Political Perspectives is produced by the students and faculty of Carleton University's School of Journalism and Communication, Canada's oldest journalism school.


The NDP and online polls – some cautions

Posted by cwaddell under All, Election 2011, Election 2011 Campaign strategy, Election 2011 Faculty links, Election 2011 Media commentary

Christopher Waddell

So the week begins with a “surge” in NDP support in two online public opinion polls. That certainly fits the media’s need to find a narrative for the campaign’s final two weeks. If there isn’t going to be a race for first place, a race for second is more entertaining than no race at all.

However precedent suggests it is worth being cautious and asking some questions about NDP performance in online polls.

In the 2006 election, Decima Research conducted a series of experiments comparing polling results from an online panel it had assembled with those obtained from traditional telephone polling. The goal was to see how accurately online polls could match telephone results and to try to figure out how online polls should be weighted compared to the traditional demographic weighting done for phone polls to ensure the pool of respondents matched the demographics of the country.

On election day, Decima asked the members of its online panel to report how they voted in an attempt to create a national exit poll. About 10,000 people responded and by 6:30 pm in Ottawa – three hours before the polls had closed everywhere west of New Brunswick – it was obvious Stephen Harper and the Conservatives were going to beat Paul Martin and the Liberals. There there was a clear gap in the share of the vote each attracted and while Liberal vote share was below 2004 results, the Conservatives were significantly higher.

By the end of the night the Liberal and Conservative vote shares in the online poll were pretty close to their actual results. The one discrepancy was the NDP which did several percentage points better in the online results than it did in actual voting results across the country. This difference was most pronounced in Atlantic Canada where the NDP was 7-8 percentage points online better than it was in the ballot box.

A conversation three or four weeks later with another polling company that had done a similar test on election night revealed the same thing. The NDP scored better online on election day than than it did in ballot boxes.

There was no simple answer but the most likely seemed to be that online panels overrepresented urban voters at the expense of non-urban voters. Urban Canadians were more likely to be heavier Internet users (with better Internet connections) and therefore more likely and more enthusiastic about participating in Internet polling panels. The possibility that online panels oversampled young voters compared to older ones could be a further explanation. The assumptions that the NDP do better with younger voters and with urban voters seemed to be possible explanations for the difference.

It is five years later and polling methodology as it relates to online samples should be much better but it is interesting that much of the NDP growth is in Quebec where there is a clear delineation between urban and cosmopolitan Montreal and the rural nature of almost all the rest of the province.

It would be worth the media’s time to ask pollsters a few questions.

How do they draw their online sample – where does it come from and how does someone get into an online panel?

How do they weight the online panel results for demographics – age, gender, income – all the usual weighting factors?

How does the weighting they do for online survey results compare with the weighting they do for telephone-based surveys?

What demographic indicators have to be weighted more online than for phone polls? Here age may be important as younger people are more likely to be Internet users so may respond to polls done during the campaign but are equally less likely to show up and vote on May 2. How does the online poll factor that behavior into its results?

Finally how do online polls weight results for rural and urban respondents as some evidence suggests that could be a more significant distinction online than it is on the phone.

Christopher Waddell is director of the School of Journalism and Communication at Carleton University. He is a former reporter, Ottawa bureau chief for the Globe and Mail and a former CBC-TV parliamentary bureau chief and executive producer-news specials for CBC TV News. You can follow him on Twitter @cwaddell27

Reader's Comments

  1. Matthew |

    It makes sense to me that NDP would show higher in the polls. Its supporters are generally more enthusiastic and probably willing to go out of their way more than a liberal or conservative supporter to get into a poll and support their party.

  2. Antonia |

    While I agree Nanos has wild swings, and figures that often don’t mesh with larger pollsters, this sample from the last few days of polling in Quebec suggest the general point of an unexpected (by the press, its been known by others for a year), surge by the NDP, is being seen by all polls.
    poll results

  3. Siobhan |

    I think it depends essentially on the digital divide which is actually NOT so much urban/rural though that is a factor in some parts of the country but also on WHO willingly takes part in polls. They tend to be younger voters who may not get out and vote even if they support the NDP so there is an age gap as well. My father may bank online but respond to a poll? Never! But he certainly votes! That generation, the one Linda Duxbury calls the Vets, likely have a higher voter turnout and a smaller partipation in polls. It is the reverse with Generation Y and even more so with Y who will say for example, “I feel like just placing a line through my ballot”

    Rural voters also have a higher turnout – they unlike urbanites depend more on their representative for govt services and in one cynical comment I was give while campaigning with someone, a member of a small community with economic challenges said to me “Don’t worry, I went and got my name off the list” Translation- they are more savvy about how parties work and know if they seek a govt job, service, contract et al, the local party will know that this individual actually voted. So even if they spoiled their ballot, they ensure they are on the list!!

    I really think the Quebec trend – if it does not peak too soon – is actually true. I live on the border of Quebec and people who typically vote Bloc have been telling me for weeks now they really like Jack Layton and somewhat tired of Gilles Duceppe (one has to wonder if this situation would exist with a new BQ Leader?) and while I do know les Quebecois et Quebecoises have the ability to vote en masse for a party, I am unsure if this support will result in the number of seats some dream of. That said, I look at NB and the 1987 election in which all seats went to one party (and NB is rather like Quebec- we tend to vote m,assively one way though that was quite an abberation!).

    It is also a quiet election – maybe not online or among the pundits but out in the trenches, it is TOO quiet, rather like the pre 1987 NB election or the one that booted out Kim Campbell, Mulroney and friends in 1993. I think many news media organizations cannot quite believe the NDP are doing as well as they are and are downplaying it to some extent – the polls and even electionprediction.org are some examples of this. They may be surprised on election night – though I do not anticipate an NDP minority govt (but to be honest, would be not unhappy if it occured) and think Harper et al will return to where they were but the possibility of an NDP Official Opposition does exist. Stranger things (ie the Bloc as Official Opposition or a party against official bilingualism in NB) have occured.