An invalid question in the census. Seriously?
Submitted by Ambrose on Tue, 2016-05-03 10:35
Statistics Canada is giving us a web survey. It actually looks good, but I’m less than impressed with some aspects of it.
I’ve never seen a survey with a hashed code that does not have the ability to save progress. If you try to log out (such as… you need to find out something you don’t know, or your computer is forcing you to restart, such as how Windows Update often does) you’re told “Any information entered will be lost.”
Not even the hack job I did for my thesis did this. If they didn’t trust their “secure access code” enough to store progress, how much should we trust it? Sigh.
And we were asked whether people can speak “well enough to conduct a conversation.”
What does that even mean? “How are you? Fine, thank you”? or does that mean a real conversation?
If even the short survey has an invalid question like this, how is the long survey giving us valid data that can help us plan social programs? Don’t they test their questions at Statistics Canada? Sigh.