Hix, S. Hortala-Vallve, R. and G. Riambau-Armet (2017): "The Effects of District Magnitude on Voting Behavior", The Journal of Politics 79 (1), 356-361. DOI: 10.1086/688889

In this short article, Hix et al. (2017) analyse the effect of district magnitude on voting behaviour with a laboratory experiment. The authors test the widely held expectation that the proportionality of the electoral system causes (in)sincere voting and that voters' likelihood to vote for their preferred party is higher as district magnitude increases. In their experimental set-up, 212 participants were assigned to a single treatment group corresponding to a different district magnitude (DM= 1, 2, 3 or pure proportional) and casted votes in 60 elections amounting to a total of 12,270 observed votes. Each participant was given a utility from each of the five parties running in the election and casted a single vote in each election. Then, participants were announced the election results and her respective, hypothetical profit in such election. New party preferences were assigned after five elections. At the end of the last election, each participant received a payment corresponding to his/her valuation of the party of each candidate who was elected by the overall group in four randomly selected elections[1]. The findings of the experiment show that in line with pervious contributions, sincere voting increases with the proportionality of electoral system. If the experiment simulated a larger district magnitude, participants more likely found that their most preferred party promised them the highest expected utility. However, while a district magnitude of 1 or 2 lead to similar levels of sincere voting, this behaviour strongly enhances as district magnitude is enlarged to 3. The experiment also uncovers a frontrunner effect, showing that a considerable share of participants voted for the winner of the last election irrespective of the district size and the utility of this party according to their personal preferences.

This article uses an interesting research design aiming to isolate the psychological effect of district magnitude on voting behaviour from factors such as voters’ preferences or the number of parties. Most of the article’s findings side with previous quasi-experimental studies and thus provides further evidence for causality. However, some of the surprising results pose the question, to what extent an experimental research design based on monetary incentives reflect the decision-making process of voters within their natural environment. In real elections, voters might not prioritize according to financial gains (which besides tend to be hard to calculate for ordinary citizens). They might cast votes based on ideological stance, general policy preferences or on descriptive cues like gender or ethnicity. The artificial setting could as well create the frontrunner effect, about which the authors say that they “do not fully understand the psychological motivations of such behaviour“ (p. 361). The rationale behind the frontrunner effect might stem from participants who aimed at increasing their monetary remuneration irrespective of district magnitude or candidate characteristics. I wonder whether and how one could adapt the experimental setting to embed the complex combination of intrinsic and rational incentives that voters face in the real world.