How do ci tizensform their political atti tudes?How do polit icalactors in fluence people’s political views in th ei r everyday lives? Do people’ s fe elings towards politic alpar ti esand polit ic iansinfluence publi c opinion? The book Understanding Policy Attitudes: Ef fects of Source Cues on PoliticalReasoningpre sents a seriesofla boratory experiments focused on the formation processes of ci tizens’ atti tudes to politi cal issues in the context of their emotional attachments to politi cal actors. It investig ateshow basic cues, in for mof af fective reactions of experi - menta l subject s to wards political parties and party leaders, influence attit udes on issues ar ticulated by these par ties andleaders . The re searchputs emphasis on negativity and negative fe el in gsof ci tizens towar dspoliti calactorsandshows that this typeofnegative attachment influences the way people think about political issues.
The re sults of thepr esente d st udiessuggest th at citizens’ id eolo gical
setup is the decisive force behind policy attitudes. Sourc e cues, th erefor e,exercise their power withi n thescope of less politi cally and ideo - logically profiled issues, more technical and obscure. In a seri es of experiments th e bookal sosets af fective cues in context of two theore ti calapproaches towards political cogniti on– politi calheuri sticand mot ivated reasoni ng– and examines what type of reasoning is triggered by af fective source cues. Outcomes of th e fir stresearch on th e to pi c in th e context oftheCzechpolitical environmentarepresented.Since the experimental method is a newcomer in the fiel dof Czech politi cal scie nce, th e volume’s ambition is also to introduce experiments as a re le vantand useful tool for extending knowledge of substantial politi cal pr ocesses and phenomena. Lenka Hrbková is Assistant Professor at th e Department of Political Science of Masaryk University . Shespecialises in th e studyof political psychology and public opinion, with a substanti vefocus on poli ti cal attitudes and emotions in politics.
OF AFFECTIVE SOURCE CUES
ON POLITICAL REASONING
OF AFFECTIVE SOURCE CUES
ON POLITICAL REASONING
MASARYK UNIVERSIT Y
The research presented in this volume was funded by the Czech Science
Foundation within the project Experimental research of voting behavior and
decision making in highly personalized elections (P408/GA13-20548S).
Scientific Editorial Board of Masaryk University
Prof. MUDr. Martin Bareš, Ph.D.
Mgr. Iva Zlatušková
Ing. Radmila Droběnová, Ph.D.
Mgr. Michaela Hanousková
Assoc. Prof. Mgr. Jana Horáková, Ph.D.
Assoc. Prof. PhDr. Mgr. Tomáš Janík, Ph.D.
Assoc. Prof. JUDr. Josef Kotásek, Ph.D.
Mgr. et Mgr. Oldřich Krpec, Ph.D.
Prof. PhDr. Petr Macek, CSc.
Assoc. Prof. Ing. Petr Pirožek, Ph.D.
Assoc. Prof. RNDr. Lubomír Popelínský, Ph.D.
Mgr. David Povolný
Mgr. Kateřina Sedláčková, Ph.D.
Prof. RNDr. David Trunec, CSc.
Prof. MUDr. Anna Vašků, CSc.
Assoc. Prof. Mgr. Martin Zvonař, Ph.D.
PhDr. Alena Mizerová
Reviewer: doc. PhDr. Lukáš Linek, Ph.D.
© 2016 Masarykova univerzita
ISBN 978-80-210-8452-0 (online : pdf )
This book is the result of my dissertation research which would not have been
possible without my supervisor Roman Chytilek, who first taught me about
experimental research in political science and let me pursue rather unusual
research topics in the context of Czech political science. I would like to give
special thanks to my colleagues Otto Eibl, Jozef Zagrapan, and Petr Voda for their
useful comments and advice on my ideas, trials, and errors throughout my entire
relationship with this project. The experiments certainly would not have
happened without the technical and practical support by Michal Tóth and Miro
Special thanks goes to David Redlawsk who was so kind to read my
dissertation as the opponent and gave me invaluable feedback and support to
continue in my search for answers to questions related to political psychology. I
sincerely appreciate Lukáš Linek and his review of my work and
recommendations for the improvement of this text. Last but not least, I would
like to thank my family and friends for their support and optimism which made
this book possible.
INTRODUCTION 1. EXPERIMENTS IN POLITICAL SCIENCE 1.1 Experimental Political Science 1.2 Internal Validity 1.3 External Validity 1.4 The Relationship between Internal and External Validity 2. INFORMATION PROCESSING AND HEURISTICS 2.1 Limits of Human Cognition 2.2 Source Cues and Political Reasoning 3. EMOTION IN POLITICS AND AFFECTIVE CUES 3.1 Affect and Cognition 3.2 Affective Politics and the Theory of Affective Intelligence 3.3 Automaticity of Affect and John Q. Public Model 4. AFFECTIVE SOURCE CUES IN SINGLE PARTY CUE EXPERIMENTS 4.1 Affective Cues, Negativity Bias, and Leader Effect 4.2 Source Cues and Prior Policy Attitudes 4.3 Summary
96 5. CUE-TAKING: HEURISTICAL THINKING OR MOTIVATED REASONING? 5.1 Motivated Political Reasoning 5.2 Affective Cues of Policy Source and Policy Target 5.3 Summary 6. CONCLUSION 7. APPENDICES 7.1 Appendix A 7.2 Appendix B 7.3 Appendix C 7.4 Appendix D 8. REFERENCES LIST OF TABLES
LIST OF FIGURES
“Hence, in order to have anything like a complete theory of human rationality, we
have to understand what role emotion plays in it.”
Herbert A. Simon, 1983, Reason in Human Affairs, p. 29
“There are no definite answers to many questions in politics, and anyone who claims
to have uncovered the “laws” of political behaviour should be treated with a heavy
dose of skepticism.”
David P. Houghton, 2015, Political Psychology, p. 17
The book you are holding in your hands explores the way people use information
in a political context from a cognitive point of view. I consider the topic to be of
an extremely high importance – after all, information and information processing
can be understood as the most fundamental element of human behavior, political
behavior included. As Steven Pinker (2002) points out, the cognitive revolution
revealed that the world of peoples’ ideas and the world of physical matter are
interconnected through information, information computation, and feedback,
and that all human behavior has reasons underpinned by information processing.
The cognitive approach has also had its impact on political science, political
psychology, and the study of political processes in general by the acceptance of
the assumption that information about the outside world is organized in citizens’
memory structures, and that these memory structures determine how people
interpret and evaluate political events and how they make decisions (McGraw
2000). To improve the knowledge of how people think and behave in politics, it
is necessary to understand how they actually handle the information surrounding
The focal point of this book is the concept of source cues and their role in
policy attitude formation. “Cues” refer to “stimuli in the persuasion context that
can affect attitudes without necessitating processing of message arguments” (Petty
and Cacioppo 1986: 18). A cue, thus, serves as a datum that people may use to
infer other information and, by extension, make decisions based on it (Bullock
2007). Even though cues may take a variety of different forms (as authors such
as Kahneman and Tversky have pointed out) in the context of political decision
making and of politics as such, the most common form of cues which citizens
use are those related to the source of an item of information. In source-cue
research, the main research interest is focused on the source of a particular
political message as the primary persuasion stimulus of interest. Source cues are
considered important in information processing because they provide an easy
tool for citizens to evaluate a message (e.g. Carmines and Kuklinski 1990;
Mondak 1993). When an individual engages in this type of cue-taking, it is
possible that in the course of the evaluation of an item of political information,
she would focus on “who” is saying it rather than “what” is actually being said
(Kuklinski and Hurley 1994).
There has been a strong tradition of source cue research in US political
science and political psychology. Jeffrey Mondak, in one of the early studies on
the topic of source cues, confirmed that once people hold opinions about political
leaders, those opinions guide their assessment of policy issues associated with
the leaders (e.g. Mondak 1993a). Mondak also suggested that the source cue
mechanism applies not only for individual citizens, but also for aggregate public
opinion (Mondak 1993b). At about the same time, Wendy Rahn showed strong
evidence of source cue effects. In her study, voters – even though they were
capable of data-driven processing of information – engaged in heuristic
processing in candidate evaluation, neglecting policy information and relying
mostly on party stereotypes when faced with information about the party
attachment of a candidate (Rahn 1993). Pamela Conover and Stanley Feldman
showed that political party cues also steer voters’ assessment of candidates’ issue
positions (Conover and Feldman 1989). Source cues matter not only in relation
to voters’ perception and evaluation of political candidates, but there is also
evidence that they affect other aspects of political judgment and behavior, such
as attribution of blame and responsibility for political failures to politicians in
context of natural disasters (Malhotra and Kuo 2008). Vote choice is naturally
one of the important aspects of political decision-making which may be affected
by cues. It has been suggested that the endorsement of candidates by well-known
interest groups may be used by some voters as a voting cue compensating for a
lack of awareness about politics (Arceneaux and Kolodny 2009). Source cues also
work as an important factor in the creation of opinions about political issues.
Available research suggests that when evaluating policy proposals, citizens tend
to rely on party cues and support policy positions sponsored by their political
party regardless of their own actual ideological positions (Cohen 2003). Less
politically aware citizens in particular rely on party cues when assessing policy
issues (Kam 2005). Source cues may also influence such core phenomena as
INTRODUCTION political values. Party cues have been proved to affect not only which basic political values citizens tend to claim, but also to promote horizontal constraint among these values, meaning a certain level of coherence of values reflecting the left-right value distribution (Goren, Federico, and Kittilson 2009).
Obviously, the study of source cues and related cognitive processes is a
relatively traditional agenda of political science and political psychology. However, an important part of the general knowledge of how source cues work – as is the case of any social phenomenon – is to take environmental contexts into consideration as well, and to apply research questions related to cue-taking across different political environments. This point can be considered to be an important but often neglected feature of political psychology research, which predominantly focuses on the micro level of political processes and tends to treat political judgment as detached from any social and political context, such as differences in political systems, ideologies, party systems, political culture etc. Such practices may even invoke concerns about political psychology being “insufficiently political” (Rahn, Sullivan, and Rudolph 2002). This concern is relevant especially (but not exclusively) for experimental research, which is inherently reductionist; it often depends on limited samples and examines human attitudes and behavior in very specific situations and therefore the validity of results is also limited (the limits of laboratory experimental research and the problem of validity are discussed in more detail in Chapter One of this volume). However, the influence of context on the individual has been identified as a significant factor altering human behavior (Mischel 1997; Rousseau and Fried 2001; Sarason, Smith, and Diener 1975).
Political scientist and statistician
Andrew Gelman emphasizes, “Once we realize that effects are contextually bound, the next step is to study how they vary” (2014: 633). Asking the same research questions across different political and cultural environments should therefore strengthen our scientific knowledge of the given phenomenon – either by establishing external validity to the results replicated across different contexts or by extension or updating of the theoretical assumptions. The failure to replicate results across various datasets in different contexts may thus contribute to the scientific inquiry as well. Many scientific discoveries can be traced to failed attempts at replication (Feldman-Barett 2015; Van Bavel et al. 2016). The use of scientific replication across different political and cultural environments may be 1
Contextual factors also play a role in the often-discussed “replicability crisis” in experimental
research in psychology. Many variables in psychology as well as other social sciences cannot be
fully understood apart from cultural and historical contexts which may result in failed replication
(Van Bavel et al. 2016).
one way to alleviate some of the concerns about the decreased emphasis put on the actual “political” factors in individual studies – which is a common feature of studies in political psychology.
This volume applies the study of source cue effects to a different, more complex environment compared to the US context, where only two major political parties and two dominant ideologies compete and present a rather simple structure of choice with mutually exclusive alternatives. The experimental research presented in this book was intended to expand the current knowledge of source cues and to add to a still rather limited body of research in a non-US context (Brader, Tucker, and Duell 2012; Capelos 2010; Merolla, Stephenson, and Zachmeister 2007; 2008; Petersen et al. 2013; Samuels and Zucco 2014). In Europe, the topic of source cues has been studied surprisingly scarcely with the exception of Ted Brader, Joshua Tucker, and Dominik Duell’s comparative experimental study of party cue effects on citizens’ issue positions in Hungary, Poland, and the UK (Brader, Tucker, and Duell 2012), and Tereza Capelos’ experimental study about foreign leader cues in assessment of foreign policy proposals, which was conducted in the Netherlands (Capelos 2010). In addition, there has also been a Danish study on cognitive processes related to cue-taking (Petersen et al. 2013). In my inquiry, I extended the source cue research in a series of laboratory experiments conducted in the Czech Republic, a country offering a very different structure of political environment compared to the USA. One of the most profound differences is the complexity of the political space, presumably increasing opportunities for citizens to engage in cue-taking. The topic of how citizens process political information in general – and how political actors might influence public opinion or political behavior in particular – has been a somewhat sidelined topic in Czech political science so far. Therefore, this volume provides an innovative approach to the study of Czech politics and public opinion.
This research searches for answers to several basic questions derived from the available literature about political source cues. First of all, do citizens use political parties as cues in their judgment about policy issues? Traditionally, scholars have studied the effect of source cues in the context of party identification or sympathy to a political party. My experimental designs enable me to extend the study of source cues with the concept of negativity – effects of which have been well documented in various aspects of human reasoning (Rozin and Rozyman 2001; Skowronski and Carlston 1989; Vaish, Grossman, and Woodward 2008). Even though the role of negative information has been studied in psychology as well as in several political contexts – such as negative political campaigns (e.g. Mattes and Redlawsk 2014; Mark 2009) or differences in political ideology (Hibbing, Smith, and Alford 2014; Malka and Soto 2014) – not much
attention has been paid to the issue of negative source cues so far. The question undergirding my research interest is simple and asks whether only those parties that citizens like influence how citizens perceive those parties’ policy positions, or whether negative feelings of parties can also drive policy attitudes. My findings confirm that a negative perception of political actors actually may influence how citizens form attitudes about various political matters and that negativity towards political actors is an important part of political cognitive processes.
The second research question explores whether the perception of political leaders works as a more salient or useful cue compared to political parties. This aspect of cue-taking may be important, since the visibility and media coverage of political leaders is increasing (Aelst, Schafer, and Stanyer 2011) and the concept of personalization of party politics – in which party leaders gain a more prominent role vis-à-vis political parties – has increasingly been identified as a trend in various countries (e.g. McAllister 2007; Rahat and Sheafer 2007; Garzia 2013).
The conceptualization of source cues in this book differs from the traditionally used partisan cues. Instead, I base source cues on affect-driven reactions of an individual towards specific political actors. Cues are, therefore, understood to be the basic likes and dislikes felt towards political actors. This approach enables me to explore whether the affective effects of leaders differ from affective effects of political parties. This approach is the result of a previous pilot experimental study and attempts to measure citizens’ discrete emotional reactions (in the sense of anxiety and enthusiasm, as predicted by the Affective Intelligence Theory further discussed in Chapter Three) towards political parties. Political parties proved to be too weak as stimuli for subjects to manifest meaningful emotional reactions in the aforementioned pilot, which prompted the idea that political personalities might possibly be more affectively valenced concepts. However, my current findings, reported further in this book, suggest that political leaders do not fit any differently in subjects’ cognitive framework than political parties with respect to the effects on reasoning about policies. In other words, sponsorship of a policy proposal by a party leader doesn’t evoke any different evaluation of the proposal than strictly partisan sponsorship.
The third research question addresses the effects of the source of an item of information as a factor influencing public opinion in light of one’s own ideological principles. Do citizens follow the party lead, or do partisan cues interact with what citizens already think about politics? Answering this question offers a more complex perspective on the source cue quandary, because source cues do not exist in an information or political vacuum. Traditionally, the effects of source cues on the policy positions of citizens have been studied in relation to
partisan ideological consistency. The main point of interest is whether citizens toe the party line even when assessing a policy position which is at odds with stereotypical party ideology (Cohen 2003; Greitemeyer 2009; Petersen et al. 2013). However, my approach does not focus on how source cues work when the policy position does not correspond to the expected position of the sponsoring political party, but on the effects of source cues in relation to citizens’ own policy attitudes. The intention behind this innovation in the experimental design is to extend the knowledge of the conditions under which source cues actually do influence policy attitudes.
Even though source cue research has traditionally focused on the influence that elites in general and political actors in particular may have on public attitudes towards various issues, more recently questions have been raised about the type of cognitive processes which actually undergird source-cue information processing (Bolsen, Druckman, and Cook 2013; Petersen et al. 2013; Leeper and Slothuus 2014). Therefore, the fourth question of this book seeks to answer is: Are source cues used as cognitive shortcuts, or do they trigger a different, more elaborate type of reasoning? Nevertheless, the goal of this research is not to evaluate the quality of the judgment based on cue-taking, but to test what level of cognitive effort source cues actually demand and whether the traditional understanding of source cues as a type of heuristic or cognitive shortcut holds. In my experiments, I follow the research on motivated reasoning in the context of partisan cues (Petersen et al. 2013), but I expand the approach with the concept of (in)congruence of one’s feelings towards a political party and a political message by the same party. My experimental designs enabled me to test people’s modes of reasoning when a message by a political party was not congruent with their feelings towards a party. In other words, the book focuses on what happens when one’s favorite party proposes a policy one has to disagree with (or when one’s disliked party proposes a policy one agrees with). This approach brings a new perspective to source cue research by focusing on what happens when citizens encounter an item of information which causes a conflict of attitudes, which can – and does – happen quite easily in real life.
This research examines the process of public opinion formation in light of two essential types of factors. The first factor consists of the political actors that form the political environment in which citizens operate and where they serve as the main source of policies. The other factor, which is related to the political actors under scrutiny, are the feelings and affective reactions these actors provoke in citizens. My research approach draws on the assumption that interactions between actors and feelings towards the actors are the basis for attitudes on various political matters. Bringing affective reaction towards political parties and
their leaders into consideration this approach breaks new ground in the study of politics in the context of Czech political science. Even though affects and emotions have been ignored by scholars of political processes in general for a long time, previous findings in psychology, cognitive science, and neuroscience have confirmed that human emotions exert an important impact on reasoning and decision making (Zajonc 1980; Damasio 1994); this conclusion naturally applies to all aspects of human life, including politics. The role of emotions has already been acknowledged to influence political processes and it has become a prominent topic of political psychology research, captured in several influential volumes (e.g. Marcus, Neumann, and MacKuen 2000; Lau and Redlawsk 2006; Neuman et al. 2007; Demertzis 2013; Lodge and Taber 2013). This volume seeks to reflect this development in the field of political psychology and to include both affective and cognitive processes into the study of politics in the Czech Republic. In order to do so I conceptualize a source cue as an affect-driven phenomenon, in which basic affective reactions towards political actors are transferred to other related objects, such as policy proposals.
I understand affect to be “an episode of massive synchronous recruitment of
mental and somatic resources to adapt to and cope with a stimulus event that is subjectively appraised as being highly pertinent to needs, goals, and values of the individual” (Scherer 2005: 314). Following the work documenting the primacy of affect in political psychology (e.g. Cassino and Lodge 2007; Lodge and Taber 2013), this research project has been based on the assumption that feelings and affective reactions
are distinguished by their valence, and that the basic affective
dimension is positivity-negativity. The main concept used in the following experiments are the basic positive feelings (or “likes”) and negative feelings (or “dislikes”) subjects feel towards political actors; this way, the experiments enabled me to test theoretical source cue assumptions for both liked and disliked political actors.
Complementary to affect and affective processes are cognition and cognitive
processes. Cognition has typically been understood as “thought-knowledge” and referred to as conscious and intentional processes (Spezio and Adolphs 2007). 2
In the following text, the terms “feelings” and “affect” are used interchangeably, though
neuroscience understands feelings explicitly as one’s conscious emotional processing (e.g.
Damasio 1994). 3
However, other approaches may define emotions differently. For example, discrete models of
emotions assume that discrete emotions are particular responses towards different circumstances.
For example, Ekman differentiates between six universal emotions: anger, fear, disgust, sadness,
happiness, and surprise (Ekman 1992), and more recently, shame has been added to the list
(Ekman and Rosenberg 2005).
However, in line with recent developments in neuroscience, the general term of cognition in this volume includes also nonconscious and automatic processes – therefore, cognitive processes are understood to be all ongoing processes of the mind, regardless of their intentionality.
This volume contributes to the knowledge of how citizens use political information within the process of attitude formation. The main contribution to the source cue research in political science and political psychology is twofold. First, it includes the concept of negativity, which has been considered an important factor influencing a vast number of human cognitive processes, however, has not been properly set in context of neither source cue effects not motivated reasoning. The results suggest that negativity bias should be included into our inferences about heuristics as well as about motivated political reasoning. Negativity in terms of negatively valenced feelings towards various political actors are a common feature of politics, especially when the politics is polarized and the feelings that citizens experience in relation to politicians, political parties or political groups are strong.
Second, the experimental approach of the presented research uses a dynamic adjustment of the treatment individually to each subject so that the affective cues can be set in the context of the individual policy attitude. The available research mostly deals with situations in which the policy proposal contradicts the expected position of the given sponsor of the policy. However, the matter of interest in this research is how citizens actually process a piece of information about policy which is dissonant with their own opinions. This approach is especially useful in such political contexts where the ideological cleavage between parties is not so clear-cut and predictable, which complicates the experimental designs. Also, citizens’ opinions are not absolutely identical to party ideologies, therefore the program of one’s favorite party and one’s attitudes may differ. Interestingly, in real life, also the situation when one’s disliked party holds a policy position which corresponding to citizens’ political attitudes might occur. Even though cases of such an informational incongruence might be a common feature of politics, particularly with the increased influx of political information on the web and social media, its effects on citizens’ attitudes and processes of attitude formation are not clear. This research, even though it is set in a laboratory, is relevant for situations which we can easily encounter in our everyday lives.
For the purposes of this type of research, laboratory experiment functions as the most suitable method, because it enables researchers to have a high level of manipulation and control of the information that the experimental subjects actually perceive and react to. Experiments have become a prominent method in the study of various political processes and phenomena (Druckman et al.