Through almost every news source, mainstream and not, I’m finding myself assailed by a constant stream of polling numbers which frame the Canadian election campain in ‘horse race’ terms. I’m also keeping a pretty close eye on twitter etc and getting a far different ‘feel’ for how people consider the various parties and candidates from what the poll numbers reflect.
Now of course my own circle (real and virtual) is going to impart its own bias. But I’m wondering if polling numbers are inherently going to favour one ‘side’ over the other just because of the nature of technology, society, and culture today.
What got me thinking about this? My phone rang a few minutes ago. Call display said ‘Quebec’. I ignored it. I generally ignore any number I don’t know on my home phone because I don’t want to talk to telemarketers, and they seem to call a lot. If someone has something important to say, they can leave a message. Of course, this means that I’m also ignoring any and all pollsters. That quebec number was likely a pollster.
Adding to this issue is the fact that I, like most folks my age, work full-time. I happen to be working from home this afternoon but that’s not typical. So the pollsters have 2 choices of when they can call me - during the work day, when I’m not around, or in the evening, when I’m least likely to want to talk to them. And I think people generally are getting less likely to answer random calls. Certainly less than 20 or 30 years ago. Many of us don’t even have land lines anymore - I think I’m in the minority in my social circle for actually having one.
So it’s pretty clear to me that there is going to be a limited, identifiable demographic associated with people who are willing and able to take pollsters’ calls. Folks with landlines, home during the day - these people are going to be far easier to reach if you’re a pollster. Housewives, seniors, etc etc etc. This demographic coincidentally seems to match up with the Conservative ‘base’ in many aspects. And it’s not surprising to look at the last election where the polls predicted a far higher number of seats going to the Conservatives than actually did.
I believe in the basic methodology of polling but this seems like a very large hole. Would love to get others thoughts on this.