Canadian National Household Survey (Long form Census) - low response rates potentially bias findings.

Not surprisingly, Statistics Canada has put innumerable caveats on the release of the long form information.  Under mandatory completion, rates exceeded 94%.  Under voluntary completion, a full third have not completed the form.

There are innumerable stories out in the last 24 hours on the identified shortcomings with Statistics Canada disclaimer on the data problems.

Census coordinators often used to hand hold individuals who were challenged in completing the form.  So, who most likely choose not to complete.   From the privacy paranoid individuals wanting to protect their secrets who have forced the change, they are likely those with something to hide like illicit income or total wealth. A very small fraction of the respondents.

No, it was the literacy challenged, those with less proficiency in English or French, those with fewer resources, those that might feel unwelcomed that are likely the ones who avoided replying.  This skews the census results to a higher socio-economic grouping, perceptions of fewer problems and challenges and makes Canadian social policy look better than it likely actually is.

Great if you are the government in power.  Great if you want to trivialize the country's problems.  Great if you disperse funds based on documented need.

Bad if you are actually trying to do the right thing.

There doesn't seem to be any appetite of the Harper government to move back to an evidence base on which to base policy decisions. It is however a great example of policy based evidence making.