Opinion: the scandal around fake polling has highlighted just why and how Irish parties engage in constituency polling

In the aftermath of the scandal about political parties conducting fake polling, it's perhaps worth delving into the world of constituency polling in Ireland. What is it? And why do political parties do it? The principle reason why constituency polls are important is because individual candidates have a relatively strong influence on voting behaviour in Irish elections.

We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences

From RTÉ Radio 1's Today With Claire Byrne, Hugh O'Connell from The Irish Independent and former Fianna Fáil advisor Derek Mooney in the political party polling controversy

The Irish National Election Study surveys voters after each general election. One question it asks is whether the voter would vote for the same candidate if that candidate had run for a different political party. Across the past five elections roughly, one third have stated that they would follow the candidate to another party, one third that they would not and a final third say that it would depend on the party. This is partly a consequence of our electoral system and is underlined by the relatively success of independent candidates, which are extremely rare in other countries.

Interestingly, as we see below, the role of individual candidates at the most recent general election is stronger for some parties than others. Support for insurgent parties such as Sinn Féin and the Green Party were based more around the party brand. This shouldn't be a surprise and reminds one of the story that the Sinn Féin candidate who topped the 2020 poll in Kildare South had been on holiday during the final week of the campaign. At the other extreme, the drubbings that Fianna Fáil and Labour faced in 2011 and 2016 respectively reduced both parties to a support level based around the strength of specific candidates.

Parties and candidates: "which, for you, is more important in deciding how you just cast your vote - the party (or the fact that the candidate was independent) or the candidate him or herself?" Broken down by party of choice.

Constituency polls are primarily used by political parties to determine their local candidate strategy. That involves testing the popularity of specific candidates (and encouraging or discouraging individuals to stand) as well as deciding how many candidates to stand.

The effect of standing too few candidates should be obvious, as Sinn Féin found out in the most recent general election and on fewer seats in the Dáil as a result. However, one of the vagaries of our electoral system is that a political party standing too many candidates can negatively affect the party's chances of winning seats. The idea is that that transfers from an eliminated candidate are never as efficient as it would be if the party stood fewer candidates.

One example of this is Cork South-West in 2011, where the two Fianna Fáil candidates together achieved 94.5% of a quota. This is normally well in excess of what would be required to win a seat. However, neither candidate was elected: once Christy O'Sullivan was eliminated, Denis O'Donovan only picked up 57% of Christy’s vote.

We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences

From RTÉ Radio 1's News At One, Kevin Cunningham discusses the fake pollsters story which has ensnared Sinn Féin, Fine Gael, Fianna Fáil and the Green Party

There is some academic debate around this logic. It may also be argued that the inefficiency is made up by the candidate’s activities boosting the party’s brand and what we observe as inefficiency is merely the personal vote of a candidate finding its natural home. However, the statistical evidence does suggest that standing too many candidates will reduce the likelihood of getting candidates across the line, all else being equal.

So how reliable is a constituency poll? Well, there are a number of reasons why a poll can deviate from the outcome. The first is that we are looking at a mere sample of voters so naturally there will be a difference between the people we speak to and the population owing to random chance. From the three TG4/Ipsos MRBI polls conducted in Donegal, Kerry and Galway West during the short campaign of the 2020 election, the average difference between poll and the result was 3.4% (for candidates winning/polling over 3% of the vote).

A second source of deviation is in the time between the poll and the outcome itself. Polls are frequently referred to as reflecting a 'snapshot in time’. Indeed, at the last general election, there was a dramatic change in support for Sinn Féin in January, so any poll prior to that would reveal a significant systematic deviation. Analysis by Gail McElroy and Michael Marsh of the 34 constituency polls published ahead of the 2002 general election revealed that the polls conducted in advance of the short campaign deviated from the outcome to a much greater degree.

We need your consent to load this rte-player contentWe use rte-player to manage extra content that can set cookies on your device and collect data about your activity. Please review their details and accept them to load the content.Manage Preferences

From RTÉ Radio 1's Drivetime, law lecturer Jennifer Kavanagh from the Waterford Institute of Technology on Sinn Féin's use of a fake polling agency name when conducting surveys

There is one additional and important source of uncertainty which owes to the fact that first preference vote share is not an immediate indication of who is likely to win the seat and transfers bring an additional source of uncertainty. For example, Katherine Zappone was elected in Dublin South West in 2011 having won 6.6% of the first preference vote, while Anne-Marie Dermody of Fine Gael was not elected even though she won 9.6% of first preference votes.

All this uncertainty reduces the overall value of an opinion poll and makes it more difficult for political parties to pay a commercial rate so it's not surprising that they have sought to replicate the work of a professional polling company. One limitation of doing so relates to something called 'social desirability bias’. This is something that is particularly acute in face-to-face polling. What social desirability bias means is that a number of those who answer the survey will be inclined to give an answer closer to what they suspect the interviewer might want, as opposed to their truer feelings.

It has been previously noted that this bias influences responses on taboo topics such as sexual activities, illegal behaviour such as social fraud or unsocial attitudes such as racism. It is for this reason that political parties conducted polls not on behalf of 'Sinn Féin', 'Fianna Fáil' or ‘Fine Gael’ but rather fabricated entities such as ‘IMRA’, ‘Pinpoint’ and 'PRAI'.

Of course, the big problem of this is the deceptive and organised behaviour of political parties

Working for the British Labour Party, we observed such differences between our canvassing returns and private polling. Canvassing would be conducted by volunteers and politicians representing the party and would notably have a much higher level of support for the Labour Party than our private polling would reveal. Closer to home, Garret FitzGerald's autobiography records that one of the few ‘well-founded’ pieces of advice he received in his first election was that he would win 90% of the votes he would consider to be ‘certain’.

Having been frequently on both sides of the arrangement in both commissioning polls and conducting them, it is clear that political parties do need to use a private entity to have any level of reliability and ascertain what is somewhat limited information. Of course, the big problem of this is the deceptive and organised behaviour of political parties. The 2020 UCD exit poll revealed that the voting public disagreed with the idea that ‘most politicians are trustworthy’ by a two-to-one majority. Given that those that do not vote are likely to have lower levels of trust, Irish democracy is more fragile than it might seem above the surface and activities such as fake polling do not help.


The views expressed here are those of the author and do not represent or reflect the views of RTÉ