Wednesday, September 15, 2010

The Others

Most federal polls these days are prompted polls - in other words, the surveyors or computer programs list the parties from which a poll respondent can choose. Invariably, "an other party" is often provided as one of the options.

It's a great "none of the above" choice and is usually one or two percentage points higher than the less than 1% the "other" parties (and independents) usually get in an election.

But what about unprompted polls? In these polls, it is up to the respondents to name the party for whom they plan to vote. This usually results in lower "other" numbers. But what are these people actually saying?

Well, our friends over at Environics have filled me in on who the "others" were in their last poll. It's actually quite interesting.

First, there are the people who legitimately list an "other" party. They know their voting options, and tell the pollster that they will be voting for the Marijuana Party, the Communist Party, the Marxist-Leninist Party, the Rhinoceros Party, or an independent. Out of the 24 responses Environics provided me, this group accounted for 33%.

Then there are the people who know who they want to vote for, but do not realize that the party they are listing is a provincial party and not a federal party. So, we get responses like the "Wildrose Party", the Saskatchewan Party, the Freedom Party, and Quebec Solitaire (sic). This group accounted for 17% of the 24 "other" responses. One can't help but wonder whether the respondents misunderstood the question or sincerely believe that these parties will be federal options on voting day.

Some people say they will vote for parties that no longer exist - the Reform Party.

Then there are those are simply name parties that have never existed: le Parti Federal, the Progressive Alliance, or "TP".

Some people say things that don't quite answer the question that was asked: McGuinty, "the red one", or "the Harper administration".

And finally, a good deal of people take the opportunity to make a political message. These included "no to PC", "tout sauf conservateur" (anyone but Conservative), and "a new party if one comes up as now all are very bad".

In the end, what we get is that out of the people who say "other" in an unprompted poll, only about 1 in 3 actually list a party currently recognized by Elections Canada. The remaining two-thirds list parties that don't exist, are provincial parties, or make a political message. Thankfully, that portion of the population is less than 1% of those surveyed.

In the wake of the high "Others" numbers in EKOS polls, the calls for the formation of "The Others" as a real party appear premature.


  1. Would it be useful ask followup questions to try to determine how these people will actually vote when presented with a ballot? It might improve the predictive accuracy of the model.

  2. It could be, I suppose, but we're talking about less than 20 people in a poll of 1,800.

  3. I've certainly told a pollster in the past that I intended to vote Libertarian.

    I've been lucky enough to live in ridings with Libertarian candidates the last few elections.

  4. The Rhino party actually exists, though it doesn't run a lot of candidates. It ran in the Outremont, St-Hyacinthe, Vancouver Quadra, and Hochelaga by-elections, and 7 candidates in 2008.

  5. You're right, I thought they were the Neo-Rhino Party now, but it seems they've reverted to their old name.

  6. I was half-expecting a picture of "the others" from lost in this article:

  7. Timely post. The last federal Environics poll showed an astounding 9% for 'Other' in the Alberta sub-sample - greater than the NDP (7%) and the Greens (4%).

    Makes one wonder if the 'Wildrose Alliance' played a big part in that 'Other' selection.

  8. Now the trick is to get "None Of The Above" onto ballots.

    It's been done elsewhere with very interesting results !

  9. SJS:

    I had more this in mind:

    Thanks for hunting this info down, Eric. I now feel justified in my mocking of EKOS for prompting for "Other".

    The clear majority of "Other" responses are for options that don't really exist on the ballot.
    This artificially lowers the results for real ballot options.

    The most accurate poll would therefore not even count these in the result, in the same way that "don't know/refuse to answer" are removed from the tally. Prompting for "other" simply encourages people to give "noise" rather than usable data.


COMMENT MODERATION POLICY - Please be respectful when commenting. If choosing to remain anonymous, please sign your comment with some sort of pseudonym to avoid confusion. Please do not use any derogatory terms for fellow commenters, parties, or politicians. Inflammatory and overly partisan comments will not be posted. PLEASE KEEP DISCUSSION ON TOPIC.

Note: Only a member of this blog may post a comment.