Voting experiments

AndrộBlaisã Jean-FranỗoisLaslier KarineVanderStraeten Editors Voting Experiments Voting Experiments ThiS is a FM Blank Page Andre Blais Jean-Francáois Laslier Karine Van der Straeten Editors Voting Experiments Editors Andre´ Blais Department of Political Science Universite´ de Montre´al Montreal, Que´bec Canada Jean-Franc¸ois Laslier Paris School of Economics Paris, France Karine Van der Straeten Toulouse School of Economics Toulouse, France ISBN 978-3-319-40571-1 ISBN 978-3-319-40573-5 DOI 10.1007/978-3-319-40573-5 (eBook) Library of Congress Control Number: 2016948875 © Springer International Publishing Switzerland 2016 This work is subject to copyright All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed The use of general descriptive names, registered names, trademarks, service marks, etc in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG Switzerland Preface This book contains 16 contributions that were presented in the workshop “Voting experiments,” organized by the editors in Montreal on March 28–29, 2014 This workshop, where presentations were selected after an open call for papers, was organized within the framework of the Making Electoral Democracy Work Project The Making Electoral Democracy Work Project is a 7-year research project, started in 2009 and funded by the Social Sciences and Humanities Research Council of Canada The goal of the study is to examine how the rules of the game influence parties and voters The study has three components: (1) an Internet survey of voters in 26 elections in five countries; (2) an analysis of party strategies in the same 26 elections; and (3) a series of experiments designed to complement the analyses of party and voter behavior The project, directed by Andre´ Blais, involves an interdisciplinary team of researchers from Canada, the USA, and Europe The experiments presented in this book are not intended to constitute a representative sample of all the voting experiments which have recently been conducted over the years by political scientists, psychologists, or economists Many chapters are coauthored by researchers who are members of the Making Electoral Democracy Work Project, and as such they reflect a particular interest in how rules and institutions affect voting behavior We have nevertheless tried in this book to offer a large perspective on voting experiments, aiming at showing the wide variety of experimental methods that can be used, as well as the wide variety of topics that can be addressed We hope that this book will be of interest to both experimentalists and scholars working on voting behavior and will inspire additional research We thank the anonymous reviewers who agreed to review the manuscripts submitted to this volume (two reviewers per manuscript), the University of Montreal for hosting the workshop, and Damien Bol for his extremely valuable help in organizing the workshop and editing the book We are also grateful to the Social Sciences and Humanities Research Council of Canada, the Centre for the Study of v vi Preface Democratic Citizenship (CSDC), the Centre d’E´tudes en Recherches Internationales de l’Universite´ de Montre´al (CERIUM), and the Center for Interuniversity Research and Analysis of Organizations (CIRANO) for their financial assistance Montreal, QC, Canada Paris, France Toulouse, France Andre´ Blais Jean-Franc¸ois Laslier Karine Van der Straeten Contents Introduction Andre´ Blais, Jean-Franc¸ois Laslier, and Karine Van der Straeten Part I Processing Information About Candidates/Voting Correctly Deciding Correctly: Variance in the Effective Use of Party Cues Jennifer L Merolla, Laura B Stephenson, and Elizabeth J Zechmeister 19 The Company Makes the Feast Party Constellations, Campaign Context and Issue Voting in Multi-party Systems Konstantin V€ ossing and Till Weber 43 Candidate Extremity, Information Environments, and Affective Polarization: Three Experiments Using Dynamic Process Tracing Mona S Kleinberg and Richard R Lau 67 Common Knowledge and Voter Coordination: Experimental Evidence from Mali Jessica Gottlieb 89 Part II Impact of Polls on the Decision to Vote or to Abstain Are People More or Less Inclined to Vote When Aggregate Turnout Is High? 117 Andre´ Blais and Rafael Hortala-Vallve Visibility and Sanctions: The Social Norm of Voting in the Lab 127 Aina Gallego, Carol Galais, Marc Guinjoan, Jean-Michel Lavoie, and Andre´ Blais vii viii Part III Contents Impact of Polls on Candidate Choice: Bandwagon Effect and Strategic Voting Experiments on the Effects of Opinion Polls and Implications for Laws Banning Pre-election Polling 149 Todd Donovan and Shaun Bowler Polls, Partisanship, and Voter Decision-Making: An Experimental Analysis 171 Shane P Singh, Jason Roy, and Patrick Fournier Coalitions, Coordination and Electoral Choice: A Lab Experimental Study of Strategic Voting 191 Annika Frede´n Patterns of Strategic Voting in Run-Off Elections 215 Karine Van der Straeten, Jean-Franc¸ois Laslier, and Andre´ Blais Strategic Voting and Personality Traits 237 Cengiz Erisen and Andre´ Blais Part IV Methodological Debate and Innovations Individual Behavior Under Evaluative Voting: A Comparison Between Laboratory and In Situ Experiments 257 Herrade Igersheim, Antoinette Baujard, Fre´de´ric Gavrel, Jean-Franc¸ois Laslier, and Isabelle Lebon Recruiting for Laboratory Voting Experiments: Exploring the (Potential) Sampling Bias 271 Damien Bol, Simon Labbe´ St-Vincent, and Jean-Michel Lavoie Measuring Perceptions of Candidate Viability in Voting Experiments 287 Simon Labbe´ St-Vincent, Andre´ Blais, Martial Foucault, Jean-Franc¸ois Laslier, Nicolas Sauger, and Karine Van der Straeten Electoral System and Number of Candidates: Candidate Entry Under Plurality and Majority Runoff 303 Damien Bol, Andre´ Blais, Jean-Franc¸ois Laslier, and Antonin Mace´ Through the Polling Booth Curtain: A Visual Experiment on Citizens’ Behaviour Inside the Polling Booth 323 Michael Bruter and Sarah Harrison Introduction Andre´ Blais, Jean-Franc¸ois Laslier, and Karine Van der Straeten Voting experiments are about voting, and the dependent variable is therefore obvious: voting behavior What are the independent variables? The answer is more complex Indeed, the process determining/leading to the vote decision can be a long and complex one First, voters have to form some opinions about the political candidates and their electoral environment In the course of the campaign, they are exposed to information about the candidates, some of this information being provided by the candidates themselves and some by the media, and social or family networks Through polls or discussions within their networks, they also receive some information about how other voters are going to behave: whether they will turnout or not, and which candidate or list they intend to vote for Eventually, they make a decision about whether to vote or abstain, and if they vote, about which candidate or party to support The objective of (most of) the voting experiments presented in this book is to understand how different facets of the voters’ political and social environment influence their voting decisions These experiments use a variety a methods—the book includes laboratory experiments, survey experiments and field experiments— but as experiments they all share a common denominator In order to study the impact of the environment on voting decisions, they vary in a controlled way some key aspect of the environment in which the voter makes decisions in order to detect and isolate its effect on voting behavior A Blais (*) Department of Political Science, Universite´ de Montre´al, Montreal, QC, Canada e-mail: andre.blais@umontreal.ca J.-F Laslier Paris School of Economics, Paris, France K Van der Straeten Toulouse School of Economics, Toulouse, France © Springer International Publishing Switzerland 2016 A Blais et al (eds.), Voting Experiments, DOI 10.1007/978-3-319-40573-5_1 318 D Bol et al Appendix: Game-Theoretical Equilibria In this Appendix, we compute the equilibria of a nine-player game, which mimics the laboratory experiment presented in this chapter The game is identical to the experiment except that it is supposed to be one-shot, whereas in the experimental sessions, we repeat the game three times with fixed positions for all players We first consider a streamlined version of the game (full information game), neglecting the uncertainty due to the random choice of 50 out of 90 voters Then, we study the game with uncertainty, taking into account this random draw of voters Below, we also discuss, in view of the results, the pertinence of the equilibrium approach In the full information game, all nine players know the payoffs with certainty If we consider that all 90 voters are turning out, the situation in which only the median player (i.e., the candidate located at position E) enters is a Nash equilibrium under both plurality and majority runoff This player would indeed defeat any other player in a pair-wise competition None of them should thus enter With a cost of entry of 1/5, as in the experiment presented in the chapter (to simplify the analysis, the gain from winning an election is normalized to one), there exist six pure Nash equilibria under plurality, and three pure Nash equilibria under majority runoff The three equilibria under majority runoff are: (such as in the chapter, the players are denoted A, B, C, I) – {E}: Only player E (the median player) enters Her payoff is of – (1/5) ¼ 4/5 – {D,F}: Players and enter with a payoff of (1/2) – (1/5) ¼ 3/10 – {C,G}: Players and enter with a payoff of (1/2) – (1/5) ¼ 3/10 It is easy to check that these three situations are equilibria It is more tedious to make sure that there is no other equilibrium; we achieved this with the help of a computer The three situations above are also pure strategy Nash equilibria under plurality However, there are three more equilibria under plurality, which involve three entering players with a payoff of (1/3) – (1/5) ¼ 2/15: – {B,E,H}: A symmetric situation with the centrist and two rather extreme players – {A,F,G}: A non-symmetric situation involving one extreme player – {C,D,I}: The mirror situation of the previous equilibrium It is interesting to observe that the reasoning that if there are five entering players or less my probability of winning is 1/5 or more, and thus that my entering cost is covered, is not sufficient It is true that if six players enter, then at least one of them has a chance of 1/6 or less to win, and should thus not enter If there are exactly five entering players, these players have a probability of winning of 1/5 only if they have equal chances As soon as they not have equal chances (and they never have), the probability of winning of at least one of them goes below 1/5 This(-ese) player(s) should thus not enter The situation where there are five entering subjects is thus really an upper bound (and a crude one) for rational entry Electoral System and Number of Candidates: Candidate Entry Under Plurality 319 Let us now consider the game with uncertainty on turnout If there were 89 participating voters instead of 90 (as in the full information game), the payoffs of the candidates would arguably be extremely close to those of the full information game, and therefore the equilibria would be the same Now, with 50 voters turning out, as in the experiment, the noise introduced in the game is more important To compute the payoffs in that case, we used computer simulations: with n independent random draws of 50 out of 90 voters, one can compute the average payoff of each candidate over these n draws By the law of large numbers, these payoffs converge to the exact payoffs when n becomes large We observed empirically that the payoffs not vary by more than % for n ¼ 10,000 draws In order to obtain the set of all equilibria in the game with uncertainty, we computed the payoff of each candidate for each configuration, and checked for each configuration if it was an equilibrium (this is the case if each candidate wins with a probability higher than 20 % and if once another candidate enters, she wins with a probability lower than 20 %) It appears that all the equilibria of the game with full information are still equilibria Moreover, there is an extra equilibrium in the game with uncertainty under majority runoff: {B,D,F,H} Besides, there exists no other equilibrium In the following paragraphs, we provide some insights on these results First, it is easy to show that the single-player (median player) equilibrium is still an equilibrium (under both plurality and majority runoff) If the other candidates not enter, player E is obviously right to pay the entry cost, as she will win In contrast, other players should not enter Consider player D, if she enters, she obtains the ‘turning out’ voters located at positions between and 40 (on the 90-point scale), while player E obtains those located at positions between 41 and 90 Player D wins if the number of her ‘turning out’ voters is strictly larger than the number of ‘turning out’ voters of player E, wins with probability of 1/2 if these numbers are equal, and loses otherwise Player D wins with an approximate probability of 12 % This probability is less than 1/5, which means that she should not enter The situation is similar (or even worse) for other players The single candidate equilibrium is an equilibrium of the game with uncertainty The equilibrium {D,F} is also still an equilibrium of the game with uncertainty (under both plurality and majority runoff) The expected payoff of player D and F is (by symmetry) (1–2) – (1/5) > 0, just like in the situation where all 90 voters are counted If player C enters, she obtains the ‘turning out’ voters located at positions between and 30, player D obtains those located at positions between 31 and 45, and player F obtains those located at positions between 46 and 90 Under plurality, the chances of player F are larger than 97 % Therefore, player C should not enter (under both plurality and majority runoff) Here again, the situation is even worse for the other players Similarly, {C,G} is an equilibrium of the game with uncertainty under both electoral systems In that case, the most dangerous challenger is player E If E enters, she wins with a probability smaller than % under plurality, and with probability % under majority runoff With three candidates, {B,E,H} is an equilibrium of the game with uncertainty under plurality, but not under majority runoff In this configuration, player B obtains the ‘turning out’ voters located at positions between and 30, player E 320 D Bol et al obtains those located at positions between 31 and 60, and player H obtains those located at positions between 61 and 90 As a result, each candidate wins with probability 1/3 under plurality, and this configuration is an equilibrium However, player E is much more likely to win under majority runoff, as she almost surely wins when she reaches the second round Under this electoral system, E wins with probability 67 %, whereas B and H win with probability 17 % each Hence, {B,E,H} is not an equilibrium, but it is not far from being one With four candidates, {B,D,F,H} is not an equilibrium of the game with uncertainty under plurality, but it is under majority runoff In this configuration, player B obtains the ‘turning out’ voters located at positions between and 25, player D obtains those located at positions between 26 and 45, player F obtains those located at positions between 46 and 65, and player H obtains those located at positions between 66 and 90 Under plurality, the two central candidates (D and F) win with probability % only, which explains that {B,D,F,H} is not an equilibrium However, under majority runoff, these central candidates are advantaged when they reach the second round, as they almost surely win It appears that players D and F reach the second round with probability 25 % and win with probability 24 %, whereas player B and H win with probability 26 % Hence, {B,D,F,H} is an equilibrium Note that this result is related to the noise introduced in the game: in the full information game for instance, players D and F receive a strictly lower number of votes than players B and H, and they never reach the second round Finally, with five candidates, one can wonder whether {A,C,E,G,I} is close to being an equilibrium of the game with uncertainty The computation yields a probability of winning for extreme candidates (A and I) of % under plurality and lower than % under majority runoff This configuration is thus far from being an equilibrium The equilibrium analysis leads us to conclude that there should be a low or moderate number of entering players: at most three under plurality, and at most four under majority runoff However, one should note that this is really a typical equilibrium reasoning If I think that few other players are running, it is not unreasonable to take a chance Suppose for instance that I am the median candidate and that I observe that in the past elections, four or five players were entering, it becomes very reasonable for me to enter under majority runoff References Abramson, P R., Aldrich, J H., Blais, A., Diamond, M., Diskin, A., Indridason, I H., et al (2010) Comparing strategic voting under FPTP and PR Comparative Political Studies, 43(3), 61–90 Blais, A., Labbe´ St-Vincent, S., Laslier, J.-F., Sauger, N., & van der Straeten, K (2011) Strategic vote choice in one round and two round elections: An experimental study Political Research Quarterly, 64(3), 637–645 Blais, A., & Loewen, P J (2009) The French electoral system and its effects West European Politics, 35(2), 345–359 Electoral System and Number of Candidates: Candidate Entry Under Plurality 321 Bouton, L (2013) A theory of strategic voting in run-off elections American Economic Review, 103(4), 1248–1288 Brusco, S., Dziubinski, M., & Roy, J (2012) The Hotelling-Downs model with runoff voting Games and Economic Behavior, 74(2), 447–469 Cadigan, J (2005) The citizen candidate model: An experimental analysis Public Choice, 123 (1/2), 197–216 Callander, S (2005) Duverger’s hypothesis, the run-off rule, and electoral competition Political Analysis, 13(3), 209–232 Cox, G (1997) Making votes count: Strategic coordination in the world’s electoral systems Cambridge: Cambridge University Press Dhillon, A., & Lockwood, B (2002) Multiple equilibria in the citizen-candidate model of representative democracy Journal of Public Economic Theory, 4(2), 171–184 Downs, A (1957) An economic theory of democracy New York: Harper Duverger, M (1951) Les Partis Politiques Paris: Colin Golder, M (2006) Presidential coattails and legislative fragmentation American Journal of Political Science, 50(1), 34–48 Guinjoan, M (2014) Parties, elections and electoral contests: Competition and contamination effects Surrey: Ashgate Jones, M P (1999) Electoral laws and the effective number of candidates in presidential elections Journal of Politics, 61(1), 171–184 Lau, R R., Patel, P., Fahmy, D F., & Kaufman, R R (2014) Correct voting across 665 33 democracies British Journal of Political Science, 44(2), 239–259 Osborne, M J., & Slivinski, A (1996) A model of political competition with citizen-candidates The Quarterly Journal of Economics, 111(1), 65–96 Reynolds, A., Reilly, B., & Ellis, A (2005) Electoral system design: The new IDEA international handbook Stockholm: International IDEA Shugart, M S., & Carey, J M (1992) Presidents and assemblies: Constitutional design and electoral dynamics New York: Cambridge University Press Strøm, K (1990) A behavioral theory of competitive political parties American Journal of Political Science, 34(2), 565–598 Taagepera, R., & Shugart, M S (1994) Plurality versus majority election of presidents: A proposal for a “double complement rule” Comparative Political Studies, 27(3), 323–348 Tavits, M (2007) Clarity of responsibility and corruption American Journal of Political Science, 51(1), 219–229 Van der Straeten, K., Laslier, J.-F., & Blais, A (2016) Patterns of strategic voting in run-off 679 elections In A Blais, J.-F Laslier, & K Van der Straeten (Eds.), Voting experiments Cham: Springer, pp 215–236 Van der Straeten, K., Laslier, J.-F., Sauger, N., & Blais, A (2010) Strategic, sincere, and heuristic voting under four election rules Social Choice and Welfare, 35(3), 435–472 Through the Polling Booth Curtain: A Visual Experiment on Citizens’ Behaviour Inside the Polling Booth Michael Bruter and Sarah Harrison The Puzzle Research in electoral behaviour relies, almost by nature, on a methodological contradiction Our main source of information is typically surveys, but surveys are about self-reporting, and which social psychologists criticise as having low internal validity If we want to use methods which are not fundamentally based on self-reporting, then we tend to use experiments, which we tend to think of as having limited external validity for a number of reasons Typically, our discipline tends to think of the limitations of surveys and experiments in generic terms rather than specifically in the context of electoral behaviour research, and what is more, it would not be unfair to say that many of us tend to pay as much attention to those reservations as we would to the terms and conditions of a new website we are signing up for (i.e not much) Yet, the situation of both many surveys and much of experimental research in the field of electoral behaviour raises issues of particular relevance, and which potential implications in terms of bias should warrant a far more profound attention than we usually give them Let us start with surveys Unless we completely dismiss the insights of psychology, the biggest issue with self-reporting is undoubtedly that we ask respondents to tell us things which they very well not know, as not everything about the vote is conscious We may remember the experiments of Fisman et al (2006) on speed dating, which, among other things, effectively show that when you ask people what they look for in a partner, they give you answers which absolutely not match the type of people whom they decide that they want to see again after the process of a M Bruter (*) • S Harrison Department of Government, ECREP Electoral Psychology Initiative, LSE, London, UK e-mail: m.bruter@lse.ac.uk © Springer International Publishing Switzerland 2016 A Blais et al (eds.), Voting Experiments, DOI 10.1007/978-3-319-40573-5_17 323 324 M Bruter and S Harrison full speed dating session.1 They might not be dishonest, they simply frankly not know what really attracts them in a mate Why would electoral mating be any different? The additional problem is related to timing By definition, surveys will be taken either before (sometimes a long time before) or after (again, sometimes, several hours or days after) a citizens has actually cast their vote While this may be dismissed as a mere time constraints in much behavioural research, in the context of electoral behaviour, it can be hugely problematic because we know that the moment people cast their vote is a very special moment Indeed, Bruter and Harrison (2014a) show that approximately 40 % of the American and French populations say that they have already changed their minds on who to vote for on the very day of a vote Indeed, if we look at the November 2012 referendum on children rights, it is even the case that nearly 80 % of the population say that they finalised their vote within a week of casting it Pre-election surveys may thus give us a biased understanding of how people behave electorally As for post-election surveys, of course, we know well that respondents rationalise their answer, especially if by the time they answer our questions, they know how the rest of the country behaved and who has won the election As for experiments, while much of the argument on the lack of external validity is based on sampling arguments, which frankly, should not matter that much in the context of the testing of causal processes (at least, as long as a sample is still diverse and no subgroup shows a significantly different reaction to a given stimulus), a much less discussed element has to with the fact that experiments—and particularly physiological experiments—still use contexts which typically feel very unnatural to participants, and that their prior knowledge of the experimental process (effectively hard to avoid for both practical and ethical reasons) may lead to a number of undesirable effects (including precisely desirability!) In that context, it is not entirely clear how existing methodological procedures would enable us to capture something which takes place at a specific time (when people vote), without purely relying on self-reporting, and using processes which would not be so obviously disruptive that participants would behave in a very artificial way So is there a solution? Let us be honest: what specialist of electoral behaviour hasn’t wondered what we would learn if we could put a little camera in polling booths and directly observe voters as they cast their vote? Of course, only the worst dictatorial regimes would ever implement such a shocking device, and political scientists fortunately respect research ethics, the privacy of the act of voting, and the ultimate taboo of democratic politics: the secrecy of the vote cast by a citizen inside the polling booth Such precepts are humanly and democratically non-negotiable, but the secrecy of the Note that those findings are not only based on the main study article but also on comments provided by some of the authors on its findings, such as an insight discussion with Itamar Simonson published online by his Stanford University business school website: http://www.gsb stanford.edu/insights/four-minute-search-perfect-mate Through the Polling Booth Curtain: A Visual Experiment on Citizens’ 325 voting act ultimately concerns a very specific aspect of the voting process: electoral choice Many other aspects of a vote are not secret at all In fact, in many countries, whether someone is registered or not, or even has voted or not is information that is publically available, other voters and election observers can see when a voter goes to the polling station, with whom, and indeed how long they spend in the booth, and many countries not use full curtains in polling booths, meaning that while nobody can see which candidate a voter chooses, their body remains in full view of other voters and polling station personnel even as they cast their vote Follows an important question for experimental political scientists: if we could see how people behave inside the polling booth seeing neither who they are nor who they vote for, could we not still learn a lot of important new elements about the act of voting in a way which is not made possible by the currently dominant methodologies in electoral research, which overwhelmingly rely on some form or other of self-reporting? At the moment, most electoral research is based on either of two options: self-reporting (survey, interview, or focus-group-based research), and experiments The latter exist in many different shapes and forms, such as rational choice experiments, increasingly popular physiological experiments, etc but not, thus far, experiments based on simple visual observation This is the very analytical and methodological premise that has led to the research presented in this chapter Teaming up with some professional film makers, we designed a visual experiment whereby we could film with great precision the shadow of voters while they were in the polling booth observing their hands and their facial and body expressions without ever betraying their facial features and identity The experiment had two main objects: (1) to identify the various emotions displayed by voters in the polling booth (do they look nervous, excited, they scratch their head or tap their feet, etc); (2) to measure the time voters spend thinking before casting their vote using three different voting procedures Those three procedures included an electronic voting machine similar to those used in many US counties; a ‘UK type’ ballot whereby voters tick a box next to the candidate that they want to elect, and a ‘French type’ ballot whereby they are given one individual ballot for each of the candidates they have to choose from, and pick one that they insert into the envelope without writing anything The experiment was conducted in Germany so that all three ballots would look similarly different from a typical German ballot Visual Observation in Experiments Visual observation is a very traditional way of looking at the effects of experimental stimuli in biology and physiology It is a “historic” technique possibly most famously illustrated by Pavlov (1927) and largely used in medicine (Scheflen 1964), and psychology (Bakeman 2000; Fleiss 1986, or controversially, the Stanford’s prison experiment—Zimbardo 2007) The observer may be present or hidden, and in psychology, the object of the observation tends to be either a form of 326 M Bruter and S Harrison human interaction (Bakeman and Gottman 1986) or subconscious reactions (Lakin and Chartrand 2003) that crucially, subjects would not have been in a position to describe in self-reporting-based research Typically, the logic behind using visual observation is thus that the researcher wants to (1) measure a behaviour which participants would not be able to consciously report, and (2) which the human eye is most apt to capture accurately and efficiently The third essential advantage of such experiments is that they also allow the experimental situation to develop in a way which is comparatively fluid and natural by experimental standards, without the need to use infrastructure or implements that would permanently remind the participants of the artificiality of the situation, such as those used in most physiological experiments Methodology For the purposes of the experiment, we teamed up with some professional filmmakers and a the Falling Walls foundation (Germany) and organised a mock election for the next ‘President of the European Union’ in November 2012 We set up the equivalent of a genuine polling station, complete with station manager, a table, ballot box, and a polling booth entirely surrounded by a curtain as would be the case in a typical election We positioned an invisible cameraman next to the polling booth in such a way that it could directly film voters from the side while they were in the polling booth We used light effects to ensure that we could get a very crisp image of the voters shadow without their identity being recognisable, and also so that the camera was not visible to participants at any time either when they entered the polling station or when they were in the polling booth However, the participants were fully aware that they may be filmed at any point while they were in the polling station (see ethics section below) Because the experiment was unprecedented, we first ran pilot tests with actors in order to discuss with the cameramen the best way to shoot the scene We also used the pilot with the two emotional display coders to evaluate how to visually recognise a number of emotional signs displayed by a human being when one cannot look at his/her eyes or facial expression but only at his/her shadow Voters had a choice between six realistic candidates that correspond to the six main political forces in the European Parliament The experiment lasted for half-a-day It was performed during the 2012 “Falling Walls” conference in Berlin (the conference commemorates the fall of the Berlin Wall on November each year but inviting about 20 scholars worldwide from all fields of knowledge—from architecture to medicine and computer technology Participants come from every walk of life and are not specifically scientists let alone academics.) During the coffee breaks, Falling Walls hosts/hostesses randomly approached conference participants asking them if they would consider participating in an experiment Most accepted in principle and were given a consent Through the Polling Booth Curtain: A Visual Experiment on Citizens’ 327 form They were then sent to the polling station Participants were randomly assigned to various voting periods, with different voting procedures in place depending on the time when they were recruited The conference was all day and each voting procedure came multiple times (in a rota system) to avoid specific time effects There were three voting procedures used.2 The first was a DRE (Direct Electronic Recording) voting machine similar to those used in many US counties for Presidential elections, whereby voters are required to choose from a list of candidates who all appear on a single screen They click on their chosen candidate and then confirm their vote The second procedure was a paper ballot copied from UK General elections It is a simple A4 piece of paper with boxes corresponding to the six candidates Voters simply tick the box corresponding to their choice Finally, the third version imitated French paper ballots in Presidential elections Each voter would be given six small sized ballots (A6 size) each with the printed name of a single candidate The voters would choose the ballot corresponding to their preferred candidate and insert it in the envelope without writing anything In all three cases, the ballots (whether paper or electronic) were awaiting the voter inside the polling booth3 and simple instructions (all the same length and structure) were provided before the voter entered the booth All instructions were pre-tested before the experiment Note that the completeness of the instruction and their pre-testing were used to ensure that differences in the time used to cast one’s vote were meaningful rather than based on technical artefacts The ballots were not equivalent to the usual German ballots (although we note their relatively greater similarity to the UK ballot), the voting procedure and ballot were fully explained in the instructions, and the list of candidates purposefully presented to the candidates in advance so that differences in voting time cannot be ascribed to simple technicalities about how easy it would be to simply read candidates’ names on the three ballot types Moreover, in terms of the measurement itself, we calculated the time between the discovery of the ballots (readily presented in the polling booth) and the moment the voter started casting their vote so as to avoid any artificial difference based on clicking/writing/folding times We are thus confident that differences in voting times are indeed meaningful and correspond to some differences in thinking and deliberative processes although in a follow up experiment we plan to fully back this up with comprehensive post-test questionnaires The data collected is the full video footage as well as the votes themselves One of the dependent variables of interest in the experiment was the emotions displayed by voters which we analysed using insights from the social psychology and kinesics For information, the ballots traditionally used in German elections are different from all three ballots used here as they have multiple columns corresponding to multiple votes (usually two) They are paper ballots and everything else being equal would be relatively closest to the UK type ballot as they are paper ballots with all candidates and parties listed and use ticking In the case of the French ballots, all six ballots were presented neatly on the polling booth table on two columns and three rows 328 M Bruter and S Harrison (see for example Birdwhistell 1970; Fast 1988), although this is not discussed in this particular chapter The second dependent variable of interest is the time that voters took to think about their vote before casting it This is defined as the time that elapses between the moment the voter discovered the ballots inside the polling booth, and the moment they cast their actual vote, both of which were precisely timed thanks to the video footage that we received Finally, a third dependent variable (not discussed here) is their actual vote choice depending on the electoral procedure used In total, 145 people participated in the experiment, all German adults 71 were allocated to the voting machine group and 37 each to the French and British types paper ballots.4 Theoretical Expectations Analytically, our chapter pertains to the impact of ballot type on thinking time in elections As part of our broader model on electoral ergonomics (Bruter and Harrison 2014a), we theorise that every small aspect in the organisation of the election (type of ballot, organisation of the polling station, design of the booth, solemnity of the procedure, home or station voting, etc), most of which have been typically deemed to be “neutral” by national electoral authorities, will actually trigger different psychological reactions on the part of voters and thus, will effectively have a potential effect on how citizens vote, how they perceive the vote (satisfaction, efficacy, legitimacy, etc) and ultimately on their long-term likeliness to participate In this particular chapter, we are concerned with the impact of ballot type on citizens’ thinking time in the voting process Our theoretical model proceeds from a double hypothesis based on our theory of electoral ergonomics and on the existing literature The first, which is based on the existing literature on electronic vs paper answering processes (e.g Shannon and Bradshaw 2002; Kiesler and Sproull 1986; Boyer et al 2002) is that paper ballots will elicit more careful deliberation than electronic ballots, which will thus lead to slower decision times overall for the two paper-based ballots The second hypothesis is grounded in priming theory and particularly on selective attention failure models (James 1890; Neisser and Becklen 1975; Miller 1987) which imply that material design will not only lead to positive object priming but to likely corollary attention failure on objects that are not primed As expressed by Eriksen and St James (1986), priming can work as a “zoom lense”, and thus while a primed object “grows” in the perception of a subject, the zoom focuses on it and The higher number of voting machine voters is solely due to a higher number of periods when it was used in our random ballots cycle Through the Polling Booth Curtain: A Visual Experiment on Citizens’ 329 therefore exclude peripheral aspects of the picture In our case, we argue that priming effects—and thus selective attention failure—will evaporate with the use of multiple ballots which make it a physically impractical possibility and effectively forces subjects to consider multiple pages (ballots) or alternatives By contrast, the synthetic presentation of electoral alternatives on a single ballot enable priming-related attention failure (or omission) and will thus speed up the voting process Overall, our model thus poses that (1) people voting using paper ballots will require longer thinking time than those using electronic ballots and (2) among paper ballot users, those needing to choose between multiple separate ballots (“French” system) will need longer thinking and deliberation time than those using a synthetic (“UK” system) ballot Ethical Safeguards We endeavoured to ensure strong ethics protection throughout the experiment Firstly, no participant could ever be personally identified nor could we ever see who they individually voted for All signed a declaration of informed consent confirming that they accepted to be filmed at any time whilst in the polling station, explaining that this was a mock election, and that the results would be used for research purposes only After their participation in the experiment, each respondent was given a full debriefing document explaining the experiment they had just participated in, the filming process, and giving them our contact details should they have any query as well as an option to choose to be deleted from the experimental footage and analysis (nobody requested that) After the debriefing, the experiment organisers asked each participant for his/her age, noted their gender, and asked whether they had noticed or suspected the presence of the camera at any point while they were in the polling booth (only two people said that they did) Findings The key substantive model discussed in this chapter is the impact of the voting procedure on the time voters spend thinking about their electoral choice before casting their vote The experimental results uncovered some very significant differences between the three procedures used in the experiment The results are presented in Fig They show that on average, voters spend approximately 1.5 times longer thinking about their electoral choice when using a ‘UK style’ paper ballot as compared to a simple electronic voting machine, and approximately times longer when they use French style individual ballots This means that on average, voters using voting machines spent 20 s thinking about their 330 M Bruter and S Harrison Fig Decision time and type of ballot vote before casting it, those using a UK-type ballot 30 s, and those using a Frenchtype ballot a full minute Those differences are statistically significant While we not discuss in detail this particular section of the study, here, the second dependent variable of interest was the range of emotions displayed by the participants throughout their time in the polling booth This data was blindly double-coded in their entirety We showed, not only that voters display significant emotional reactions while in the polling booth, but also that those reactions are significantly affected by the type of ballot that they are asked to use (Bruter and Harrison 2014b) For example, participants using one of the two paper ballots were more likely to express visual signs of happiness and visual signs of solemnity than those who used electronic ballots By contrast, participants asked to use the DRE electronic ballot showed greater signs of being worried and nervous than those using paper ballots Finally, there was much less variation in emotional valence amongst those who used electronic ballots compared to those who used paper ballots (Bruter and Harrison 2014b) Finally, the difference in the votes cast by participants using the three different types of ballots were not (although almost) statistically significant There is a realistic possibility that the differences that seem to be suggested within our sample may become statistically significant if tested on a larger group Methodological Discussion One of the key issues of electoral behaviour studies is that generally, they need to rely on respondents self-reporting their decisions as well as the procedures and motivations that led to them Even a majority of experiments in the field are survey- Through the Polling Booth Curtain: A Visual Experiment on Citizens’ 331 based experiments, which again imply a self-reporting component The exception to this rule are neurological experiments which, while also very useful to the discipline, and allowing precise clinical measures of citizens reactions in given conditions, necessitate exposing participants to highly unnatural and sometimes uncomfortable conditions, which may, in themselves, affect their reactions and behaviour In this context, the method that we have developed can serve as a useful alternative for specific research questions whereby the dependent variable is a behavioural element that can actually be observed Because of the taboo surrounding the polling booth, ethical precautions are particularly critical, but we believe that the system that we developed and which is based on filming shadows provides extremely strong guarantees of both anonymity and protection of the privacy of the electoral choice Direct observation allows us to accurately and directly capture a number of ‘facts’ about the act of voting, such as how long people think before casting their vote By contrast, questionnaires and self-reporting can only give us informations about voters’ self-perceptions of their own behaviour on such elements with a necessary element of approximation and in-built error Moreover, while physiological experiments would be a credible alternative, they imply imposing a (less and less but still) invasive capturing device to the participants which reminds them of the artificiality of the exercise By contrast, while the experiment that we conducted was obviously also artificial in the sense that participants knew that they were voting in a simulated election, visual capture itself did not interfere with their voting experience and there was no obvious physical anomaly to the voters’ usual experience Arguably, the polling booth moment therefore benefitted from a comparatively more natural setting Visual experiments have their own obvious limits They can be relatively complex to put together to ensure an optimal semblance of realism and unobtrusiveness to the participants, they impose upon us to be extremely rigorous in terms of ethical and technical safeguards to ensure the full respect of subjects’ privacy and consent while interfering as little as possible with their natural behaviour, and there is an element of interpretation of the visual evidence which typically requires strong coding guidelines and the use of multiple coders to ensure reliability of the measured data We are also not claiming that visual experiments would replace other methods when it comes to understanding voting behaviour—including other experimental protocols and self-reporting Indeed, even though crucially our research design enabled us to have a direct visual observation of what happens inside a polling booth without fitting cameras inside it, being able to recognise the participant’s identity, let alone their actual voting choice which is never visible to the researcher, it is of course hard to imagine systematically introducing such processes in real elections, if only because it is easy to know how they could be wrongly manipulated by ill-intended groups What we are saying, instead, is that this is a useful additional method that can be used to put together a different and complementary way of observing voting processes, and that it might enable us to answer particularly 332 M Bruter and S Harrison important questions in the fields of electoral psychology and electoral ergonomics, as well as other (non electoral) processes of political psychology in even easier and most obvious ways This is really the limited but important ambition of our methodological contribution Thus, in this chapter, we argue that this type of experiment, which has been long used by biologists, psychologists, and anthropologists in particular, can occasionally shed a different light on a number of aspects of political behaviour, and in particular on the highly secret world of voters’ behaviour in the polling booth References Bakeman, R (2000) Behavioral observation and coding In H Reis & C Judd (Eds.), Handbook of research methods in social and personality psychology Cambridge: Cambridge University Press Bakeman, R., & Gottman, J (1986) Observing interaction: An introduction to sequential analysis Cambridge: Cambridge University Press Birdwhistell, R (1970) Kinesics and context Philadelphia: University of Pennsylvania Press Boyer, K., Olson, J., Calantone, R., & Jackson, E (2002) Print versus electronic surveys: A comparison of two data collection methodologies Journal of Operations Management, 20(4), 357–373 Bruter, M., & Harrison, S (2014a, February) Memory, emotion, and ergonomics: The electoral psychology of voters in perspective Paper presented at the LSE for the ECREP Workshop Series Bruter, M., & Harrison, S (2014b, March) The emotional act of voting: Results from a visual electoral experiment Paper presented at the LSE for the ECREP Workshop Series Eriksen, C., & St James, J (1986) Visual attention within and around the field of focal attention: A zoom lense model Perceptions and Psychophysics, 48, 225–240 Fast, J (1988) Body language New York: Simon and Schuster Fisman, R., Iyengar, S., Kamenica, E., & Simonson, I (2006) Gender differences in mate selection: Evidence from a speed dating experiment Quarterly Journal of Economics, 121 (2), 673–697 Fleiss, J (1986) The design and analysis of clinical experiments New York: Wiley Hartmann, D (Ed.) (1982) Using observers to study behavior San Francisco: Jossey-Bass James, W (1890) Principles of psychology New York: Holt Kiesler, S., & Sproull, L (1986) Response effects in the electronic survey Public Opinion Quarterly, 50(3), 402–413 Lakin, J., & Chartrand, T (2003) Using nonconscious behavioral mimicry to create affiliation and rapport Psychological Science, 14(4), 334–339 Long, J (1996) Video coding system reference guide Caroga Lake, NY: Long Miller, J (1987) Priming is not necessarily for selective-attention failures: Semantic effects of unattended, unprimed letters Perception and Psychophysics, 41(5), 419–434 Neisser, U., & Becklen, R (1975) Selective looking: Attending to visually specified events Cognitive Psychology, 7, 480–494 Pavlov, I (1927) Conditioned reflexes: An investigation of the physiological activity of the cerebral cortex (G Anrep, Trans & Ed.) London: Oxford University Press Scheflen, A (1964) The significance of posture in communication systems Psychiatry, 27, 316–331 Shannon, D., & Bradshaw, C (2002) A comparison of response rate, response time, and costs of mail and electronic surveys Journal of Experimental Education, 70(2), 179–192 Through the Polling Booth Curtain: A Visual Experiment on Citizens’ 333 Tapp, J., & Walden, T (1993) PROCODER: A professional tape control coding and analysis system for behavioral research using videotapes Behavior Research Methods, Instruments, and Computers, 25, 53–56 Weick, K (1985) Systematic observational methods In G Lindzey & E Aronson (Eds.), Handbook of social psychology (3rd ed., Vol 1, pp 567–634) New York: Random House Zimbardo, P (2007) The Lucifer effect: Understanding how good people turn evil New York: Random House ... includes laboratory experiments, survey experiments and field experiments but as experiments they all share a common denominator In order to study the impact of the environment on voting decisions,... Jean-Franc¸ois Laslier, and Karine Van der Straeten Voting experiments are about voting, and the dependent variable is therefore obvious: voting behavior What are the independent variables? The... (most of) the voting experiments presented in this book is to understand how different facets of the voters’ political and social environment influence their voting decisions These experiments