Friday, July 16, 2010

Census Issues No One Wants To Consider

Canada has had a long version of the census which was distributed to Canadians on the basis of one household out of five. This long form was obligatory and only 3 complaints were ever recorded on the intrusiveness of this long version. Today we are discussing how we will transform this by making the long form completely voluntary and distribute it to Canadians on the basis of one household out of three.

Voluntary versus Obligatory

Anyone with an inkling of how statistics are processed knows that a voluntary basis of acquiring information is the most flawed methodology you can imagine. Those who are predisposed to answer may also be introverted chess players who love to fill out long forms and will express how we don't have a national program to teach chess in kindergarten while the extroverted hockey fan would rather be outside playing street hockey with his friends instead of filling out such boring statistics and then wonder why the national hockey budget has been cut by half to help fund some chess for toddlers competition.

This issue has been discussed in the news but not thoroughly explained. We are told that its "bad" for voluntary census forms and that we should enforce compliance with jail terms for those who refuse to do their civic obligation. I suspect there is a reason why we are not told the "why" its bad.... more on that in a few paragraphs.

Is it intrusive?

Another issue that has been discussed is the intrusiveness of the information gathered. Even though only three recorded complaints on that issue have been revealed to us we are claiming that this is the worry of the government and this is why we make the long form voluntary. If the government was truly worried about intrusive questions and how they are recorded and used by others perhaps they should look into private corporations who probably have better specific information on very specific issues on most of us.

With the right hacker, an astute psychologist and your Airmiles number I can tell you what kind of person you are based upon your purchases in the past 5 years. Do you really believe that the coupon you received in the mail last week addressed to you personally was sent out to everyone else on your street? Do you really believe that 1 airmiles point for the information you provide to marketing statistics causes them a loss or do you realize that they make a profit from your information freely given through manipulation?

The government may be untrustworthy on handling some information but I would trust a corporation even less on most issues. Again, this issue of intrusive questions and what is done with the information was also dealt with in the news but not nearly as thoroughly as they could have gone and with the wrong target. Statistics Canada is probably more worried about security of its data and personal information being linked to any specific individual than lets say (hypothetically) Costco with your purchasing habits. But we don't really want the news to tell us what to really worry about, do we?

Probable Issues

There has been a situation in the last census where a group of like-minded individuals were communicated virally through email and internet chats with instructions to deliberately lie on their long version of the census. Apparently they obeyed because from one census to the next we had an illogical turn of events. People who were bilingual suddenly unlearned their second language, being English, and were only able to speak French. This set of instructions had the purpose to force the Federal Government to place more money in specific programs to insure that the French had services in their mother tongue. They used the census form to manipulate government policy. This part has been ignored by the news.

Now as much as I agree that making the long form a voluntary issue is bad for statistical studies, it is nonetheless great to witness who answers the voluntary form and see if they are also using the census to manipulate government policy. If all the bagpipers of Canada are instructed to mention in their long form how they have nowhere to practice their loud music and no one else bothers to fill in their voluntary census.... then we will see that this group is trying to manipulate the government in granting them parks specifically to play the bagpipes.

I know bagpipes are not really something to worry about, how about a cultural group or a religious group who instructs their members to lie so that their group can receive genuine concessions? It has already happened with one linguistic group.... that we know of. It is not unusual to think that a minority cultural or religious group to try a similar tactic. The voluntary form would spot this quite easily because they would all fill out the form above and beyond the average Canadian. So this would not give us an accurate picture of Canadians but it would give us an accurate picture of which social groupings are attempting to manipulate government policy.

Now before someone accuses me of being unfair against A or B or C..... If any of us were part of a minority no matter what it is.... we would be tempted to use this tool of lying to get an extra advantage.... and if you are part of a truly exceptional minority where you would not lie then you must also realize how rare you are because your very neighbors don't have this same exceptional quality of being truthful as you..... and if you deny that then you are just lying to yourself.

Do we drop the entire census because we are afraid that a few would lie? Not at all, this may be an interesting short term solution as we develop a better methodology to track immoral instructions to lie on census forms. Will this throw back Statistics Canada with inaccurate information? Which is better, knowing the info is flawed because the gathering methods are flawed or believing the info is accurate when we don't know if there was a deliberate intention to misrepresent answers? Too bad we are not discussing this issue openly and rationally amongst an educated electorate.