Speaking at last weekend's 'anti austerity' march in London Martin McGuinness said 'The Tories were elected on the lowest share of the vote ever for a Tory government - 24% is no mandate for austerity'.
First error with this is he should have said 'share of the electorate' not 'share of the vote' - The Conservatives got a 37% share of the vote (though it could be said that 37% isn't much of a mandate either!).
However, the key question is this - is it correct to spin the election result in this way and quote % share of the whole electorate rather than just of those that voted?
In terms of polling the normal convention is to present the main 'headline' poll results excluding the No Opinion/Non-Voters/Don't Knows. This is the way all the UK and Irish national polls are presented - we hear it regularly e.g. currently the Conservatives and Labour are around early to mid 30's in the polls, Ukip on 11-14% etc., and all these poll scores are excluding the No Opinion/Non-Voter/Don't Know's. It's presented this way because in a real public election the 'votes' of the non-voters aren't counted, Why? - because they haven't voted.
Yes, of course it's important to reference No Opinion/Non-Voter/Don't Know's in a poll, and to note the level of non-voters at an election, because this gives some sort of measurement of apathy, particularly if the figure gets particularly high. In polling terms this would be around 45% or above. However this figure i.e. the No Opinion/Non-Voters/Don't Knows should always be quoted separately, as structurally they are a different group.
Once a person has expressed a preference e.g. say in a poll for DUP, Alliance etc., then they are recorded as a 'vote' in terms of forthcoming elections. They have taken part in the poll, and they have chosen to take part, in a process similar to what a voter does when they go to a polling station in a real public election. We now know something about them - they support a particular party and intend to vote (i.e. at this time) for that party in forthcoming elections. If someone wants to be recorded as a 'No Opinion/Non-Voter/Don't Know' in a poll, or doesn't vote in an actual election, then we don't have that information and they could be choosing that option for one of several different reasons. Non-voters are a complex grouping, and are made up of many separate parts.
Some people assume that all non-voters at elections are sort of voting against the system, disillusioned with politics, and therefore we can make assumptions about them and sort of count their non-votes as votes for something e.g. anti-austerity. But non-voters, and those who don't want to express a preference in polls, are not all apathetic or disillusioned about politics. Indeed some of these non-voters could get very annoyed if presumptions were made about them, with cries of 'How you dare you analyse my non-vote - go away'!
Yes there is a growing group of non-voters who say they are disillusioned with politics, and would like something different. However by non-voting in a public election, and/or to a lesser extent non-voting in an opinion poll, they have opted out. If they are truly disillusioned as to what's on offer in terms of politics then the solution is to get in there, join a political party, start a new political party, stand for election, join a lobby group, start a lobby group, and basically get up on a soap-box somewhere and blow their own trumpet. If they just stay at home on election day and moan, or say 'No Opinion' to a pollster, then they've only themselves to blame.
In the same way that you can't put non-participants in a poll on the same status level as those who have expressed an opinion or preference in relation to a particular poll question, you can't put non-voters at a full public election on the same level as voters. Once you choose to not vote at an election, or to not express a view to an opinion pollster - that's it - you're out of the game. And Martin McGuinness, polling companies, political parties, media, and others, shouldn't start allocating these 'non-vote's' to one particular party, cause, or group (e.g. anti austerity). This would be presumptuous, dangerous, and from an elections and polling science point of view is wrong. You can't count non-voters on the same status as voters. Non-Votes are not Votes. Martin McGuiness should note this.
Bill White is Managing Director of Belfast polling and market research company LucidTalk, polling partners to the Belfast Telegraph. You can follow LucidTalk on Twitter: @LucidTalk.