Compared to western democracies, the polling industry in Georgia is relatively young. Our very first poll results didn’t start appearing in the media until 1990. However, the industry isn’t so young that it shouldn’t be able to accurately predict the winner of an election. But in order to accurately forecast voter behavior, two basic things need to be in place: A) you need to have a knowledgeable polling organization, and B) you need a fair and free pre-election period. Both professionalism and freedom work hand-in-hand in this case.
In established democracies, major newspapers and television networks commission survey firms to conduct electoral polls. In addition, there are also several universities, think-tanks and political parties who invest in polls prior to elections. Many of these polls are made public, and in most instances, they accurately predict the outcome of the elections. This is largely due to the logical conjunction that states “A and B” is true only if A is true and B is also true.
Over the last 10 years in Georgia, most large-scale pre-election polls have been sponsored by international donor organizations, and conducted by local survey firms. As such, GORBI has extensive experience in undertaking electoral polls both before and after elections in Georgia, and in several post-Soviet countries for American-based universities and other clients. As a pioneer of commercial opinion polling in Georgia, I can’t recall any instance when the government has interfered directly in GORBI’s work. However, in other countries where we have worked, this was not always the case. In fact there have been instances where state security officials have either requested that we omit “sensitive” questions (you can’t ask presidential job approval questions in Uzbekistan or Belarus), or where they arrested data collectors and raided the offices of survey research firms.
In Georgia, when it comes to electoral surveys or exit polls, the public tends to harbor an attitude of skepticism, and this attitude is not always without grounds. As mentioned above, if elections are conducted within a fair and free environment, and votes are not illegally stolen by one or the other party, there is no reason that polling firms can’t accurately predict election outcomes.
Pre-election polls carry with them two main weights – forecasting election outcomes, and understanding voter behavior. In general, the first weight is most important, as the life and reputation of the survey firm rests on this. A failed prediction could result in the death of a polling firm. In other words, when a competitive environment for voting exists, those who are trying to predict the election’s outcome can be likened to mine seekers – they can only make a mistake once.
The 2012 Parliamentary Elections in Georgia represented one of the most interesting cases for polling firms when logical conjunction did not work. This was because at least one component in formula (B) failed to occur.
In the meantime, by using the power of post-election surveys, we can assess the transparency of any election. For this column, I will use the data gathered by a GORBI survey immediately following the 2012 October Parliamentary Elections. For readers not familiar with the current political history of Georgia, I will say this – with regard to the 2012 elections, Georgia avoided what could have easily been another civil war or situation that led to public unrest.
Here is why:
Most of the pre-election polls failed to predict the winner, and this happened despite the fact that the winner collected significantly more votes than the ruling party at the time. On Election Day, the results of several exit polls conducted by both reputable and unknown survey firms differed greatly.
Election polls are used to predict election outcomes and interpret the meaning of the results. There is a third dimension in Georgia however – assessing the transparency of the elections (B). And as illustrated in Table 1, there was a large gap between the official results and the post-election survey data, as well as some exit poll outcomes. This is a clear indicator that “something was rotten in the state of Georgia”. Thankfully, Georgia avoided another disaster, as power was handed over peacefully to the opposition party.
While there are a plethora of online casinos accessible to Georgians, political betting markets are uncommon here, and this is perhaps because until recently, it was very risky to invest in this business since there was no way of accurately calculating what each political horse should have as a stake/share to bet on.
Creating a fair political environment would lead to a quantum jump towards accurately predicting election outcomes and analyzing voter motivation. This already happened during the 2013 presidential elections. GORBI was able to accurately predict (and this was televised prior to the actual elections) that not only was a second round of elections not going to be necessary, but also the percentage of votes the top-four candidates would receive. We were also accurate in our prediction of the 2014 local elections. Both became possible because part B of the formula was working.
Note: If there is a close race in the election, the accuracy of the prediction will be at the mercy of the voters list, which is sometimes of marginal quality, and will also be contingent on the official statistics used by pollsters for modeling and weighting the final poll.
GORBI is a regional hub for partner organizations and international clients. Since 2003, GORBI remains an exclusive member of Gallup International research network for its two decades of experience in survey research in post-Soviet Union countries, as well as Mongolia and Iraq. This data was provided exclusively to the Financial.
Discussion about this post