A lot has been made about the supposedly high levels of undecided voters in recent polls published in Ireland, with some even suggesting that this points to the desire for a new party, but is this really the case?

Firstly, it is important to make a distinction between an undecided voter and someone who states that they will not vote at all. At the last election turnout stood at 69%, which confirms that 31% of all eligible voters did not make it to vote. So to get an accurate result it is important we take this into account in our polls, or we will be including the opinion of people who will never get to the polling booth.

In RED C polls we start by asking voters how likely they are to vote at the next election on a 10 point scale where 1 is not at all likely and 10 is extremely likely. We then remove those who suggest they are between 1 and 3 on this scale, as it is highly likely that these people will not go and vote. In February 2011 just before the General Election, 9% of those polled suggested they were unlikely to vote (1-3 on the 10 point scale). While this is somewhat less than the 31% who didn’t vote, removing them at least takes likelihood to vote into account. It is also the case that those who take part in polls are somewhat more likely to vote, so we would never end up with 31% saying they wouldn’t vote.

The proportion who stated they were unlikely to vote in our most recent poll in January 2013 was 12%, so there hasn’t been a significant change in this number over time, although it does of course fluctuate based on how strongly people feel about politics at the time.

The other group excluded from our core voting figures are those that are classified as “undecided” at this time, how they will vote. These people are very different to those that say they won’t vote. If an election was held tomorrow, it is likely they would go to the polling station and vote. At the moment however, they either suggest they don’t know who for, or alternatively refuse to tell us how they will vote. It has been shown through post election analysis in the past that in fact they frequently end up voting for the party they supported at the last election, as humans are creatures of habit. But it could also be that they don’t want us to know how they will vote, and so tell us they don’t know or refuse to answer.

To combat this RED C re-allocates half of the people who claim they don’t know how they will vote to the party they supported at the last election. This group of real undecided voters, tends to reduce as you get closer to an election when people make up their mind, and gets larger when elections are further away and people are not ready to make an informed or considered opinion. We have to remember of course that for many people, how they will vote is really only thought about when an election presents itself, rather than something they are interested in all the time.

Further confusion on this topic comes around the fact that different polling companies use different methods to classify undecided voters. Ipsos MRBI do not analyse by likely voters at all, and so end up with quite high levels of what they call undecided’s, which would of course include those who wouldn’t vote. B&A remove ineligible voters first (who would be in our “would not vote” category) then publish undecided voter levels on this sample, but after this they also remove unlikely voters in their adjustment.
The fact is however that our most recent poll saw the levels of real “undecided” voters stand at 20% of all likely voters, which equates to about 18% of ALL adults aged 18+. This current level of undecided voters is quite high, but we have seen it at or close to this level in the past (see table below), while just before the last General Election it was still standing at 15%.

So in total immediately before the last election we had about 9% who wouldn’t vote and about 14% undecided, or in total 23% not included in the core figures. In our most recent poll we have 30% not included.

This and the table above do suggest a slight increase in undecided voters over time, but not a significant jump. As such there is little evidence from this data that people would like to see a new political party. That is not to say that this desire is not out there, simply that judging this by the number of undecided voters, is not evidence enough on its own.