The 2011 Canadian Election Study is available for download. You can get the file here.

(If you don’t have SPSS, you can load it into R using the SPSS import functions.)

I invite you to explore it.

What is it?

An entire generation of Canadian market researchers and pollsters grew up on the Canadian Election Study (CES). And there are a lot of them! There were federal elections in Canada in 1997, 2000, 2004, 2006, 2008, and 2011. That’s 6 elections in 14 years! It generated an incredible amount of publicly available data about political attitudes in Canada.

Research uses aside, the CES is used to teach Canadians about electoral behavior. It is among the most studied data sets in Canadian political science. By comparison, Marketing Scientists have over 400 openly available data sets to pick over over the course of their undergrad and graduate studies. As a result, it’s one of the most thoroughly mined datasets I know of.

The three major dependent variables are the outcome – “who won?”, “who voted?” and “who participates?”. There are hundreds of independent variables to choose from. Check out a portion of a crosstab below.

This entire area of research is called electoral behavior, and it is based on understanding how the attitudes and attributes of people cause voting actions.

Market Research is a natural extension of these university experiences.

  • The CES is a survey of reported attitudes, behavior, and attributes.
  • Market Research is based on surveys of reported attitudes, behavior, and attributes.

  • The CES comes in SPSS format.
  • Market Research is traditionally done using SPSS.

  • The CES seeks to explain voter behavior.
  • Market Research seeks to explain consumer behavior.

It may be of some surprise that many of those in Market Research don’t really have their roots in pure marketing, but rather, electoral behavior.

Isn’t this an analytics blog?

Yes. Yes it is.

Market Research is not Analytics. However, just because the instrumentation is different, that doesn’t mean that the inputs from surveys are entirely useless. The entire VOC industry in digital analytics exists, doesn’t?

So why should I care?

You would have stopped reading if you weren’t interested.

By spending some time on a relatively clean, nice data set, you may come understand more about why (many) (Canadian) Market Researchers people think the way they do. It may change the way you think.


I’m Christopher Berry.
I tweet about analytics @cjpberry
I write at

Invitation: Compare the actual voter turnout figure with that reported in the CES for a key insight.