Tải bản đầy đủ - 0trang
1 Issue Information, Cross-Pressure, and the Campaign Context
ossing and T. Weber
Fig. 2 Interaction of policy match and reinforcement, pre-campaign (from Model 4)
larger reinforcement and thus the less pronounced cross-pressure becomes, the
lower the effect of policy match, until it vanishes entirely for perfect reinforcement.
A statistically significant effect is found for levels of agreement up to 0.51 on the
0–1 scale. These are the participants experiencing cross-pressure on their voting
decision stemming from the contradiction between their patterns of party identifications and the assigned constellation of party conflict. They resolve cross-pressure
by engaging in information processing, indicated by the increasing effect of policy
match on vote choice, in contrast to the other participants who find their preferences
echoed by the assigned constellation. Unaffected by intensifying partisan conflict
and prior to making up their mind about their voting intentions in the upcoming
election, individuals are willing and able to process the policy information they
receive to productively engage the cross-pressuring forces coming from the contradiction between constellations of party conflict and multiple partisan attachments. Information that merely reproduces patterns of identification, however, is
ignored at this stage.
By contrast, testing the competing expectations of information discarding versus
information processing during an ongoing campaign, model 5 reveals a significant
positive interaction of policy match and reinforcement that is consistent with
hypothesis 2. While model 4 provides support for the occurrence of information
processing in response to cross-pressure, as predicted by hypothesis 3, model
5 demonstrates that during the campaign, voters will discard inconvenient information, as predicted by hypothesis 2. Illustrated in Fig. 3, the smaller the reinforcement of individual patterns of party identifications through the assigned
constellations of party conflict, the lower the effect of policy match, until it vanishes
entirely for the complete absence of reinforcement. Note that the smallest level of
The Company Makes the Feast. Party Constellations, Campaign Context and. . .
Fig. 3 Interaction of policy match and reinforcement, during campaign (from Model 5)
reinforcement stands for the most pronounced cross-pressure through the largest
possible contradiction between patterns of party identifications and constellations
of party conflict, while perfect reinforcement identifies the complete absence of this
kind of cross-pressure. The decreasing effect of policy match on vote choice that
occurs as the result of increasing contradictions demonstrates that individuals
resolve cross-pressure by discarding issue information during a campaign. Model
5 thus shows that in the context of a campaign, when voters are subjected to
intensifying partisan pressure and when it is more likely that they already made a
decision about their intended vote choice, they are clearly more prone to discard
issue information that contradicts their existing patterns of party identifications. In
summary, then, finding evidence for information discarding during and information
processing before the campaign supports our hypothesis 4.
Alternative Pathways: Reinforcement or Affirmation?
To further support our finding that the incidence of either information discarding or
information processing depends on the campaign context, we have conducted two
additional tests for a possible alternative explanation. These tests are shown in
models 6 and 7 in Table 1. While we found opposite directions in the interaction
between policy match and reinforcement for each of the two stages of the campaign
context in our previous analyses, models 6 and 7 show that there is no interaction
between policy match and what we call the “affirmation of preference order”. The
difference between “reinforcement” and “affirmation” is that the latter measures
ossing and T. Weber
whether the assigned policy positions should please the participant (i.e. whether the
parties with positions close to the participant’s are those with high thermometer
scores) whereas the former measures whether the assigned positions pit parties
against each other that are also far apart in their thermometer scores.
As explained above, reinforcement is operationalized as the R-squared of a
regression of stacked party identification on the binary policy information, divided
by the maximum possible R-squared given the subject’s preference order. Affirmation then is the correlation of stacked party identification with the binary policy
information, divided by the maximum possible absolute value. Affirmation is
therefore negative when high-ranking parties are assigned disliked policy positions.
This makes affirmation more restrictive than reinforcement because it requires a
positive correlation of policy positions and party identifications, whereas reinforcement merely requires any correlation (positive or negative).
According to the evidence presented in models 6 and 7, the alternative hypothesis that policy information only matters if the preferred parties happen to support a
subject’s prior views should thus be rejected.14 What is happening is not cherrypicking; the responses of voters to issue information are clearly linked to varying
constellations of party conflict. Prior to the campaign, information cutting across a
subject’s preference order has the most impact, showing that voters engage crosspressuring considerations through information processing. During the campaign,
the same kind of information has the least impact, showing that individuals discard
information now to avoid inconvenient cross-pressure.
8 Conclusion and Discussion
This chapter finds issue information that allows voters to infer the match between
party positions and their own preferences to have a meaningful impact on vote
choice. However, by investigating patterns of party identifications and constellations of party conflict emerging from the distribution of issue positions between
parties, we show that parties need to mind the company they keep to effectively use
information about their issue positions for generating electoral support. Keeping
company that voters find unusual can diminish the impact of issue information on
vote choice. Voters find the company parties keep unusual, when they feel widely
divergent degrees of sympathy for two parties that express the same issue preference, or alternatively, when two parties they equally like or dislike express different
The analyses shown in Table 1 evaluate the two alternative explanations separately (models 3–5
for “reinforcement” and models 6 and 7 for “affirmation”). Testing both variables at the same time
in one single model produces the same results. The coefficients of the “affirmation” variables also
remain weak and non-significant when tested on the entire sample (before and after the campaign).
We included a table showing these complementary empirical tests in the online appendix.
The Company Makes the Feast. Party Constellations, Campaign Context and. . .
Unusual company diminishes the impact of information about parties’ issue
positions on vote choice only during an ongoing election campaign. As a result of
the unexpected connection between issue positions and parties under conditions of
intensifying partisan conflict and solidifying vote intentions, voters are less willing
and able to process information about parties’ issue positions to guide their vote
choices. Unusual company through disliked parties sharing the same issue position
constitutes a dilemma for party strategy in multi-party systems, as parties have very
few ways of convincing contenders disliked by voters to not share their own point
of view. By contrast, before an election campaign, unusual company increases the
relevance of issue information. We show that contradictions between constellations
of party conflict emerging from the distribution of issue positions on the one hand
and individual patterns of party identifications on the other hand prompt individuals
to engage in increased processing of policy information when the campaign has not
yet started. This is good news for the quality of political communication outside of
campaigns, as voters seem to be able to incorporate parties’ policies during an
electoral term into their voting preferences. At the same time, it represents a
dilemma for both a deliberative and a representative view of democracy that
election campaigns—after all the main occasion for political communication—do
not seem to be a good context for getting voters to consider information about issues
and for gathering electoral support by expressing policy positions.
To reach our conclusions about party conflict and issue voting in multi-party
systems, we elaborate on theories of electoral cross-pressure on the basis of dualprocess models of preference formation and research about campaign effects. We
conduct an experiment that randomly assigns political parties to different positions
about the controversial issue of political authority for employment policy, favoring
either Europeanization or national responsibility. Using an experimental approach
to test our theoretical expectations gives our findings a high degree of internal
validity. Beyond a general claim that experimentation produces inferences with
high internal validity, we argue specifically that our research overcomes two types
of endogeneity problems that exist in prior research on party conflict and voting
behavior. The first type of endogeneity occurs in existing observational research
because voters might rationalize their own issue preferences and perceptions in
response to learning about the position of a favored party (Granberg and Holmberg
1988). We eliminate this particular endogeneity threat by gathering information
about individuals’ issue positions pre-treatment and by randomly assigning different policy positions to different parties. The second type of endogeneity in prior
observational research concerns the other side of the mass-elite divide. Political
parties, at the elite level, react to information about the preferences of the mass
public by strategically adjusting their policy positions (De Sio and Weber 2014), for
example to pander to their core supporters or to win a new source of support. Again,
the experimental research design that we implement prevents this type of
endogeneity by practicing random assignment of issue positions to different parties.
In addition to facilitating high internal validity, our research is also strongly
concerned with guaranteeing the external validity of our findings. This objective is
motivated by a desire to close the gap between non-experimental observational
ossing and T. Weber
studies on the one hand and research using a highly stylized experimental approach
on the other hand. Our claim to external validity is based on using a sample of
students and general population participants, real political parties that constitute the
German party system, the context of a real election campaign contested over the
elections to the European Parliament in May 2014, and a real issue, which, at the
same time, has unique features that lend themselves to experimental manipulation.
Speaking of this issue—the Europeanization of employment policy—our findings are symptomatic for the state of the political representation of attitudes toward
the EU. Our participants show clear reactions to party positions that are entirely
random. To put it more bluntly, any party may stand for any policy. Arguably, this
is the case because parties in the real world have failed as of yet to formulate clear
stances about European integration and communicate them to their voters in
responsible ways. The “cartel” of shallow pro-European positions among
established parties (cf. Hix and Lord 1997; Van der Eijk and Franklin 2004;
Weber 2007) still seems to be alive and well, despite increasing journalistic and
scholarly talk of a politicization of European integration after the failed constitution
and the Euro crisis. Our experiment shows that preferences about European integration could inform voting behavior if parties were able and willing to formulate
meaningful alternative positions. This suggests that our study has implications for
many other, often more controversial issues, where the same mechanisms should
apply in at least equal measure.
Our finding that voters are less willing and able to consider issue information for
their vote choices during an election campaign complements the findings of
Hillygus and Shields (2009) about the different yet related concept of electoral
defection. They observe that campaigns in a two party system can be conducive to
defections from routine patterns of voting behavior guided by partisanship in
response to “wedge issues”, on which a less favored party holds a position that
appeals to the cross-pressured voter. Yet we find that processing of issue information becomes less prevalent during campaign time. Considering the conclusions
from both contributions in conjunction with one another shows, first of all, that
cross-pressure in multi-party systems, through “wedge issues” and other mechanisms, works differently than in two party systems. With only two parties present,
exercising cross-pressure is the only way, in the domain of substantive political
conflict, for a party to actively steer support away from its competitor. Our research
shows that in a multi-party system, the effectiveness of such strategies can be
diminished by unusual company in an unexpected constellation of party conflict.
Second, placing the two contributions in context also highlights the difference
between the effect of the campaign context at the micro level and the aggregate
impact of an election campaign. On the one hand, Hillygus and Shields (2009) show
that election campaigns can push voters in certain directions, in the aggregate, and
this occurs as the result of the sheer amount of information unloaded on them with
ever more surgical accuracy. Our research, on the other hand, shows that for each
individual voter and each separate piece of information, the campaign context in
and of itself appears to be inadequate in fostering a suitable environment for
productively processing issue information.
The Company Makes the Feast. Party Constellations, Campaign Context and. . .
Acknowledgments We are grateful for the opportunity to present our research at the Montre´al
voting experiments workshop and the MSPA conference 2014. Thanks for helpful comments go in
particular to our discussants Richard Lau and Christopher Lawrence, as well as two anonymous
reviewers and the editors of this volume. Friederike Talbot provided excellent research assistance
and feedback. We also benefited from valuable feedback from our pre-testers: Nicoleta Bazgan,
Steffen Beigang, Pauline Defant, Dominik Duell, Cosima Ingenschay, Dorina Kalkum, Thomas
Maruhn, Jochen Rehmert, Birgit Reinhold, Marc Reinhold, Carina Schmitt and Marcel Skaun.
Konstantin V€ossing acknowledges support from the Deutsche Forschungsgemeinschaft (DFG)
[German research foundation], grant number VO 1990/1-1.
Alvarez, R. M., Levin, I., Mair, P., & Trechsel, A. (2014). Party preferences in the digital age: The
impact of voting advice applications. Party Politics, 20(2), 227–236.
Andersen, R., Tilley, J., & Heath, A. F. (2005). Political knowledge and enlightened preferences:
Party choice through the electoral cycle. British Journal of Political Science, 35(2), 285–302.
Bafumi, J., & Shapiro, R. (2009). A new partisan voter. Journal of Politics, 71(1), 1–24.
Barabas, J., & Jerit, J. (2010). Are survey experiments externally valid? American Political
Science Review, 104(2), 226–242.
Brambor, T., Clark, W. R., & Golder, M. (2006). Understanding Interaction Models: Improving
Empirical Analysis. Political Analysis, 14(1), 63–82.
Campbell, A., Converse, P. E., Miller, W. E., & Stokes, D. E. (1960). The American voter.
New York: Wiley.
Carmines, E. G., & Stimson, J. A. (1980). The two faces of issue voting. American Political
Science Review, 74(1), 78–91.
Chaiken, S. (1980). Heuristic versus systematic information processing and the use of source
versus message cues in persuasion. Journal of Personality and Social Psychology, 39(5),
De Sio, L., & Weber, T. (2014). Issue yield: A model of party strategy in multidimensional space.
American Political Science Review, 108(4), 870–885.
De Vries, C. E., van der Brug, W., van Egmond, M. H., & van der Eijk, C. (2011). Individual and
contextual variation in EU issue voting: The role of political information. Electoral Studies, 30,
Dinas, E., & Pardos-Prado, S. (2012). A hidden giant? Exploring the centrifugal dynamics of
attitudes towards the European unification. Acta Politica, 47(4), 378–399.
Downs, A. (1957). An economic theory of democracy. New York: Harper.
Erikson, R., & Wlezien, C. (2012). The timeline of presidential elections: How campaigns do (and
do not) matter. Chicago: University of Chicago Press.
Fiorina, M. (1981). Retrospective voting in American national elections. New Heaven, CT: Yale
Gelman, A., & King, G. (1993). Why are American presidential election campaign polls so
variable when votes are so predictable? British Journal of Political Science, 23, 409–445.
Granberg, D., & Holmberg, S. (1988). The political system matters. Social psychology and voting
behavior in Sweden and the United States. Cambridge: Cambridge University Press.
Hillygus, D. S., & Shields, T. G. (2009). The persuadable voter: Wedge issues in presidential
campaigns. Princeton, NJ: Princeton University Press.
Hix, S., & Lord, C. (1997). Political parties in the European Union. New York: Macmillan.
Hobolt, S. B., & Wittrock, J. (2011). The second-order election model revisited: An experimental
test of vote choices in European Parliament elections. Electoral Studies, 30, 29–40.
ossing and T. Weber
Hooghe, L., & Marks, G. (1999). The making of a polity: The struggle over European integration.
In H. Kitschelt, P. Lange, G. Marks, & J. D. Stephens (Eds.), Continuity and change in
contemporary capitalism (pp. 70–97). Cambridge: Cambridge University Press.
Huckfeldt, R., Levine, J., Morgan, W., & Sprague, J. (1999). Accessibility and the political utility
of partisan and ideological orientations. American Journal of Political Science, 43(3),
Lazarsfeld, P. E., Berelson, B., & Gaudet, H. (1944). The people’s choice. New York: Duell, Sloan
Lijphart, A. (1968). The politics of accommodation. Berkeley, CA: University of California Press.
Lijphart, A. (1999). Patterns of democracy: Government forms and performance in 36 countries.
New Haven, CT: Yale University Press.
McGuire, W. (1985). Attitudes and attitude change. In G. Lindzey & E. Aronson (Eds.), Handbook
of social psychology (3rd ed., Vol. 2, pp. 233–346). New York: Random House.
Miller, W. E., & Shanks, J. M. (1996). The new American voter. Cambridge, MA: Harvard
Mutz, D. C. (2002). The consequences of cross-cutting networks for political participation.
American Journal of Political Science, 46(4), 838–855.
Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. Advances
in Experimental Social Psychology, 19, 123–205.
Powell, G. B., Jr. (1976). Political cleavage structure, cross-pressure processes, and partisanship:
An empirical test of the theory. American Journal of Political Science, 20(1), 1–23.
Rabinowitz, G., & Macdonald, S. E. (1989). A directional theory of issue voting. American
Political Science Review, 83(1), 93–121.
Reif, K., & Schmitt, H. (1980). Nine second-order national elections: A conceptual framework for
the analysis of European election results. European Journal of Political Research, 81, 3–45.
Tilley, J., & Wlezien, C. (2008). Does political information matter? An experimental test relating
to party positions on Europe. Political Studies, 56, 192–214.
Van der Brug, W., Franklin, M. N., & To´ka, G. (2008). One electorate or many? Differences in
party preference formation between new and established European democracies. Electoral
Studies, 27(4), 589–600.
Van der Eijk, C., & Franklin, M. (2004). Potential for contestation on European matters at national
elections in Europe. In G. Marks & M. Steenbergen (Eds.), European integration and political
conflict (pp. 32–50). Cambridge: Cambridge University Press.
Van der Eijk, C., van der Brug, W., Kroh, M., & Franklin, M. N. (2006). Rethinking the dependent
variable in voting behavior: On the measurement and analysis of electoral utilities. Electoral
Studies, 25(3), 424–447.
Vassil, K. (2011, August 25–27). Role of self selection in estimating the effects of voting advice
applications: Empirical Evidence on the basis of Swiss Smartvote Data. Paper presented at the
6th ECPR General Conference, Reykjavik, Iceland.
V€ossing, K. (2005). Nationality and the preferences of the European public toward EU policy
making. European Union Politics, 6(4), 447–469.
Weber, T. (2007). Campaign effects and second-order cycles: A top-down approach to European
Parliament Elections. European Union Politics, 8(4), 509–536.
Weber, T. (2009). When the cat is away the mice will play: Why elections to the European
Parliament are about Europe after all. Politique Europe´enne, 28, 53–71.
Zaller, J. (1992). The nature and origins of mass opinion. New York: Cambridge University Press.
Candidate Extremity, Information
Environments, and Affective Polarization:
Three Experiments Using Dynamic Process
Mona S. Kleinberg and Richard R. Lau
Research suggests that politics in the United States have noticeably polarized at the
elite level. Over the past several decades, the two major parties have become
ideologically more distinct (Jacobson 2005) and while there once were liberal
Republicans and conservative Democrats, today the most liberal Republican is to
the right of the most conservative Democrat (Poole and Rosenthal 2001). The two
major parties now differ not only on economic issues, but also on cultural, foreign
policy, racial, and religious stances (Layman 2001).
While there is research to support the notion of political polarization among
elected officials and between the parties (commonly referred to as “elite polarization”) the evidence is less clear when it comes to political polarization at the mass
level. Some argue that the attitudes of the general public have not followed political
polarization at the elite level and cite data showing that citizens are no more divided
on most social issues and policies today, than they were during earlier times in US
history (DiMaggio et al. 1996; Evans 2003; Fiorina et al. 2005), though there are
some exceptions to this rule. Citizens’ attitudes on abortion, for example, have
polarized to some extent, and there has been a small shift in ideology (Evans
et al. 2001).
Others argue that what has happened at the mass level over the past decades can
best be described as “sorting.” Citizens have brought their ideology and their
partisan identifications more in line with each other, so that the overall distribution
of opinion in the country has not changed much over the years, but its partisan
implications have (see Layman et al. 2006; and Fiorina and Abrams 2008, for recent
M.S. Kleinberg (*)
Department of Political Science, University of Massachusetts, Lowell, MA, USA
Department of Political Science, Rutgers University, New Brunswick, NJ, USA
© Springer International Publishing Switzerland 2016
A. Blais et al. (eds.), Voting Experiments, DOI 10.1007/978-3-319-40573-5_4
M.S. Kleinberg and R.R. Lau
A third perspective, which is the perspective we focus on in this chapter, is
grounded in social identity theory (Tajfel and Turner 1979) and asserts that the
masses have polarized much more on an emotional, or affective, level than they
have on a policy level. For example, research shows that the extent to which
partisans like their own party has remained fairly constant over the past 40 years,
yet how much they dislike the other party, has increased noticeably (Iyengar
et al. 2012). Classic “social distance” measures, such as being upset when a family
member marries someone with a different party affiliation, reveal similar trends.
1 The New Media Environment and Affective Polarization
Affective polarization is primarily driven by partisan identity, yet scholars speculate that the media environment also has an effect. Iyengar et al. (2012) demonstrate
that affective polarization is only partially a function of policy attitudes, and that
partisan identity is much more important in explaining the emotional disdain of the
opposing party. Moreover, the authors show that the ideological nature of the media
environment matters. Using survey data, the authors demonstrate that the number of
negative political ads that aired during the election in the respondent’s state changes
how much he or she likes or dislikes the opposing party. Iyengar et al. conclude by
suggesting that selective exposure to news sites and partisan blogs, a form of
one-sided or biased engagement with information on the Internet, is probably an
even more important cause of affective polarization, though they present no evidence to this point. In this chapter, we present data from three experiments to
examine the effects of the media environment on affective polarization.
The media landscape has changed dramatically since the “broadcast news” days
of the mid-twentieth century. In contrast to the era of network TV, when choice was
limited, we now live in a “high-choice” media environment (Prior 2007). While the
media environment offers many choices, and has become more diverse in many
ways, the news diet of many individuals has in fact become increasingly homogeneous. Research shows that viewers take advantage of the many choices available
to them by customizing their information content via selective exposure. Individuals now consume the “daily me,” as Sunstein (2008) puts it. Algorithms further
narrow media exposure by providing social media news feeds tailored to the
recipient. Hence, many people reside in a news “bubble,” without necessarily
knowing just how small their information landscape has become (Pariser 2011).
If partisans are disproportionately exposed to partisan (that is, ideologically
compatible) political information by consuming more information about their
preferred candidate, or by using media sources that reflect their ideology, then
increased affective polarization at the mass level seems like the inevitable outcome.
In this chapter we use three experiments (all three are simulated election campaigns) to test whether selective exposure to information and changes in the media
environment have an impact on affective polarization. In keeping with traditional
Candidate Extremity, Information Environments, and Affective Polarization:. . .
research, we also test the effects of political polarization at the elite level on
We test the impact of the Internet-era news environment on affective polarization
by experimentally manipulating the media environment in our control and treatment groups. To test the effects of increasing polarization at the elite level, we vary
the ideological extremity of the competing political candidates. We ask first,
whether either of these manipulations influence selective exposure to information,
resulting in biased information search; and second, whether either of these manipulations by themselves, or the presumed mediating factor, biased information
search, contribute to greater affective polarization.
We measure selective exposure by tracking how individuals search for information (e.g. what information items they select and read, and what they ignore). We
conceptualize biased engagement with information as spending more time with
information about the in-party candidate than the out-party candidate.1 Although a
quantitative measure, we take the amount of search about one’s preferred candidate
to also reflect a qualitative dimension because we assume that the amount searched
for the in-party candidate reflects desire for qualitatively good or favorable information. A balanced search would be retrieving an equal amount of information
about both candidates. Affective polarization is defined as the magnitude (absolute
value) of the difference between the post-election feeling thermometer (0–100)
evaluations of the two competing candidates.2
We expect affective polarization to be greatest among “sorted” individuals (that
is, people whose partisan and ideological identifications are in line) and by subjects
who engage in biased information search. Furthermore, we expect it to be exacerbated by polarization at the elite level, and by a more partisan media environment.
Specifically, we test the following three major hypotheses:
In our experiments this conceptualization makes the most sense. An alternative conceptualization
would track what sources subjects select and if they prefer a certain ideological orientation. Given
our particular experiments however, we chose a different conceptualization that measured selective exposure to content (e.g. the candidate) though we encourage others to conceptualize selective
The feeling thermometer question taps into how warm or how cold individuals feel toward
candidates. Thus, it is an affective measure because it taps into people’s feelings toward candidates. Respondents rate how they feel about a candidate using a “feeling thermometer,” which is a
scale of 0–100. Higher numbers represent warmer and more favorable feelings towards the person,
and lower numbers represent colder or less favorable feelings. Feeling thermometers are common
in American Politics research (see for example the ANES time series). We modeled our feeling
thermometer question after the ANES.
M.S. Kleinberg and R.R. Lau
• H1: Ideological sorting leads to both biased information search about the preferred candidate, and affective polarization in evaluations.
• H2: Elite polarization leads to both biased information search about the preferred
candidate and affective polarization in evaluations.
• H3: A more diverse and partisan media environment heightens both biased
information search about the preferred candidate, and affective polarization in
This chapter proceeds as follows. First, we provide a brief introduction to the
Dynamic Process Tracing Environment experimental platform, as it was utilized to
program all three of our experiments. Then we describe in turn three different
experiments that provide valuable insight into the processes by which political
polarization may occur. A final section summarizes our results across these three
different experiments and provides suggestions for future research.
A common criticism of experiments is that they can demonstrate that a certain
relationship exists, but not why it exists. Static approaches (e.g. surveys, most
standard experiments) limit the extent to which we can study how the information
environment, political candidates, and information search interact. The Dynamic
Process Tracing Environment (DPTE) platform, developed by Lau and Redlawsk
(2006), is designed to answer the “why?” question, to observe decision makers as
they are making a decision.
DPTE is a computer-based system that runs in any browser and presents lab
(or online) subjects with a simulated social environment such as an election
campaign that unfolds over time (Lau 1995; Lau and Redlawsk 2001). To enter
the experiment, participants log on to DPTE from a lab, or their home computers. In
a typical election experiment, subjects first read a description of the study, acknowledge a consent form, and answer standard survey questions. Next, participants enter
the campaign phase during which they are offered various pieces of information
about the candidates in various formats (e.g. via text, images, short video clips,
etc.). The experimenter can control how much new information to introduce at a
time. While some information can be programmed to be seen by all participants
(a common approach is to show those items as a pop-up window), other items
(information boxes) scroll down the computer screen and are up to the subject’s
discretion whether to select. Similar to search engine results, information boxes
usually bear a headline and a short description of the content, offering just enough
information to allow subjects to decide whether to click on and read the full text, or
to ignore it. Information boxes can be customized to bear a logo, providing a cue
about its source (e.g. a news organization, the candidate’s website, etc.). After the
campaign phase of the experiment, subjects vote, are given a post-questionnaire,
and are debriefed. Please see Fig. 1 for a screenshot of information boxes in DPTE.