Let's scrap the long-form census!

The status quo for Canadian censuses is that one-fifth of households complete the “long form” of the census in its entirety and the remainder are given a much-abbreviated version of it.  The completion of the long-form census, until now, has been mandatory for those to whom it has been delivered under threat of fine and jail.  Census information is important to all Canadians; it is used by different levels of government, businesses, historians, scientists, and others for reasons as varied as planning infrastructure to creating employment opportunities.  The current governing party of Canada has decided, unilaterally, that punishing its citizens for failing to fill out some survey is unfair and that the penalty be abandoned.

The issue of whether it is fair to punish citizens for not filling out a census form is a red herring in the brouhaha that has ensued.  The government demands that income tax returns be filed for those making above a certain threshold; this is a non-issue for most people.  If the penalty for failure to comply was a life sentence in jail, I doubt it would trigger the same level of outrage from the current governing party.  Paying property taxes and filling out censuses are simply part of the price Canadians pay for living in homes.

Instead, the real concern is the integrity and value of the census.  With a voluntary form, self-selection, the perennial problem plaguing people-studying researchers, rears its ugly head.  Just as one might infer that everyone was returning iPhone 4s because of antenna problems based on the vocal few, plans will be made based on responses from those that opt to respond to the census.  Sure, this may sound like a boon to those who end up filling out the form, but, dare I say it?  This sounds most undemocratic.  Without knowing how to correct for response biases, the census loses much of its value.  In fact, some studies with voluntary responses use the census to account for response bias!  The quality of the data collected, should these changes be implemented, should still remain above the science that many if not most scientists studying people engage in simply due to the number of responses; however, if experience designing experiments has taught us anything, a voluntary census will be but an elementary school science fair project in comparison to a mandatory census.

Another argument in favour of scrapping the mandatory census that has been trumpeted by the government is reduced cost.  While moving to a voluntary model may reduce costs (after the $30 million advertising campaign designed to improve response rates), what are we getting for that reduced cost?  If the data is unreliable, why even bother paying for a census anymore?  There’s a saying that goes “you get what you pay for,” but this is more a case of it’s an idiot that pays twice.  Further, since more individuals (one-third voluntarily instead of the mandatory one-fifth) will get a chance to voluntarily fill out a full-form census (which, if the governing party gets its way, I think it is in our collective and individual best interests to fill it out), there is a potential for more data to be collected than under the old system — which results in higher coding costs (i.e., to record the data in a usable form).  In fact, we’ve now (voluntarily) wedged ourselves between a rock and a hard place:  depending on the number of voluntary responses, we’ll either end up collecting useless census data and waste money doing so OR have a useful sample size and spend more money than we do now ((A large proportion of the one-third of people that can voluntarily fill out the census must do so — I’m no statistician, but I think more completed responses than the status quo 20% of households is required — which not only means added cost, but added time commitment on the part of Canadians, albeit on a voluntary basis.  And this assumes everyone fills in the short form which, which may still be mandatory due to bizarre logic on behalf of the governing party.)).  While we will not immediately experience many ill effects of a crippled census, it’ll happen eventually as plans based on the new census data begin to materialize based on inaccurate data.

So, by all means scrap the long-form census and all the benefits that come with it.  But, please, if you’re going to do that, save our tax dollars and scrap the inefficient and/or ineffective voluntary census that you’re planning on foisting on us in its place.  Just get Paul the octopus to make up some numbers instead.

3 Replies to “Let's scrap the long-form census!”

  1. It is my opinion that making something mandatory will not in any way improve the accuracy of the information gathered. I think the opposite is true. Our government (or anyone) has no business forcing us to answer those questions. There are many ways to gather this information, let us get on with it and make the most of our ever changing technology. There is no need to threaten people in order to gather information. I thought that our great country was much more savvy than what I have been hearing in the MSM these days. Many countries are way ahead of us on this matter. The media and other influential people should do some homework before they criticize.

  2. Fortunately, we do not need to resort to conjecture and opinion to guess whether voluntary responses are biased. Unfortunately, the answer is that they are. Countries that have accurate data use other mandatory means of extracting information that we gather by census (for example, via income tax forms).

  3. The long form census gathers information on ethnicity, religion, education, work, housing and disability. This information is already in other government databases. The long form census is most certainly redundant. Perhaps the short form census is too.

Leave a Reply

Your email address will not be published. Required fields are marked *