It looks like the kind of problematic outcomes that had been anticipated when the Harper Conservatives interfered with the information gathering format of Statistics Canada’s National Household Survey are now becoming evident. Findings from the data gathered in the 2011 Household Survey were supposed to be released on Wednesday, but are now being held back until September 11 because StatsCan found serious errors in the data.
It would be hard to overlook this issue since it wouldn’t take too much of a stretch in logic to see it connected with making response to the long-form census voluntary, while also altering the instrument and decreasing the number distributed. At the same time, distribution of the short-form survey, which remains mandatory, was increased.
Can anyone recall any previous time that StatsCan has cited serious errors in their data calculations? They have gone so far as to place a disclaimer on the back of the latest publications that have gone out so far, warning of “potentially higher non-response error than those derived from the 2006 census long form”.
The information gathered from those living in this country, the nature of the figures and patterns that are uncovered through these regular Household Surveys are vital for making spending and policy decisions at many levels, and are valuable to organizations and interested parties even beyond the government.
Why would a government choose to dumb-down data that are critical in decision making for the entire country? The optics may win some political support but the reality is that this kind of error occurrence is the real result.
Aside from the notion of protecting privacy, reiterated by Industry Minister Christian Paradis, and appealing to those libertarian types who feel less government involvement – read interference – is better, the changes in the census collection were rejected by demographers, statisticians and just about anyone who knows anything about the value of these kinds of data.
According to top-level statisticians with StatsCan, non-response to the long-form household survey was around five to six per cent before the rules changed to make response voluntary. Non-response has now climbed to 30 per cent.
Not only that, despite the short-form census being more widely distributed, there are groups that are more likely to not respond, such as Aboriginals and those in the low-income category, which leaves their numbers and other vital questions regarding such groups unanswered. If they are not represented, then decisions on programs directed at these groups are not going to accurately reflect their true need.
When statisticians try to compensate for this by manipulating the numbers at five per cent, the effect is closer to the true demographics. However, when they try to aggregate numbers at a 30 per cent non-response rate then “serious errors” occur. What we get are census data telling us, for example, that between 2006 and 2011 the proportion of immigrants to this country from the Philippines was higher than any other immigrant group, which is categorically false when lined up against Immigration Canada figures.
This altering of a key statistical instrument that serves the national interest is not a first for the Harper Conservatives as they attempt apparently to weaken the influence of certain areas of government that do not fit into their plan for recasting Canada in their vision, with a focus on energy and mining and less emphasis on social development. A government that does not make strengthening the social safety net a high priority would then be better able to take advantage of findings that under represent the most needy.
The attack on employment insurance recipients shows just how this federal government regards some Canadians. The Conservative government’s EI reforms, which were met with public protests, took eligibility away from thousands of seasonal workers, including education workers and students. It would not take much of a push for them to point to the findings of the newly engineered census to go even further.
Conversely, we might infer that if statistical information had ever held much sway with them there would not be as much focused resources and rhetoric in the direction of crime and imprisonment at a time when the national crime rate is falling.
If the goal in changing the rules at Statistics Canada was to weaken the institution’s credibility, then the Harper Conservatives are well on their way.