Has Stats Canada Screwed Up job Numbers?

#ldnont has never had a month where 4000 people left the labour force. This month's data? 5000 gone. It's an error. @LFPress @KevinLabonte

— Mike Moffatt (@MikePMoffatt) May 8, 2015

Same day we got new warnings about bad data in the census.

In An Information-Based Economy When You Provide Crap Information You Hurt The Economy

From here.  Original stats-can report is here.

The Macdonald Laurier Institute On Canadian Crime

So, it is pretty clear that the Macdonald Laurier Institute (MLS), which put out this study criticizing Stats Canada's reporting of crime statistics, has close connections with the current federal government, whose ideological views the study seems to reinforce. I don't want to make too much of that, because its hard to argue against some of their more generic methodological recommendations: Stats Canada should report more fully and completely and in a more open fashion. So should every body that produces statistical information. But life is tough, and times are hard, and money is short.

So I am only going to touch on a couple of issues mentioned in the study (which, I am proud to say, I have actually read from end to end). Here's the first one:

The annual report on crime statistics, know as Juristat, routinely revises crime statistics from previous years upward in any given year’s report, making annual crime decreases appear more significant than they are.

OK. Below is the relevant time-series from the Juristat report in question. It displays the ebbs and flows in Canada's Crime Severity Index (note: a chart showing the more traditional measure of crime by volume can be found here. There is no accompanying graph, but it would look generally similar to the one below):
What the analysts from MLI are saying is that, when Stats Can reports the data point for 2009, it will be measured against a point on the graph representing 2008 that has been upwardly revised. Thus any decline year-over-year will be accented, and any increase muted. That's not surprising: an initial StatsCan crime report is issued, more data comes in that was not available at the publication date, and the report and accompanying graph is updated. It isn't as though there is anything that could be done to get around this this, other than making publication date so late that revisions are unnecessary.

But, more importantly, the issue MLI identifies only really applies to the last couple of data points on the X-axis of the graph above. Stats Canada figures might be revised once or twice, but eventually they become pretty much set in stone, so the revisions MLI notes will not alter the overall shape of the graph. Which is to say, the revisions will not show a long-term downward trend if there is an upward trend in the underlying crime data. The downward trend in crime that Stats Canada notes really is there.

There's one other point in the MLI study that I would like to touch on, because it seems representative of a number of the criticisms that they make:

This is, as the MLI admits, a methodological complaint. Stats Canada chops up and recombines its numbers based on a series of methodological decisions. But these are not the only decisions that could have been made, and indeed when you make re-cut the numbers as per MLI, you get table 7 above. The question then becomes whether this, as MLI asserts, is a "much better" way of chopping up and recombining the numbers or not, and whether the result is worthy of highlighting in the Stats Canada report, or not.

I will not defend the agencies various methodological decisions in this space, for they have already done on their own behalf, here.

Against Knowledge

A door to Canada's past has slammed shut, leaving future Canadians with little information about their own families and the country's history, in a move the government says was prompted by privacy concerns.

Statistics Canada has quietly made major changes to the country's census in time for the forthcoming round of national sampling in 2011. The long census questionnaire that provided information on a broad range of topics such as ethnicity, education, employment, income, housing and disability has been eliminated. Instead, those questions will be asked on a new, voluntary National Household Survey (NHS) and the results will never be released, in contrast with the treasure trove of census data that become public after 92 years.

Everyone hates filling out the long form, and the only reason you do it is because its illegal not to. So expect to see the response rate to StatsCan's new voluntary survey fall towards nil, which in effect means that the nation has taken a scalpel to its own brain and removed that part devoted to memory.

A truly stupid decision, probably the dumbest since Brian Mulroney decided to try turn Stats Canada's publications into revenue generators back in the 1980s. If you wondering why so much U.S. data is floating free about the Internet to be played with in interesting ways, whereas in Canada you have to pay $1 for every digit, blame him.

Lorrie Goldstein: Things Were Better Long Ago

In which Lorrie Goldstein opines the Good Old Days, and suggests that Stats Can statisticians are the same variety of Commie as Climate Scientists.

Jesus. People were whining about the good old days back in the good old days when I was young.