There are costs and benefits to rotating the names of candidates on ballot papers

There is a small but comprehensive literature on the impact of candidate and party position of election outcomes, with a consensus that appearing in particular positions a ballot paper can influence voter choice. Here, Kamil Marcinkiewicz adds to this literature, arguing that while name position does have an impact, it is not a uniform phenomenon, with different impacts being felt in different electoral systems and indeed countries.

Credit: Ministry of Defence, CC BY NC ND 2.0

Credit: Ministry of Defence, CC BY NC ND 2.0

In my research article published in the “Electoral Studies”, I discusses the impact of ballot position on performance of candidates in the elections to the lower house of the Polish parliament, the Sejm. In this piece, I demonstrate the existence of three types of ballot position effect. The most pronounced of them is the benefit of appearing at the top of the ballot. It corresponds with the phenomenon long observed in the survey research. Respondents confronted with a lengthy list of items tend to pick the one which appears in the first position. A similar mechanism applies to the ballot papers which include in Poland up to 40 names per party.

The second type of the effect results from the spillover of the first position benefit. It means that proximity to the top of the ballot paper improves a given candidate’s performance. The second-placed candidate receives a higher share of votes within his or her party than the third-placed, whereas the third-placed is supported by more voters than the fourth-placed. This pattern is observed among all other positions, except for the last one. Thirdly, appearing at the bottom of the ballot paper provides Polish candidates with a weak benefit as compared to appearing in the second-to last position. The existence of this phenomenon indicates that the voters tend to pay more attention to the margins than the centre of the list of items.

Despite identifying a significant impact of the ballot position on performance of a candidate within her or his party list, I do not advocate the introduction of name rotation in the Polish case. The reason for this is the beneficial role of the ballot position plays in the Polish electoral context. Poland uses the open list proportional representation with a single obligatory preference vote. This means that candidates compete not only between, but also within the parties. Due to appearance of different party lists on separate sheets of the ‘ballot booklet’ handed out to the voter at the polling station, the danger of supporting a candidate from the ‘wrong’ party (i.e. accidentally picked by a voter) is marginal.

Furthermore, in the Polish case, the party leadership decides who appears in the positions perceived traditionally as focal (e.g. top, second, last). Assignment of a candidate to a given slot is hence a signal for the voter conveying information on quality of a candidate in the eyes of party leadership. Introduction of a rotation mechanism would hence deprive worse informed voters a parsimonious strategy of selecting the optimal vote option. Since it is impossible to cast an overall party vote (as opposed e.g. to the Czech Republic), the efficiency of usage of a vote by low-information voters might decrease if the name order on the ballot was random.

What might be the implications of the aforementioned study for the elections in the United Kingdom? The seats in the House of Commons are distributed according to the plurality rule in the single-member constituencies. The consequence is that every party places only one name on the ballot paper. The candidates are, thus, perfectly distinguishable from each other based on their party affiliation. The fact that they appear in the alphabetic order may, but does not have to, be problematic.

The studies comparing performance of candidates in the British parliamentary election and the Australian federal election in the 1970s and early 1980s did not find evidence of ballot position effect in the United Kingdom. It was shown, however, that ballot position biased the results of election in Australia, which at that time also used alphabetic name ordering and where voting is compulsory. After 1984 Australia adopted a system of random assignment of candidate to a ballot position, which however, still did not eliminate the problem of ballot order completely. Researchers suggest therefore adoption of the name rotation in Australian federal election.

The research on ballot position effect in the United States produced evidence for the significant bias in favour of higher placed candidates in the primaries and local elections. The findings related to national and state elections are more ambiguous. Even though it is widely assumed that benefits of ballot position in high-visibility elections are small, many US-states rotate candidate positions on ballots also in state and national elections.

Alphabetical ordering of parties in the European Parliament elections also does not have to necessarily bias the result. Due to the usage of the closed list system in the EP elections, a voter has to merely make a decision between parties, not individual candidates. The list of parties is not exceedingly long and party labels are usually informative. Since voting is not compulsory, it is unlikely that the magnitude of the bias caused by indifferent voters casting a ballot for the first option on the list may be sufficient to have significant impact on the result.

Both under the closed list PR system and in case of the elections in single-member districts rotation of candidate or party names may be beneficial and is highly recommended when voting is compulsory. It generates, however, also some costs. The cost of producing ballot papers with many different forms of layout is nowadays a minor problem, especially when voting computers are in use. A more serious issue is increase in the cost of making a decision faced by a voter. Inspecting a ballot with rotated name order in search for a candidate of one’s choice may pose a challenge for some sections of electorate (e.g. elderly citizens).

It may be argued, however, that both under the closed list PR and when single-member districts are in use the benefits of rotation exceed the cost. Under the open list PR, as used in the election to the Polish Sejm, the number of vote options may be overwhelming for an ordinary voter (choosing a party and then an individual). Non-random positioning of candidates reflecting preferences of the party elite may then offer a useful heuristic shortcut.

Note: this post is based on the article ‘Electoral contexts that assist voter coordination: Ballot position effects in Poland?’, Electoral Studies, 33(1), pp. 322-334

Note: this post represents the views of the authors and not those of Democratic Audit or the LSE. Please read our comments policy before posting. The shortened URL for this post is:

Marcinkiewicz_Kamil_photoDr Kamil Marcinkiewicz is a lecturer in quantitative methods in political science at the University of Hamburg, Germany. His full academic profile page can be found here. 


Similar Posts