ANALYSIS:THE ROLE of opinion polls in charting the ebb and flow of a campaign was never more in evidence than for the recent presidential election. They tracked the changing tides from the first polls that showed David Norris as a front-runner right up until the final poll that confirmed a surge in support for Seán Gallagher.
However it has also been suggested they played a role in influencing opinion and somehow interfered with the democratic process. Nothing could be further from the truth.
Opinion poll criticism is part of the political game and polling organisations tend to ignore the protestations of politicians, even when they question the robustness or relevance of polls. An easy way to deflect attention from an unflattering result is to downplay the significance or credibility of the poll.
Understandably, politicians are unwelcoming of polls. To have your performance evaluated in such a public way is one of the most challenging aspects of political life.
It is of course ironic that politicians publicly dismiss polls when privately the larger parties and some candidates commission their own surveys.
But it seems the public may be listening to politicians when they undermine the validity of polls. When last measured by Ipsos MRBI, one in three Irish people said they did not trust polling companies. Thankfully polling companies are more trusted than politicians but trust levels fall behind that afforded to other professions such as doctors and teachers.
While the polling industry could easily chalk any distrust down to the link between polling and politics, or polls that only pretend to be scientific, even relatively low levels of distrust, for whatever reason, can leave it vulnerable to challenge.
The most recent challenge to polls has come from an unusual source – the media. Writing in last Friday's Irish Times, John Waters acknowledged the reliability of polls but called into question their role in elections generally, and specifically in the recent presidential election. He noted in the article that polls had "acquired an overwhelming sway over our elections" and hinted that they may in fact be unconstitutional.
If or how polls influence how we vote has been the topic of previous research. The most extensive report on the subject was published by Prof Wolfgang Donsbach in 2001. According to his study “Who’s Afraid Of Election Polls?”, there is no conclusive evidence that polls exert any influence on how we vote, and if there is an effect it is at such a minimal level as to be harmless. He notes that polls are usually low down the list of sources of information accessed or processed by a typical voter.
While the evidence that polls by themselves sway elections is thin, it is accepted that polls influence the media narrative. The presidential election perfectly illustrates this point.
From the outset, pre-campaign polls made David Norris the fron-trunner, until "lettersgate" derailed his campaign. When the first Irish Timeselection poll showed Martin McGuinness in contention, the focus of the media shifted to his suitability bearing in mind his past IRA membership. Attention then shifted to Seán Gallagher when the final Irish Timespoll gave him a substantial lead over Michael D Higgins with less than a week to go.
When we consider this sequence of events, two essential truths are revealed. Firstly, the polls never followed a trend: they reflected voter reaction to developments in the campaign and not prior poll ratings. And secondly, leading candidates were overtaken by events, not polls.
Did the polls sink the hopes of less popular candidates because the polls showed they had very little chance of winning and thus the prophecy of the polls was self-fulfilling? They weren’t for Gallagher who started out at the back of the field. Other candidates simply failed to connect with voters and drifted in the polls as a consequence.
It has been hypothesised that candidates who went into election day with no realistic chance of winning were less attractive to some voters because they were written off. This may have been the case – the bandwagon effect. It may also have been the case that some voters were more sympathetic to those who were struggling in the polls – the underdog effect. The arguments are complex in each direction.
When we examine the chronology of the campaign it becomes very clear that polls by themselves had a relatively benign impact on the outcome. But it would be wrong to say polls do not make a valuable contribution to elections, and by extension, to the democratic process.
Imagine an election campaign without published polls. The candidates and parties themselves, through private polls, would know the state of play, but voters would not. The most popular candidate or party would be the one with enough resources to be able to make themselves appear to be the most popular.
Published polls level the playing field. Information is power and there is nothing more anti-democratic than restricting information access to insiders and elites.
In an election, it is easy to lie with statistics. It is even easier to lie without them.
Damian Loscher is managing director of Ipsos MRBI which conducts polling for
The Irish Times