Wednesday, 19 October 2016

Findings from a Voting Advice Application in Georgia: Ideology matters by Dr Jonathan Wheatley

On 14 September, the Preference Matcher Consortium,
of which Oxford Brookes University is a member, launched a Voting Advice Application (VAA) in the run-up to the Georgian parliamentary elections of 8 October.
VAAs are web platforms that allow users to compare their positions on a set of concrete policy issues with those of parties or candidates in elections. They then provide their users with visual displays showing their positions with respect to parties or candidates. The Georgian VAA Xmamkvlevi, which can be loosely translated from the Georgian as “Vote Survey”, asked Georgian voters to position themselves on thirty issues using the categories “completely agree”, “agree”, “somewhere in the middle”, “disagree”, “completely disagree” and “no opinion”. It then matched their responses with the positions of the eight largest political parties in Georgia. These positions had already been determined by a panel of experts from Ilia State University in Tbilisi. The issues related to economic policy, relations with the West and Russia, the role of the Church and gay rights, amongst other topics. Xmamkvlevi provided its users with three displays: (1) a bar chart comparing users' proximity to all parties based on all thirty VAA items; (2) a two-dimensional map, based on users' responses to selected items that were deemed to “belong” to a particular ideological dimension (economic left versus right and liberal versus conservative), and (3) a bar chart that indicated how other users who answered the thirty items in a similar way intend to vote. This third feature depended on another attribute of Xmamkvlevi; by means of supplementary questions it collected (anonymous) data from users on age, gender, education, political interest, party identification and vote intention. These supplementary questions had options that included “I do not intend to vote”, “I am undecided” and “I prefer not to say”.

The purpose of VAAs such as Xmamkvlevi is twofold. First, they help undecided voters decide how to vote and, in less established democracies such as Georgia, provide voters, politicians and political parties with incentives to consider matters of policy. In former Soviet republics such as Georgia, scholars have often assumed that voters are either to be drawn to charismatic leaders (of which Georgia’s former president Mikheil Saakashvili is a prime example) (Enyedi 2006) or simply vote for the party or leader that will provide (or promises to provide) material goods to their village or family (Stefes, 2006). The normative rationale of a VAA is to contribute to competitive and programmatic politics by counteracting this tendency. Promoted through advertising on Facebook, Xmamkvlevi was accessed by a relatively large number of voters. According to Google Analytics, 13,510 users accessed the site and meaningful data were obtained from 10,531 users. This should be considered a rather large number given that the number of registered voters in Georgia is just 3.5 million, VAAs are not well-known and Internet use in Georgia, although growing fast, is far from universal. The relatively high uptake rate suggests that policies matter for Georgian voters and this sends a clear message to Georgian politicians that their policy positions (insofar as they have any) will either appeal to or repel voters.

Second, VAAs provide the researcher with a wealth of data on voters' political preferences. Provided due care is taken to control for the fact that the data is self-selected, and therefore not representative, this data can help us to address some of the core research questions in political science. In this respect Xmamkvlevi has provided some fascinating insights into whether Georgian voters choose parties according policy or ideology and, if they do, what are the main dimensions of political competition that divide them. Drawing on a subsample of approximately 4,500 users from the Xmamkvlevi data that were representative of the population as a whole in terms of vote intention (using “abstainers” and undecided users as a proxy for the 49% of the population who did not turn out to vote), I carried out a dimension reduction technique called Mokken Scale Analysis to identify latent political dimensions. This showed that a single ideological dimension appeared to split Xmamkvlevi users. One pole of this dimension is represented by those who take a conservative position on social-moral issues such as gay rights, believe in a greater role for the Georgian Orthodox Church, oppose integration with the EU and the West more generally and adopt a statist position with regard to the economy. Those that cluster around the opposite pole, on the other hand, are more liberal on social-moral issues, feel that the Church already has enough power, strongly favour integration with the West and support free market economic policies.

Interestingly, this dimension differs radically from the traditional left-right dimension that is used to explain political competition in Western Europe. In Georgia social conservatism is associated with the economic left, while a liberal or progressive position on moral-cultural issues is associated with the economic right. This largely conforms to the findings by Gary Marks and his colleagues that the economic and cultural axes of political competition correlate in a different way in much of post-communist Europe than in Western Europe (Marks et al., 2006). A possible explanation for this is that the multiple transformations that many post-communist societies have undergone in the past twenty-five years as a result of economic transformation from a state-run to a market-run economy and more recently of both economic and cultural globalisation has created a divide between “winners” and “losers” of globalisation that is even deeper than that identified by Hanspeter Kriesi and his colleagues in western Europe (Kriesi et al., 2006). According to this logic, the so-called “winners” embrace the liberal economic and cultural changes that globalisation brings, while “losers” reject these changes.

Even more interesting is the fact that those users who self-identified as party supporters by declaring that they identified with a particular party and that they were going to vote for that party occupy distinctive niches along that dimension. The graphs below show density maps of the supporters of the three parties that at the time of writing appeared to have overcome the 5 percent electoral barrier required to enter parliament through the party lists (party lists provide 77 out of 150 seats, while the remaining seats are elected by means of a majoritarian system over up to two rounds). These three parties are the governing Georgian Dream party (GD in the graph), which appears to have won a majority in parliament, the main opposition United National Movement (UNM, in second place) and the small pro-Church Alliance of Patriots of Georgia (PAT). The graph clearly show that the three groups of party supporters have common ideological characteristics with supporters of the Alliance of Patriots of Georgia closest to the (economic) left, conservative pole, supporters of the UNM closer to the right-liberal pole and supporters of the governing party rather closer to the middle, but tending towards the left-conservative pole. This suggests that Georgian politics is not only a struggle for power and resources, but that ideology also matters, especially amongst the rather well-educated sectors of society that used this online application.

No comments:

Post a Comment