How Kijiji’s data threw off Ottawa’s math on skills shortages – The Globe and Mail

Interesting story on the difficulties of getting accurate information, and the weaknesses of some of the social media sites like Kijiji and double counting. Bad data can lead to faulty conclusions:

Kevin McQuillan, deputy provost and professor of sociology at the University of Calgary, has written a paper challenging claims of a Canadian labour shortage and says the move to online job postings continues to give statisticians headaches.

“We are struggling to deal not only with changes in the labour market, but changes in how people hire,” he said. “We haven’t really gotten on top of this new way of hiring that’s done in online postings, [where] the same notice of a job appearing on multiple sites, or social media. So counting that can be difficult.”

How Kijiji’s data threw off Ottawa’s math on skills shortages – The Globe and Mail.

And a follow-up piece with the Government’s reaction to the story:

Mr. Kenney said critics should recognize the challenge of producing reliable labour data in a world of online job boards.

“Here’s the bottom line, everyone who is dealing with this debate should have a little bit of humility and admit that none of us know exactly what is going on in the labour market of today.”

Economist Don Drummond said better information can be produced at a cost of about $39-million a year. He was part of an advisory panel in 2009 that made dozens of recommendations to improve labour-market data, yet few suggestions were implemented.

The former TD chief economist would like to see one entity, such as Mr. Kenney’s department Employment and Social Development Canada or Statscan, “pick up the baton” and take responsibility for more detailed and current labour market data at the national and provincial level.

http://www.theglobeandmail.com/news/politics/tories-defend-use-of-kijiji-data-in-face-of-opposition-ridicule/article17690737/

Better for the government to spend some money for better data than, as it did in the change from the Census to the National Household Survey, spend more for poorer quality data (couldn’t resist!).