% who received the form in the first place. Who is to say that everyone fills it our correctly in the first place?

Whether or not he retires with a big pension isnt really all that relevant to his views on statistics I wouldnt think.

Correct me if Im wrong, but isnt the short form mandatory for everyone anyway? If its the compulsory nature of the things thats an afront to your civil liberties then whether its long or short form surely makes no difference, compulsory is still compulsory, even if its only a five minute compulsion! One suspects that its simply the brief inconvenience of having to fill out the from that's the problem, rather than the nature of the information provided itself...

Anyway, you've mistaken me for a statistician, Im not qualified to argue whether the new way is less reliable or not. I know at least one high ranking statistician who thinks its not though

A psychologist friend of mine once explained that there are ways and means to weed out the bogus responses in a survey. Wont bore you with the details.

Whats harder to deal with is when the sample of respondents itself is skewed. Kind of like all those people who declared themselves Jedi on the last census after an email campaign. Problem is its not always going to be that easy to spot the results of an orchestrated campaign of voluntary respondents I guess.

Statistically I dont suppose the end results are going to be hugely different, so lets look at it from an economic standpoint. Why go to the expense of sending out the long form survey to even more people than are necessary if half are destined for the recycling bin? Its a waste of money to change a system that works based on the whining of less than one percent of those required to complete the long form. One wonders what the take up rate on a non compulsory long form would be. Didnt the US try it and abandon the idea?

So you can look at it from the point of view of the statistical validity, or you can look at it from the cost to carry it out, and either way the compulsory form makes more sense. Last I checked the short form was compulsory anyway, so the argument that making people respond is an infringement on their civil rights (or whatever) doesnt hold any water at all, unless they intend to make the short form optional too.

No, actually I don't. For a government to serve the community cost effectively it needs to know who's in it. This is obvious really. Maybe certain ethnicities are shown to be relatively worse of financially than others - how would you address this without knowing?

Well, as far as addressing ethnicity with lower incomes, what do you suppose the federal gov't do? What do they do now? It doesn't serve the community cost effectively now with the mandatory census. Is it a secret that Natives are on average worse off then the rest? I can guarantee that the mandatory long census didn't suddenly uncover this fact.

You are assuming that making the census mandatory for 20% of the population and not the other 80% means that people will fill it out more honestly as opposed to sending it to 33% of the population and making it voluntary? Can't your psychology buddy spot the difference in either case? Why are you assuming that groups of people will turn evil with a voluntary census? Couldn't they do that with a mandatory census?

I don't see a conflict between the collection and analysis of census data compared to policy making based on the results. You could even make an argument that it's good that they are separate.

I've no idea what they do now however.

No it isn't a secret, but you need to measure it - otherwise how do you know if what they are doing is having an effect? It's not just the absolute figure that is important, it's also the trend.

I'm with Alan, the government needs a snapshot of the country it governs in order to have something accurate to work from. Much better than governing based on tabloid headlines for example, the census is a pure model of the country to shape and mold.

even to the point of it made mandatory rather than voluntarily? The most important info such as age, sex, kids, families, language, income, employment, etc are all gathered in other ways besides the long census forum. What else is needed to know and what and how would the gov't change/make policies based on those things?