According to the NDI polls 37% of likely voters say that they would vote for the national movement and 12% say they would vote for the Georgian Dream. Does this tell us anything about actual likely election results and, if so, what?
Obviously, no-one sensible, particularly NDI, would suggest that Georgian Dream will only get 12% of the national vote but that does not mean that the poll should be ignored.
22% of those polled said they did not know how they would vote and 21% refused to answer. Therefore any attempt to make predictions requires that we make some assumptions about how to assign these voters.
There are a number of ways that these numbers can be apportioned. First, one can simply assume that all 43% will eventually vote in the same proportions as those who DID answer. The ratio of NM:GD:Other is about 37:12:8. So, on that simple mathematical apportionment you would expect a final result of 65% for the National Movement, 21% for Georgia’s dream and 14% for ‘other’. Clearly, this would be a very bad result for Georgian Dream.
But does that seem a reasonable approach? Many people assume that the overwhelming majority of ‘refuse to answer’ must be supporters of Georgian Dream. Otherwise, why refuse to answer? This seems particularly reasonable given that NDIs poll has a DK/RA level that is about 20% higher than most of the other polls, and that it gives a far higher proportion of the vote to the National Movement than any of the others.
If that is true and you assign all the 21% of the refuse to answer entirely to the opposition, and then use the simple apportionment for the rest., then we would get National Movement 47%, Georgian Dream 31% and other 19%.
Third, one can assume that all the 'dont know' AND 'refuse to answer' are supporters of the opposition. That does not, however, give an extra 43% of the vote to the Georgian Dream since, in the NDI poll, 8% go to various 'other' categories. So, distribute the 43% according to these proportions then Georgian Dream would get 38% of the national vote.
Finally, if one assumes that all of the ‘don’t know’ and the ‘refuse to answer’ are really secret supporters of Georgian Dream (and not just opposition supporters), they will get 55% of the vote.
What is one to make of these different options? There are two other sets of numbers in the NDI polls that seem to give us some opportunity to see which of these scenarios might apply. First, one can look at the previous NDI polls and see what happened to the ‘don’t knows’ and ‘refuse to answers’ in those.
NDI offer a side by side comparison between their polls results and the actual election results; in the 2010 Sakrebulo elections nationally and the 2010 Tbilisi Mayoral elections.
In the case of the national polls for the Sakrebulo elections 19% did not know, and 10% refused to answer and, on election day, this 29% seems to have apportioned proportionally across the parties roughly in line with the polling. If this same logic applied in the parliamentary elections then the results would be those found in the first scenario with 65% for national movement and 21% for Georgian dream.
In the case of the Tbilisi Mayoral election only 6% refused to answer and 14% were undecided, but in this case almost all of this 20% went to the opposition. As Alasania has pointed out, he was polling at 7% of the vote, but actually won 19%, Chanturia polled at 4% but got 11% and Dzidziguri polled at 2% and got 8%. Note that in the Tbilisi Mayor’s election, this re-assignment was not big enough to affect who won. But, if the same thing happened in the upcoming parliamentary elections we would be closer to the third or fourth scenarios and possible outright victory for Georgia’s Dream (ignoring the effect of majoritarian voting).
So, looking at previous polling does not seem to help.
One other option for how to assign the undecided/refuse to answer is to look at how they answered other questions. NDI provides the answers to ‘satisfaction’ questions, broken down by future parliamentary vote. When asked if the country is ‘going in the right direction’, 39% of undecided and 37% of ‘refuse to answer’ say that it is ‘not changing’ or ‘going in the wrong direction’. When asked if ‘the Government is making changes that matter to you’ then 33% of ‘don’t knows’ and 34% of ‘refuse to answers’ say ‘no’.
This may be a good window into voting intention, as the National Movement supporters are overwhelming positive and Georgian Dream supporters are overwhelming negative in their answer to both questions. If this is correct then we are, once again, in the same territory as the first scenario – since the first scenario also apportions approximately 1/3 of the ‘don’t know’/’refuse to answer’ to the opposition. This is, again, an overwhelming victory for the national movement.
All of this does not bode well for the opposition. For them to come close to achieving a straight victory they would need almost all of the undecided and refuse to answer to side with them, even though 50-70% of this group who expressed general satisfaction with direction of the government and the country.
There is, however, a reading of these results that is more optimistic for the opposition. The election is still two weeks away and that is a long time in Georgian politics. If we interpret the 43% undecided/refuse to answer as a group who are genuinely undecided about who they will vote for then that makes the campaign incredibly important and it means that, even though they are behind, the opposition could still win.
If this opportunity is real, then to take advantage of it, they really need to start paying attention to polling in order to shape their message, rather than ignoring polling because they don’t like the results. The poll does offer ideas about where the opposition could be focusing their attention. While the polls show that the population is generally happy with the government and the direction of the country, it also shows that a significant proportion of the electorate have little idea about the economic plans set-out by either Merabishvili or Ivanishvili, though they are far more ignorant of Ivanishvili. They also show that the government/opposition are more or less equally trusted in a couple of important areas of policy.
If the opposition paid more attention to the polling, it would then allow them to run a more effective campaign and maybe even shift the undecided voters. Ignoring the polls, however, just because you don’t like what they say, is unlikely to help.