Wednesday, July 21, 2010

Will Censusgate Cost Tories Votes?

I sincerely doubt it, but that is one of the politics headlines at the Globe (and today's poll question). There have been a lot of things this past year that could have conceivably cost the Tories votes, issues far more serious than this. As a stats junkie I support making Canadians fill out the long form census, but I know that a lot of normal Canadians don't like filling out the long form with all that detailed information. Maybe putting the G20 in Toronto will cost the Tories some votes in Toronto, but how is relieving a large proportion of the population from doing something they don't want to do going to cost the government votes? It seems like a silly question to ask, which is why I'll ask it...

Look, it is the summertime and not much is going on. Most political pundits undoubtedly experience a decrease in traffic during the summer. People spend more time outside during the summer. I suspect that those friendly to the Liberals in the media are trying to pump up the volume to deflect attention away from Ignatieff's 3 hour bus tour. Kady O'Malley for example has written 8 consecutive posts on censusgate, and the last non census piece was "munk ado about nothing" denying the Iggy U of T job offer rumours. You have the Liberal leader touring the country making an ass of himself, and all Kady wants to write about is something that only academics seem to care about. Again, I support making Canadians fill out the big census, but if the form is dumped I won't lose any sleep over it...

Iggy doesn't lose sleep over war crimes, but the long census, that's another story! Or is the media creating this separate from the Liberal party because the summer news cycle has been slow, and the only interesting political story has been Ignatieff fumbling through his first attempt at campaigning. The G20 protests backfired on the protesters. Fake Lake was massively popular with the thousands of visiting foreign media members. Is the only reason this is a story because the news cycle has been slow?

12 comments:

  1. "— 49 per cent in favour and 51 per cent against — on whether the Conservatives' move is "good" or "bad," according to the results of an Ipsos Reid poll conducted exclusively for canada.com. ...

    Quebecers showed the strongest support for abolishing the long-form census in 2011, with 62 per cent saying the move is "good," ...

    Overall, one in five (19 per cent) Canadians say they won't fill out the long form if it's voluntary, ..."


    link

    Let's do some math. Under the old rules, 20 % of the population received the long form, and let's say they all returned it (in fact 160,000 refused to fill it out).

    Under the new rules, 33 % will receive it, and 81 % of those will supposedly return it. Multiply .33 x .81 and you get .267 !

    So Stats Can gets 6.7 % MORE returns to key in and analyze. But is this larger sample less representative ? I took stats at university, and this kind of situation never came up.

    But my sense is, if you sample 26.7 % of the population, you're going to get a pretty good representation, even if you just blindly send the form to every third household, everywhere in the country.

    Now, in the case of polls, which sample 1,000 to 2000 out of, say 20,000,000 then the sample has to be stratified among regions, different demographics have to be represented, etc--you gotta be a LOT more careful in HOW you take that sample.

    ReplyDelete
  2. More than 160,000 Canadian refused to fill it out, a five-fold increase over 1981. More than 60 of those who refused were referred by the government for prosecution.
    There are some people, including the leader of the Liberal party, Michael Ignatieff, who believe Canadians should be forced to divulge intimate, private details about their personal lives to the federal government. We disagree. -Tony Clement
    Wilson gave me link.

    We are wising up to the Central Planners who used intimidation to secure details about our legal activities.

    Just like prorogue was blown out or proportion and Omar Khadr these are not campaign vote winners.

    Less taxes, less waste, less intrusive narrative has taken hold in Europe as the left are losing all support. Obama may follow Jimmy Carter, this November how many Democrats are going to fall?

    ReplyDelete
  3. Well that poll should quiet the media/Liberal elite driven census-gate!

    Organizations don't vote, citizens do.

    ReplyDelete
  4. If I get the long form I will fill it out - I have nothing to hide ... opps white male, 60ish, retired, english speaking, BC Lions supporter, red/white wine drinker, needs vacation (What - you are retired!!) ... and I vote conservative.

    ReplyDelete
  5. Once again I must ask the question. Is the long form census being eliminated or is it simply the criminality of not completing it being eliminated?

    ReplyDelete
  6. The long form is going to be voluntary, not eliminated, so there will be no census police hunting yah down, threatening you to fill it out or else....

    The clincher to all of this, and where the Liberal elites and their media get the hook,
    is that most other census' that organizations do depend heavily on, maybe even more so than this long form,
    are voluntary.

    So how is information from other 'voluntary' census' reliable,
    but wouldn't be in this case?
    This is simply wafer chasing by Liberals and their media.

    Canadians are not stupid,
    and organizations don't vote.

    ReplyDelete
  7. My understanding is the criminality of not filling it out is being eliminated, and it will be voluntary.
    With all the media coverage of this I would not be afraid to guess that if it is still mandatory next time, a lot more than 160,000 will not fill it out completely, and we will go public with all intimidations we get.
    Mary T

    ReplyDelete
  8. Doesn't anyone see a pattern here? The Liberals love making criminals out of law abiding Canadians. Gun Registry, Census Long Form etc

    ReplyDelete
  9. One of StatsCan's recommendations to keep the data as reliable as a mandatory census, was a larger sample of voluntary respondents + more promo. Cost $30m (2006 census cost $567 million spread over seven years, it takes that long for the census police to get the info !).

    So now we have an equally reliable data source,
    and those end users that pay for the data should be charged 5% more to pick up the cost.

    ReplyDelete
  10. I found it interesting that the 'Optional Long Form Deniers' used the US expample of sticking with a mandatory form as ammunition for their cause.

    Canadians are very different than Americans when it comes to civil compliance. Just look at the seat belt compliance rates between the two countries.

    Canadians will complete the optional long form in droves and Stats Can will get better data than ever before - and those Canadians who object to the intrusion won't become criminals.

    This is a non issue.

    ReplyDelete
  11. We have a list of who voted Liberal in the last elections.Send them the long form,with the threat of jail if they don't fill it out.Send the short form to the Conservatives and the Blockheads and NDP could just be listed as trouble makers.

    ReplyDelete
  12. The problem with the voluntary form is that it screws up the statistical accuracy. For example, people with busy lives and demanding jobs obviously won't want to spend what precious little free time they have filling out a long form, where the unemployed are more likely to have the spare time and fill out the form. If one group is more likely than another to "volunteer" then your data will be representative of one group instead of the entire population. (then policy based on the data will be skewed to the unemployed, not the entire population).

    Another example, I do polling on my website, but they are all biased. It is voluntary for people to visit my site, and the people most likely to visit are Conservatives. Ergo my polls are representative samples of the demographic most likely to visit my website. To remove the bias from the poll question, I would have to randomly contact people directly and ask them the question. The voluntary nature means that the long forms will represent the people willing to do the extra paperwork and not be a representative sample of the entire population.

    But again, I am a mathematician who wants to merge census data with election data and write a book about it. Perhaps that's selfish of me, but if it helps elect Tories, one could argue that it would be great for the country.

    ReplyDelete