The big longread event in this weekend’s papers was, without question, Shannon Gormley’s piece in Macleans on the Thai Cave Rescue (if you haven’t read it yet, stop everything and do so.  I’ll be here when you get back.  Amazing, right?  OK, let’s move on.)

Anyways, the second most important longread was the big Globe and Mail feature on Canada’s data gaps, which was actually two pieces, one on data gaps generally and one specifically on Statistics Canada and why the agency is not very good.  There was much applause in the twitterverse, which was mostly deserved.  But having read both pieces three times now, I can’t help feeling that while these pieces are very good at describing the symptoms of our data problems, they haven’t correctly identified the disease.  Specifically, they’ve fallen a little too easily into the trap of “blame Statistics Canada”.

Now as you likely know, I am not a huge fan of Statscan’s work in education.  They had a really good run at data collection during the late 1990s/early 2000s, with the advent of products like the Youth in Transitions Survey, the Post-Secondary Education Participation Survey, the Survey of Approaches to Educational Planning, etc.  But while these were Statscan-run projects, they weren’t given core funding by Statscan. Rather, they were paid for by HRDC/HRSDC/ESDC/whatever that Ministry is called this week as one-offs.  If ESDC loses interest (as it did under the Harper government), Statscan is back to just running its administrative data collection role, which consists basically of student unit records, surveys of tuition and institutional finances, and (more recently) counts of faculty.  Oh, and the National Graduates Survey, which it continues to mess up. 

But to be fair to Statscan, as a country we have some bizarre expectations of what a national statistics agency is supposed to do.  In most countries, national statistics agencies have nowhere near as big a mandate as Canada’s.  Take the United States, for instance.  They have a Census Bureau, which also does the biannual American Communities Survey (which is way cool).  But it’s the Department of Education which does all the surveys and admin collection related to education, the Department of Commerce which produces trade statistics, and the Department of Labour which does all the Labour Market stuff.  Most countries have similar set-ups, though with different divisions of responsibility.  In some countries, the central agency publishes and acts as a clearinghouse for statistical information, but it is the ministries themselves that handle collection.  In others, the central agency handles the population surveys (there are spillover benefits from specialization), but the line ministries get the admin data. 

The benefits of having line ministries involved in data collection are clear – unlike an abstract statistical agency, they possess in-house subject matter expertise about the data they are collecting.  But in Canada, by sloughing it all off on Statscan, line ministries can largely ignore data collection and interpretation.  The current Statscan set-up is what happens when Homer Simpson’s winning slogan for sanitation commissioner (“Can’t Someone Else Do It?”) becomes an actual principle of data management.

If we’re stretching the “to be fair” bit, there aren’t many pieces of data in the possession of federal ministries that anyone actually cares about: the important stuff is all in the provinces, and it’s their fault the data doesn’t get to Statscan.  Many provinces are, in fact, irritatingly unwilling to share data or contemplate minor changes to data collection procedures in order to make cross-provincial standardization a reality. From the provinces’ point of view, the marginal benefit they get from making these cross-provincial comparisons is so tiny, it’s not surprising that they don’t bother.  The problem is that while it is true researchers should often be happy with provincial datasets, the fact is that most provinces often fail at data collection and transparency.

And it’s not just provinces.  Cities: they’re terrible too.  Universities and Colleges?  Don’t get me started.  One day soon I will publish my critique of every institution’s data transparency practices and it ain’t going to be pretty.  We have allegedly world-class universities whose transparency practices consist of 2-page infographics or who password protect their factbooks.  Only a handful of universities actually publish data in machine-readable format and most regularly delete data more then five years old.  And colleges are uniformly worse.

See, here’s the thing.  Collecting and publishing data is an act of transparency, and transparency, in practice, invites accountability.  Canadians hate accountability in public administration.  Just look at post-secondary education: everyone blabs about access or affordability, but no one ever defines these terms or sets systems goals because then someone would be held accountable.  Truth and Reconciliation?  The absolute key to this is hiring more Indigenous faculty, but no one sets actual targets because then someone would be held accountable.   Canadians don’t like holding each other accountable because we’re too nice.  It messes with the I’m OK/you’re OK ethic that pervades everyday life. 

So yeah, there is lots to complain about with respect to Statscan, for sure.  But the basic problem is way deeper than that.  Where good data is concerned, Canadians have met the enemy and he is us.  The Globe pieces hint at that a bit, but my overall impression is still that Statscan is being scapegoated.  Lord knows I like to rant about Statscan but leaving it at that won’t solve our deeper data woes.