Chronopolitics, et al

[Jennifer Bell]

OK, "assumption" is strong given past data [1]...  but I do think the Jedi thing is probably more serious.  The last British survey was the first to see the rise of Jedi's, which it's true wasn't statistically signifigant, but I'm sure it got a lot of press locally.   ... and 1% is a lot when you consider that it was probably an internet-based citizen fringe campaign.

It's inevitable that the Canadian press will hit on the Jedi thing just before the next census.   Suddenly it will become cool in certain demographics -- aided by twitter, blogs, and other social media -- to stick it to the man by lying on the mandatory census.. eg. race: Purple. religion: Jedi.  education: PhD., thus exposing threats of punishment as empty.  Then, at StatsCan, what do you do with your 5% of purple jedi doctors?  Throw them out?  As a govt., what will you then say about the validity of your awfully expensive census?

Maybe the govt. is trying to avoid an even bigger media issue when the census does arrive.

I guess I object to this b/c I've never filled out a long form, and I didn't know it had questions on, for instance, on the # of rooms in your home.  (Nice to see Tasha Kerridan taking this line too in the Post [2]).  Being threatened with jail time if I don't fill out a series of arguably  intrusive questions seems, a priori, completely counter to the usual benign/happy Canadian value system.  

If you really want to incentivize people, why not offer a tax break for the particular segments that you want to see fill out the census correctly, cross checked by a small # of spot interviews?  I wonder if that's been experimented with anywhere?

Jennifer

[1]http://www.nationalpost.com/Voluntary+version+census+proved+unreliable+costly/3297456/story.html

[2] http://fullcomment.nationalpost.com/2010/07/19/tasha-kheiriddin-the-great-census-crisis-of-2010/#ixzz0uESiR65U

--- On Thu, 7/22/10, Michael Mulley <[hidden email]> wrote:

From: Michael Mulley <[hidden email]>
Subject: Re: [CivicAccess-discuss] Economist: the traditional census is dying, and a good thing too.
To: "civicaccess discuss" <[hidden email]>
Received: Thursday, July 22, 2010, 5:11 PM

On Thu, Jul 22, 2010 at 4:56 PM, Jennifer Bell <jenniferlianne@...> wrote:
>
> Really, the whole uproar seems to be resting (as Heather points out) on the assumption an
> un-enforced 'mandatory' data collection process is more accurate/better than a voluntary
> one.

I don't think 'assumption' is fair here. That a more-or-less-mandatory
census sample and a voluntary sample have very different
characteristics is a basic empirical result in social sciences that's
been demonstrated over and over again. And not only is one type of
sample indeed more accurate, but you just can't do reliable
time-series comparisons between the two types of samples.

There's also no reason to think that this will cease to be the case.
'Jedi' and so on have had essentially no effect on data quality in
countries where that's happened ('Jedi' maps quite nicely to either
'No religion' or 'Declined to respond', for example), and I don't see
any reason to suspect that a movement will arise that would lead a
significant number of Canadians to give misleading answers.
_______________________________________________
CivicAccess-discuss mailing list
CivicAccess-discuss@...
http://lists.pwd.ca/mailman/listinfo/civicaccess-discuss

[Tracey P. Lauriault]

Jennifer;

Have you read this yet?

http://datalibre.ca/2010/07/19/uses-of-census-long-form-data-question-justification/

I can get much more detailed information about your property from the land registry. I can find out the cost of your property and how you bought and with whom you have your mortgage and when you took it.    And then I can go and get the plans and look at the drawings and follow that up by going to the property assessment office to see how much your house is worth. 

That question about the number of bedrooms in the Census is cross tabulated with year of construction to know trends, also with the number of people living in the house to know over housing and housing density issues, also to assess what is available in the market and what is not.  The Canada Housing and Mortgage Corporations uses that data in their analysis.  I have used it to understand the # of rental accommodations w/ x number of bedrooms in an area, with average rents to assess affordability of housing for those on fixed incomes.  It is also used at the city scale to guage how the rental and buying market it.  This information can also be cross tabbed to assess if there is enough housing in a city or certain areas, or how seniors are faring as often there are issues where many elderly people are living in large homes alone and the condition of housing deteriorates as they cannot maintain it, so size, age, number of bedrooms, year of construction are cross tabbed to help with that assessment.

I have used that data primarily to study homelessness and poverty.  when I get the data from statcan the data are aggregated at various geographies to maintain privacy.  Also, if I ask for a cross tab, it goes through another level of privacy screening to assess that individuals cannot be identified.  If the sample size is too small, then StatCan will actually leave that geography blank again to ensure privacy. 

So, really, the number of bedrooms question is really not a problem.  And I am amazed how people feel afraid of that.  I know more about people by reading their twitter feeds than what the census knows about you as an individual. 

On Thu, Jul 22, 2010 at 7:53 PM, Jennifer Bell <[hidden email]> wrote:

OK, "assumption" is strong given past data [1]...  but I do think the Jedi thing is probably more serious.  The last British survey was the first to see the rise of Jedi's, which it's true wasn't statistically signifigant, but I'm sure it got a lot of press locally.   ... and 1% is a lot when you consider that it was probably an internet-based citizen fringe campaign.

It's inevitable that the Canadian press will hit on the Jedi thing just before the next census.   Suddenly it will become cool in certain demographics -- aided by twitter, blogs, and other social media -- to stick it to the man by lying on the mandatory census.. eg. race: Purple. religion: Jedi.  education: PhD., thus exposing threats of punishment as empty.  Then, at StatsCan, what do you do with your 5% of purple jedi doctors?  Throw them out?  As a govt., what will you then say about the validity of your awfully expensive census?

Maybe the govt. is trying to avoid an even bigger media issue when the census does arrive.

I guess I object to this b/c I've never filled out a long form, and I didn't know it had questions on, for instance, on the # of rooms in your home.  (Nice to see Tasha Kerridan taking this line too in the Post [2]).  Being threatened with jail time if I don't fill out a series of arguably  intrusive questions seems, a priori, completely counter to the usual benign/happy Canadian value system.  

If you really want to incentivize people, why not offer a tax break for the particular segments that you want to see fill out the census correctly, cross checked by a small # of spot interviews?  I wonder if that's been experimented with anywhere?

Jennifer

[1]http://www.nationalpost.com/Voluntary+version+census+proved+unreliable+costly/3297456/story.html

[2] http://fullcomment.nationalpost.com/2010/07/19/tasha-kheiriddin-the-great-census-crisis-of-2010/#ixzz0uESiR65U

--- On Thu, 7/22/10, Michael Mulley <[hidden email]> wrote:

From: Michael Mulley <[hidden email]>

Subject: Re: [CivicAccess-discuss] Economist: the traditional census is dying, and a good thing too.
To: "civicaccess discuss" <[hidden email]>

Received: Thursday, July 22, 2010, 5:11 PM

On Thu, Jul 22, 2010 at 4:56 PM, Jennifer Bell <jenniferlianne@...> wrote:
>
> Really, the whole uproar seems to be resting (as Heather points out) on the assumption an
> un-enforced 'mandatory' data collection process is more accurate/better than a voluntary
> one.

I don't think 'assumption' is fair here. That a more-or-less-mandatory
census sample and a voluntary sample have very different
characteristics is a basic empirical result in social sciences that's
been demonstrated over and over again. And not only is one type of
sample indeed more accurate, but you just can't do reliable
time-series comparisons between the two types of samples.

There's also no reason to think that this will cease to be the case.
'Jedi' and so on have had essentially no effect on data quality in
countries where that's happened ('Jedi' maps quite nicely to either
'No religion' or 'Declined to respond', for example), and I don't see
any reason to suspect that a movement will arise that would lead a
significant number of Canadians to give misleading answers.
_______________________________________________
CivicAccess-discuss mailing list
CivicAccess-discuss@...
http://lists.pwd.ca/mailman/listinfo/civicaccess-discuss

_______________________________________________
CivicAccess-discuss mailing list
[hidden email]
http://lists.pwd.ca/mailman/listinfo/civicaccess-discuss

--
Tracey P. Lauriault
613-234-2805