FlashReports
Political ideology, exploration of novel stimuli, and attitude formation
Natalie J. Shook a,*, Russell H. Fazio b
a
Virginia Commonwealth University, Department of Psychology, P.O. Box 842018, Richmond, VA 23284-2018, United States
b
Ohio State University, 1835 Neil Avenue, Columbus, OH 43210, United States
a r t i c l e i n f o
Article history:
Received 23 January 2009
Revised 19 March 2009
Available online xxxx
Keywords:
Attitudes
Political ideology
Conservatism
Liberalism
Exploratory behavior
a b s t r a c t
In this study, the relations among political ideology, exploratory behavior, and the formation of attitudes
toward novel stimuli were explored. Participants played a computer game that required learning whether
these stimuli produced positive or negative outcomes. Learning was dependent on participants’ decisions
to sample novel stimuli and discover the associated valence. Political ideology correlated with exploration
during the game, with conservatives sampling fewer targets than liberals. Moreover, more conservative
individuals exhibited a stronger learning asymmetry, such that they learned negative stimuli better
than positive. Mediational analyses revealed that the differences in learning were due to the extent of
exploratory behavior during the game. Relative to liberals, politically conservative individuals pursued
a more avoidant strategy to the game, which led to their development of a more pronounced valence
asymmetry in learning and attitude formation.
Ó 2009 Elsevier Inc. All rights reserved.
Over the past 50 years, a substantial amount of research has focused
on distinguishing conservatism or right-wing ideology from
liberalism or left-wing ideology (see Jost, Nosek, & Gosling, 2008,
for a review). The goal of much of this work has been to identify
personality traits, values, and motives that differentiate these
political ideologies. Generally speaking, conservatives are viewed
as relatively traditional, dogmatic, and conforming, whereas liberals
are viewed as more unconventional and flexible. Measures of
conservatism have been associated with self-reported intolerance
of ambiguity (Sidanius, 1978), greater need for closure (Webster
& Kruglanski, 1994), and greater perceptions of the world as dangerous
(Altemeyer, 1998). Liberalism has been associated repeatedly
with the Openness to Experience dimension of the Big Five
(Carney, Jost, & Gosling, 2008; Jost, 2006). Liberals also score more
highly on sensation-seeking scales (Jost, Glaser, Kruglanski, & Sulloway,
2003) and report being more egalitarian (Conover & Feldman,
1981).
Relatedly, it has been proposed that political ideology may be
associated with differences in cognitive style (e.g., Adorno, Frenkel-Brunswik,
Levinson, & Sanford, 1950). Conservatives may come
to understand and organize their world in a more structured and
rigid way, whereas liberals appear to be more open to complexity.
To test this hypothesis, Tetlock (1983) coded 20 policy statements
from US senators for the complexity of the arguments. Conservative
senators exhibited less complex cognitive styles than liberal
senators. That is, they were more ‘‘prone to rely on simple evaluative
(good versus bad) rules in interpreting policy issues” (p. 123).
Moreover, conservatives have been found to be more rigid in their
interpretation of situations and more severe in their judgments of
others. In a series of studies, Tetlock and colleagues (2007) found
conservatives to hold perpetrators of crimes more responsible for
their acts and to endorse harsher punishments. Thus, more conservative
individuals see the world as involving fewer ‘‘shades of
gray” and also are more resolute in their judgments of others.
Wilson, Ausman, and Mathews (1973) further suggested that
the psychological basis for political ideology may rest in the extent
to which individuals fear the unknown. To test this possibility, participants
rated a number of paintings that varied along two dimensions:
simplicity–complexity and representational–abstract. More
complex and abstract paintings were intended to represent less
familiarity and more ambiguity. Conservative participants liked
simple representational paintings the most, whereas liberal participants
preferred the complex and abstract paintings. The authors
interpreted these results as reflecting liberals’ greater tolerance
of uncertainty or unfamiliarity.
More recent work adds to the proposition that conservatives
display greater sensitivity to fear and threats. Lavine and colleagues
(1999) presented individuals with a message promoting
voting behavior (an ideologically neutral topic) and manipulated
message framing. The appeal was framed in terms of either the
benefits of voting (e.g., ‘‘Voting allows one to be heard.”) or the loss
associated with not voting (e.g., ‘‘Voting prevents your values from
being undermined.”). Conservative participants rated the loss
framed message as more persuasive and valid than the benefit
framed message, whereas the opposite was true for liberal participants.
The results may highlight a fundamental sensitivity to
threats of loss among more conservative individuals.
0022-1031/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jesp.2009.04.003
* Corresponding author.
E-mail address: njshook@vcu.edu (N.J. Shook).
Journal of Experimental Social Psychology xxx (2009) xxx–xxx
Contents lists available at ScienceDirect
Journal of Experimental Social Psychology
journal homepage: www.elsevier.com/locate/jesp
ARTICLE IN PRESS
Please cite this article in press as: Shook, N. J., & Fazio, R. H. Political ideology, exploration of novel stimuli, and attitude formation. Journal of Experimental
Social Psychology (2009), doi:10.1016/j.jesp.2009.04.003
Thus, the available literature has outlined personality and motivational
characteristics associated with political ideology. However,
the evidence to date relies heavily on self-reports on traitrelated
inventories or inferences from participants’ stated preferences.
Questions still remain as to the broader implications of
these tendencies for basic cognitive processes and behavioral decisions.
If conservatives fear the unfamiliar and are relatively more
sensitive to threat, whereas liberals are more open to novel experiences,
then there may be fundamental differences in how individuals
with varying ideologies approach their social world, acquire
information, and form attitudes.
In a series of experiments, Fazio, Eiser, and Shook (2004) examined
how individuals explore their environment and form attitudes
toward novel objects. The research involved a computer game,
BeanFest, in which participants imagined themselves in a world
of beans. In order to succeed at the game, participants had to learn
which beans were good (i.e., produced positive outcomes when approached)
and which were bad (i.e., produced negative outcomes).
Interestingly, participants consistently exhibited a learning asymmetry.
Negative attitudes were learned better than positive
attitudes.
The learning asymmetry was found to be largely the result of
sampling behavior during the game. One of the BeanFest parameters
involved feedback being contingent upon approach behavior. If
approached, the true nature of the specific bean was revealed, because
selection affected the participant’s points. If participants
chose to avoid a bean, however, they learned nothing regarding
its value. Thus, any negative misconceptions that participants
developed would be corrected only if they were willing to risk further
sampling. Mediational findings indicated that individuals who
engaged in more approach behavior during the game learned more
about the true nature of each bean and exhibited a smaller learning
asymmetry. Moreover, an experimental variation in which participants
were provided full feedback (information about what would
have happened if the bean had been selected) significantly reduced
the learning asymmetry, to the point that it was not statistically
evident.
If conservatives are more fearful of the unknown and liberals
more open to new experiences, then political ideology may relate
to attitude formation via direct experience. Politically conservative
individuals may be relatively more cautious, limiting their exposure
to novel targets, whereas liberals may approach new stimuli
more indiscriminately. If true, liberals will expose themselves to
a greater number of potentially negative stimuli, but will gain
information from these experiences. The more avoidant behavior
of conservatives will limit their negative experiences, but may lead
to reduced information gain and, in particular, a lesser probability
of correcting invalid assumptions of negativity. Thus, conservatives
may come to hold a greater number of negative attitudes (and,
hence, a less balanced perception of their world) than liberals.
The current research uses the BeanFest paradigm to explore the
relation between political ideology and fundamental processes of
attitude formation via exploratory behavior. We expected participants
who endorsed political conservatism to be less exploratory
during the game (i.e., approach fewer game beans).1
This less frequent
sampling should, in turn, affect learning and attitude formation.
Hence, more conservative individuals were expected to
display a more pronounced valence asymmetry than more liberal
participants.
Method
Participants
Fifty-eight introductory psychology students (67% male) participated
for course credit. Average age was 19.38 years old
(SD = 2.19; range = 17–30).
Measures
Political ideology was assessed via two measures. The first concerned
agreement with 13 belief statements involving clearly identifiable
ideological stances (e.g., ‘‘Abortion is wrong, because
everyone, even unborn babies, has the right to life.”). Participants
responded on a À2 (disagree strongly) to +2 (agree strongly) scale,
with some items later being reverse-scored so that higher numbers
reflected stronger agreement with conservative positions
(M = À.18, SD = .58). The measure displayed satisfactory internal
consistency (a = .71). Participants also indicated their political
self-identification on a scale ranging from Liberal (1) to Conservative
(7) (M = 3.97, SD = 1.73). The two measures correlated highly,
r(56) = .59, p < .001, so a single composite was created by averaging
the two standard scores.2
BeanFest
The participant’s goal in the BeanFest game is to accumulate
points by making judicious decisions about which beans to accept
(approach) and which to reject (avoid). Beans have either a positive
or negative value. When accepted, the participant’s point value,
which ranges from 0 to 100, is adjusted according to the bean’s value.
If rejected, the participant’s points remain unchanged. However,
the participant does not receive any information about the
rejected bean.
Participants begin the game with 50 points. Reaching 100 represents
winning the game, and reaching zero represents losing.
In either case, the game restarts at 50 points. Although the number
of games played depends on the individual’s success, participants
complete the same number of trials (108).
The beans differ by shape (10 levels from circular to oval to oblong)
and number of speckles (1–10). During the game, 36 beans,
from six different regions of the 10 Â 10 matrix, are presented
(see Fazio et al., 2004, for details). These were carefully selected,
and assigned point values of +10 or À10, in such a way so as to
avoid any linear relations between shape or number of speckles
and bean valence. Participants must associate each type of bean
with the outcome it produces. To increase participants’ engagement,
a monetary incentive was provided. The payoff scheme
was structured such that participants would receive $1.00 for each
win. For each loss, 50¢ would be deducted. Participants could receive
$0–$10.
Procedure
Upon arrival, participants were seated in individual cubicles
and provided instructions for BeanFest. The game began with a
practice block of six trials. The actual game was divided into three
blocks of 36 trials; thus, all 36 beans were seen three times. During
a trial, participants had to indicate whether they wanted to accept
or reject the presented bean by pressing one of two buttons on a
response box.
After the game, participants completed the test phase, which
assessed their learning of the beans. Participants were presented
1
Our hypotheses were informed by the political ideology literature and by a
serendipitous finding. A BeanFest study was conducted with participants who had
completed a political ideology scale in an unrelated study. We found a correlation
between approach behavior and ideology, r(60) = À.25, p < .05. More conservative
participants approached fewer beans during the game. This finding encouraged the
current study.
2
All analyses were performed separately with the political beliefs scale and the
identification measure. Similar results were found with each measure for both the
correlational and mediational analyses.
2 N.J. Shook, R.H. Fazio / Journal of Experimental Social Psychology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article in press as: Shook, N. J., & Fazio, R. H. Political ideology, exploration of novel stimuli, and attitude formation. Journal of Experimental
Social Psychology (2009), doi:10.1016/j.jesp.2009.04.003
with the beans in a random order and indicated whether they believed
each bean to be ‘‘good” (i.e., increased points during the
game) or ‘‘bad” (i.e., decreased points during the game). Following
the test phase, participants completed a series of questionnaires
for an ostensibly separate study, including the political ideology
measures. Participants were then paid and debriefed.
Results
First, participants’ learning of the game beans was examined.
Phi coefficients were computed assessing the relation between
the bean valence (positive/negative) and a participant’s classification
of that bean during the test phase (positive/negative). The
average phi coefficient was .30, which is much better than chance,
t(57) = 9.09, p < .001, indicating that, on average, participants did
learn. To determine whether learning varied by valence, as in previous
studies, the proportion of positive and negative beans correctly
classified in the test phase was calculated. Overall, learning
was well above chance for both positive beans (M = .60),
t(57) = 4.69, p = .001, and negative beans (M = .69), t(57) = 6.83,
p < .001. However, the negative beans were learned better than
the positive beans, t(57) = 2.24, p < .05, indicating a learning
asymmetry.
Next, we examined the relations between the BeanFest indices
and political ideology. The correlations are presented in Table 1.
More conservative participants approached fewer beans during
the game. Throughout the game, politically conservative participants
pursued a more cautious strategy. Although political ideology
did not relate to overall learning, it did predict the learning
asymmetry (see Fig. 1). That is, more conservative participants
exhibited a larger difference in their learning of the negative versus
positive beans (in favor of the negative).
To determine if the relation between political ideology and the
learning asymmetry was due to participants’ game behavior, mediational
analyses were performed. The average level of approach
behavior during the game correlated with the learning asymmetry,
r(56) = À.60, p < .001; the more one approached, the smaller the
learning asymmetry. Controlling for approach behavior, the correlation
noted earlier between political ideology and the learning
asymmetry was no longer significant, r(56) = .13, p > .30. A Sobel
test (MacKinnon & Dwyer, 1993) revealed that the reduction in
the correlation was significant, z = 2.16, p < .05. Thus, the relation
between political ideology and the learning asymmetry was mediated
by participants’ game behavior.
Discussion
Utilizing the BeanFest paradigm (Fazio et al., 2004), the present
study explored the relation between political ideology and attitude
formation through direct experience. Political ideology was found
to correlate with exploration in a novel situation. When confronted
with novel objects that could be beneficial or harmful, politically
conservative individuals tended to approach fewer objects than
more liberal individuals. That is, conservatives adopted a more
cautious strategy to playing the BeanFest game and learning about
new targets.
Participants’ decisions during the game ultimately had consequences
for their learning of the game beans. Although there was
no difference in overall learning, politically conservative participants
demonstrated a greater disparity in learning of negative
and positive beans than liberal participants. That is, political ideology
predicted the learning asymmetry. Moreover, the correlation
between the learning asymmetry and political ideology was mediated
by approach behavior during the game. Participants’ game
strategies affected the information they gained about the bean
world and, thus, the attitudes that they formed. As conservatives
chose a more avoidant game strategy, they did not gain as much
information about the beans, and any negative misconceptions
were not corrected by subsequent sampling. Consequently, more
conservative individuals tended to correctly identify more negative
beans, but to miscategorize relatively more positive beans as negative.
In effect, they overestimated the number of negative beans in
the world. As liberals were willing to approach more beans, they
gained more information about the beans and did not exhibit as
strong of a learning asymmetry. They developed a more balanced
perception of the world.
The findings from this study highlight some of the broader correlates
of political ideology. Conservatives’ intolerance of the unfamiliar
(Wilson et al., 1973), perceptions of the world as dangerous
(Altemeyer, 1998), and fear of loss (Lavine et al., 1999) were reflected
in the cautious strategy adopted when playing BeanFest
and learning about the novel objects. Liberals demonstrated more
openness to new experiences (Jost, 2006) by exploring the new
bean world to a greater extent. These different approaches to interacting
with one’s environment led to differences in attitude formation
and participants’ perceptions of the bean world.
The present study is limited by its correlational nature. Future
longitudinal research could explore the development of conservatism
versus liberalism and thus more clearly assess the relation between
political ideology and willingness to explore. Nevertheless,
this study provides clear evidence for the existence of relations
among political ideology, exploratory behavior, and attitude formation.
Moreover, the fundamental asymmetry between approach
Table 1
Correlations between BeanFest indices and political ideology.
Game behavior and learning indices Political ideologya
Overall game approach behavior À.30*
Block 1 approach À.27*
Block 2 approach À.27*
Block 3 approach À.25+
Overall learningb
.08
Learning asymmetryc
.28*
a
Higher numbers reflect greater conservatism.
b
Phi coefficient between actual valence of bean and participant’s classification of
the bean during the test phase.
c
Proportion of negative beans correctly classified minus proportion of positive
correctly classified.
+
p < .08.
*
p < .05.
Fig. 1. Scatterplot and regression line relating political ideology (higher numbers
reflect greater conservatism) to the learning asymmetry (proportion of negative
beans correctly classified minus proportion of positive correctly classified).
N.J. Shook, R.H. Fazio / Journal of Experimental Social Psychology xxx (2009) xxx–xxx 3
ARTICLE IN PRESS
Please cite this article in press as: Shook, N. J., & Fazio, R. H. Political ideology, exploration of novel stimuli, and attitude formation. Journal of Experimental
Social Psychology (2009), doi:10.1016/j.jesp.2009.04.003
and avoidance has significant implications for an individual’s
worldview. The reluctance to explore that characterizes more
politically conservative individuals may protect them from experiencing
negative situations, for they are likely to restrict approach
to known positives. However, as a result of information gain being
dependent on sampling, this reluctance also promotes a presumption
that the world is a relatively harsh place. Any invalid negative
attitudes that they might develop are unlikely to be tested by further
sampling. Hence, more politically conservative individuals believe
the world to consist of more objects of which they must be
wary than objects they desire. The greater approach behavior of
more liberal individuals results in more information gain, but
leaves them vulnerable to negative experiences as they sample objects
about which they may be uncertain. Liberals may be more
likely to find themselves in unsafe situations than conservatives.
Thus, a more balanced worldview may come at the cost of experiencing
more negative events.
References
Adorno, T., Frenkel-Brunswik, E., Levinson, D., & Sanford, R. (1950). The authoritarian
personality. Oxford England: Harpers.
Altemeyer, R. A. (1998). The other ‘‘authoritarian personality”. Advances in
Experimental Social Psychology, 30, 47–91.
Carney, D. R., Jost, J. T., & Gosling, S. D. (2008). The secret lives of liberals and
conservatives: Personality profiles, interaction styles, and the things they leave
behind. Political Psychology, 29, 807–840.
Conover, P. J., & Feldman, S. (1981). The origin and meaning of liberal/conservative
self-identification. American Journal of Political Science, 25, 617–645.
Fazio, R. H., Eiser, J. R., & Shook, N. J. (2004). Attitude formation through exploration:
Valence asymmetries. Journal of Personality and Social Psychology, 87, 293–311.
Jost, J. T. (2006). The end of the end of ideology. American Psychologist, 61, 651–670.
Jost, J. T., Glaser, J., Kruglanski, A. W., & Sulloway, F. (2003). Political conservatism as
motivated social cognition. Psychological Bulletin, 129, 339–375.
Jost, J. T., Nosek, B. A., & Gosling, S. D. (2008). Ideology: Its resurgence in social,
personality, and political psychology. Perspectives on Psychological Science, 3,
126–136.
Lavine, H., Burgess, D., Snyder, M., Transue, J., Sullivan, J., Haney, B., et al. (1999).
Threat, authoritarianism, and voting: An investigation of personality and
persuasion. Personality and Social Psychology Bulletin, 25, 337–347.
MacKinnon, D. P., & Dwyer, J. H. (1993). Estimating mediated effects in prevention
studies. Evaluation Review, 17, 144–158.
Sidanius, J. (1978). Intolerance of ambiguity and socio-politico ideology: A
multidimensional analysis. European Journal of Social Psychology, 8, 215–235.
Tetlock, P. E. (1983). Cognitive style and political ideology. Journal of Personality and
Social Psychology, 45, 118–126.
Tetlock, P. E., Visser, P. S., Singh, R., Polifroni, M., Scott, A., Elson, S. B., et al. (2007).
People as intuitive prosecutors: The impact of social-control goals on
attributions of responsibility. Journal of Experimental Social Psychology, 43,
195–209.
Webster, D., & Kruglanski, A. (1994). Individual differences in need for cognitive
closure. Journal of Personality and Social Psychology, 67, 1049–1062.
Wilson, G., Ausman, J., & Mathews, T. (1973). Conservatism and art preferences.
Journal of Personality and Social Psychology, 25, 286–288.
4 N.J. Shook, R.H. Fazio / Journal of Experimental Social Psychology xxx (2009) xxx–xxx
ARTICLE IN PRESS
Please cite this article in press as: Shook, N. J., & Fazio, R. H. Political ideology, exploration of novel stimuli, and attitude formation. Journal of Experimental
Social Psychology (2009), doi:10.1016/j.jesp.2009.04.003