Psychological Dispositions, Political Engagement, and Social Welfare
Christopher D. Johnston
Department of Political Science
Duke University
christopher.johnston@duke.edu
Acknowledgements: I would like to thank Stanley Feldman, Howard Lavine, Matthew Lebo,
Chris Federico, Ben Newman, and James Wilfred Jenkins III for comments and discussions on
earlier versions of this manuscript.
ABSTRACT
INTRODUCTION
Recent theorizing on the relationship of stable dispositions (e.g. personality traits,
psychological needs, core human values) to political ideology has largely assumed a model
whereby a set of core traits relate in similar ways to values and preferences across ideological
domains, thus providing a “bottom-up” structure to mass belief systems (e.g. Jost et al. 2003;
Jost, Federico and Napier 2009; Mondak 2010). In other words, the same traits which predispose
one to find an affinity with (say) conservative ideas in the social realm also structure positive
responses to conservative ideas in the economic realm. Mondak (2010), for example,
highlighting the consistent findings of associations between openness to experience and
conscientiousness with liberalism and conservatism, respectively (see, e.g., Carney et al. 2008;
Jost et al. 2003; McCrae 1996; Thorisdottir et al. 2007), theorizes, “the openness and
conscientiousness hypotheses are best understood in terms of traditional views in which
liberalism corresponds with a willingness to see government tackle new and varied problems,
while conservatism implies a more cautious approach in which presumption favors the status
quo” (p. 127). While the extent of such bottom-up constraint is, perhaps, contingent on the
citizen’s degree of exposure to ideological ideas (Federico et al. 2011; Federico and Goren 2010;
Jost, Federico and Napier 2009), the qualitative nature of the relationship is expected to be the
same across domains according to these theories.1
This theoretical perspective is appealing as it represents an alternative to the historically
dominant, “top-down” approach to understanding the emergence of ideological constraint in the
mass public (Converse 1964; Zaller 1992). If citizens’ stable traits structure their response to
Some might quibble with this portrayal of the literature, noting that examinations of the relationship of “Big Five”
traits to issues across domains have yielded some variation, for example, with respect to the effect of agreeableness
on economic preferences (Gerber et al. 2010). Such research remains the exception rather than the rule, and the
major theoretical treatments within this literature have highlighted common antecedents of social and economic
conservatism (Feldman and Johnston forthcoming).
1
political ideas at the level of actual substance (as opposed to more superficial features such as
partisan labels), then the emergent structure of ideology is, perhaps, more normatively pleasing
than what is implied by the top-down approach (e.g. Shapiro and Jacobs 2010). Despite its
appeal, and apparent parsimony, empirical anomalies cast doubt on the sufficiency of this
theoretical framework for understanding mass ideology. Most importantly, variation exists in the
degree of empirical support for the relationship of core psychological dispositions to political
ideology across domains, with support much stronger for relationships to social and cultural
conservatism than to economic conservatism (e.g. Carney et al. 2008; Feldman and Johnston
forthcoming; Thorisdottir et al. 2007; Van Hiel, Pandelaere and Duriez 2004).
A recent response to this apparent counter-evidence has been to posit political
engagement (e.g. sophistication, attention to elite discourse) as a limiting factor to the emergence
of cross-dimensional constraint (Federico et al. 2011; Jost, Federico and Napier 2009). From this
perspective, we should expect the same dispositions to structure ideology in all domains, but
exposure to ideological ideas is a precondition to such structure, and thus should be observed
more readily among the politically sophisticated and attentive. This model is thus a hybrid of the
bottom-up and top-down perspectives: ideology is constrained in meaningful ways by stable
psychological dispositions, but ideology itself is a construction of elites and is thus obtained
through exposure to elite discourse.
While an important observation, and one which will be leveraged below, the present
paper argues that this theoretical extension is only one slice of a more important theoretical point
regarding the translation of psychological dispositions into political preferences and ideology,
namely, the fact that citizens may apprehend issues and broad issue domains by means of more
than one organizing idea; in other words, citizens may represent political issues in multiple ways
(e.g. Conover and Feldman 1984). If such variety in representation is psychologically relevant, in
the sense that it different views resonate with dispositions in different ways, it suggests the
potential for variation in the qualitative relationship of dispositions to preferences, not just the
extent of the relationship observed. In a sense, political psychologists have imposed a single
understanding of issues and ideology on the citizenry, and have thus neglected potential
heterogeneity in the qualitative relationship of dispositions to preferences.2 The present paper
argues that this neglect is at the heart of the empirical anomalies above, and moving beyond this
approach represents a first step toward understanding the psychology of preference formation
and ideological constraint at a deeper level.
In the following sections I argue that the relative lack of support for theoretical
expectations in the economic domain is a result of the theoretical imposition of a single
representation of these issues on the public by theorists, and thus a failure to consider a distinct
view which has very different implications for the translation of psychological dispositions into
liberal and conservative orientations within the economic domain. I first review recent literature
on the dispositional influences on political ideology, highlighting a common theme regarding the
importance of needs for certainty and security to the development of conservative political
preferences. I then argue that at least two core understandings of economic policy are extant in
the mass public: a symbolic view which sees economic policy in terms of institutional stability
and change (Jost et al. 2003), and a more concrete view, which sees economic policy in terms of
social protection against the risks intrinsic to free market capitalism. I argue that these two views
One potential exception is Gerber et al. (2010), who examine variation in the influence of “Big Five” traits
between Blacks and Whites. They make a similar argument to the one made in the present paper, namely, that the
differential economic experiences of Blacks relative to Whites should alter the way the economic domain is
understood for the former, and thus the role of personality in structuring preferences. Even here, however, the
argument is largely one of extent, as the relationship for Blacks is expected to be minimal relative to Whites.
2
suggest exactly opposite relationships of conservatism to epistemic and existential needs which,
when ignored, manifests as minimal or null relationships in aggregate statistical tests.
When seen through the lens of institutional stability and change, needs for certainty and
security resonate with economic conservatism and its emphasis on traditional American values
and hierarchies. When seen through the lens of social protection, such needs resonate with
economic liberalism and its provision of a safety net against income loss in the face of adverse
labor market conditions. Finally, I argue that variation in how the economic domain is
understood should manifest as differences across levels of political engagement, with highly
engaged citizens showing the former relationship, and the unengaged showing the latter.
I test these hypotheses empirically with data from three years of American politics, and
with two distinct measures of psychological needs for certainty and security, finding strong
support across all years and tests. In the last section of the paper I consider the broader
implications of these dynamics for American politics, and for the provision of social welfare in
this country, tying my findings to recent debates over the importance of the “culture war.” As
needs for certainty and security are negatively associated with socioeconomic status, I suggest
that my theoretical model implies that low SES White citizens should become increasingly
economically conservative as they become more politically engaged, contrary to what might be
expected on the basis of “self-interest” models of preference formation. This counter-intuitive
result follows directly from the model, and suggests that culture war theorists (e.g. Frank 2004),
highlighting the potential decline in social welfare provision as a function of the move of lower
class Whites into conservative camps, may be more correct in their insights than political
scientists have been willing to admit (e.g. Ansolabehere, Rodden and Snyder 2006; Bartels 2006;
2008; Gelman 2008). While, as these latter scholars note, the economic domain has surely not
been displaced by the rise of cultural conflict, the interaction of citizens’ stable traits with elitelevel politics may shape their economic ideologies in a way which leads low SES Whites, for
better or worse, to support conservative economic policies.
RESEARCH ON DISPOSITIONS AND IDEOLOGY
Research within the dispositional paradigm has been quite diverse, utilizing concepts
from biology and genetics (e.g. Alford, Funk and Hibbing 2005; Amodio et al. 2007; Settle et al.
2010), personality and social-cognitive psychology (e.g. Carney et al. 2008; Gerber et al. 2010;
Jost et al. 2003; Mondak 2010), and core values (e.g. Smith et al. 2011; Thorisdottir et al. 2007).
Most of this work, however, despite the diversity of language, converges on a common
understanding of the psychology underlying political orientations and ideology. This conceptual
convergence is made explicit by John Jost and his colleagues within their theory of political
ideology as motivated social cognition (Jost et al. 2003; Jost, Federico and Napier 2009). They
argue that political conservatism (as opposed to liberalism), in its emphasis on institutional
stability, and in its provision of a justification for the socioeconomic status quo, is palliative for
citizens with psychological needs for epistemic certainty and existential security. As Jost and
Hunyady (2002) argue, “There is a good match between needs to reduce uncertainty and threat
and system justification, because preserving the status quo allows one to maintain what is
familiar while rejecting the uncertain prospect of social change. For many people, the devil they
know seems less threatening and more legitimate than the devil they don’t” (p. 262).
This general hypothesis finds empirical support in a number of studies. Jost et al. (2003),
in a meta-analysis of over eighty studies on the topic, find that psychological indicators related to
these needs (e.g. intolerance of ambiguity, mortality salience) are associated with several
indicators of political conservatism. More recent work also shows converging evidence.
Hetherington and Weiler (2009) find that needs for security and certainty (in the form of
“authoritarianism,” see also Feldman 2003; Stenner 2005) are strongly related to preferences
over social and foreign policy, and have become an increasingly important predictor of partisan
orientations in the American mass public since 2000. Political scientists working within the “Big
Five” personality framework have found strong and consistent relationships between the traits of
openness to experience and conscientiousness and political ideology (Gerber et al. 2010;
Mondak 2010), which is notable as these traits are similar conceptually to the needs identified by
Jost and colleagues (see also Carney et al. 2008; McCrae 1996; Thorisdottir et al. 2007).
Nonetheless, recent work fails to provide a full account of political ideology in the
contemporary American public. Of particular importance is the tendency of this research to
ignore or downplay the multidimensionality of ideology (Feldman and Johnston 2009; Treier and
Hillygus 2009; but see Gerber et al. 2010), or in other words, the observation that political
preferences across ideological domains (e.g. economic, social) are only moderately related to one
another. While such ideological “constraint” has increased in recent years (Abramowitz 2010;
Levendusky 2009), it has not achieved a level where a single, left-right continuum constitutes an
accurate representation of mass preferences.
Moreover, empirical research that has explicitly distinguished ideological domains in
investigations of dispositional antecedents has generally shown variation in support for
expectations. More specifically, while a great deal of work finds strong associations between
needs for certainty and security and social and cultural preferences (Jost et al. 2003), and in more
recent years, preferences over foreign policy (Hetherington and Suhay 2011; Hetherington and
Weiler 2009), such needs show little association with conservatism in the economic domain (e.g.
Carney et al. 2008; Chirumbolo et al. 2004; Feldman and Johnston forthcoming; Van Hiel and
Mervielde 2004; Van Hiel, Pandelaere and Duriez 2004), or show associations in the opposite
direction expected by extant theory (Golec 2001; Johnston XXXX; Thorisdottir et al. 2007).
Studies which do find support for expectations in the economic domain share
characteristics which suggest the need for a more comprehensive model. Gerber et al. (2010) find
that both openness and conscientiousness are associated with economic preferences, but utilize a
sample of registered voters which are likely distinct on both observable and unobservable
characteristics related to political engagement. Federico et al. (2011) find that political
sophistication strongly moderates the relationship of needs for certainty and order on
egalitarianism, a particularly symbolic indicator of economic preferences. Finally, studies within
psychology (reviewed by Jost et al. 2003) which find support for expectations have most often
utilized Social Dominance Orientation (Sidanius and Pratto 2000) as their operationalization of
choice; a construct related to the economic domain in a highly symbolic fashion. These findings
suggest that both political engagement and the symbolic nature of the operationalization utilized
may heighten the relationship of these dispositions to conservatism. In the following section, I
present a broader theoretical model which can subsume these disparate findings.
THEORY
The basis of the claim that psychological needs for certainty and security will promote
economic conservatism lies in an abstract and symbolic representation of this domain. More
specifically, as argued by Jost et al. (2003), economic conservatism qua support for the
institutional status quo, in terms of both traditional American values related to individualism, and
in terms of prevailing socioeconomic hierarchies, should be palliative for citizens averse to
uncertainty and change. For such citizens, liberal innovation in the form of government programs
to redistribute wealth and provide social welfare should be seen as threatening in the same way
that liberal policy in the social domain (e.g. support for gay marriage) will be seen as threating to
the sociocultural status quo. In both cases, citizens high in needs for certainty and security prefer
the stability and predictability of traditional institutions and values. Jost and colleagues argument
is stated nicely by Gerber et al. (2010) in their application of the “Big Five” to political
preferences, “It follows that this attraction to novelty and tolerance for complexity encourage not
only overall liberalism, but also support for liberal social and economic policies, which typically
involve new programs or interventions that overturn existing practices” (p. 116).
This is only one possible understanding of the economic domain, however. A second way
of viewing government intervention into the economic domain concerns the social protective
functions of redistribution and social welfare (e.g. Moene and Wallerstein 2001; Sinn 1995).
Social welfare, by redistributing wealth from the rich to the poor, smooths the income
distribution, which reduces the variance in expected income and thus serves as a form of
insurance. As Sinn (1995) argues, “Every insurance contract involves a redistribution of
resources from the lucky to the unlucky, and most of the redistributive measures of the state can
be interpreted as insurance if the time space between judging and taking these measures is
sufficiently long” (p. 495-496). Furthermore, social welfare programs such as unemployment
insurance or government provided health insurance, in addition to their contemporaneous
redistributive function, serve to reduce the risks associated with the free market by reducing the
extent of lost income as a function of adverse labor market conditions, the devaluation of skills
through technological advance, or structural adjustments in the composition of the labor market
due, for example, to international trade (e.g. Ehrlich and Maestas 2010; Iversen 2005; Iversen
and Soskice 2001; Rhem 2009; Scheve and Slaughter 2004). As Moene and Wallerstein (2001)
explain, “In the second view, the essence of welfare policy is the public provision of insurance,
and self-interested voters support welfare policy to obtain protection against risks that private
insurance markets fail to cover” (p. 860).
If both “pure” redistribution and social welfare programs serve social insurance
functions, then a second potential relationship between needs for certainty and security may be
considered. Specifically, citizens most averse to uncertainty and insecurity generally should
prefer programs designed to reduce such uncertainty and insecurity in the economic domain
specifically. As Iversen (2005) states, “workers…do not know with certainty how they will fare
in terms of future employment and income. In this situation, risk-averse people will demand
insurance against loss of employment and income” (emphasis mine). While Iversen (2005)
considers risk aversion as a general disposition of the public, in the present context we can
consider variability as a function the psychological needs for security and certainty highlighted
by recent work (see also Ehrlich and Maestas 2010; Kam and Simas 2010). Put simply, citizens
high in these needs should prefer liberal intervention in the economy to increase the security,
stability, and predictability of their income and place in the labor market.
It is my sense that both perspectives on the relationship of these dispositions to economic
liberalism and conservatism are prima facie theoretically reasonable. The present paper argues
that they are both operative in the mass public, but vary in importance across levels of political
engagement. Consider that the hypothesis connecting needs for certainty and security to
economic conservatism relies on the assumption that citizens represent the economic domain in
abstract, symbolic terms. There is, however, little reason to expect this to be the case for all
citizens. Indeed, as political scientists have long noted (e.g. Converse 1964), and political
psychologists have recently explored (Federico et al. 2011; Jost, Federico and Napier 2009), such
abstract reasoning is not pervasive, and is likely isolated to those who are most “sophisticated”
politically. Citizens who are engaged with the elite political space are more likely to listen to and
accept discourse which attempts to draw out the symbolic aspects of political issues. Highly
attentive citizens should thus display greater levels of abstraction when it comes to “hard” issues
like those of the economic domain (Carmines and Stimson 1980; Pollock, Lilie and Vittes 1993).
On the other hand, citizens who pay little attention to politics should be more likely to
represent such issues in terms of considerations which are available to them, namely personal
ones. Converse (1964; p. 213) made this point nicely:
At the same time, moving from top to bottom of this information dimension, the character of
objects that are central in a belief system undergoes systematic change. These objects shift from
the remote, generic, and abstract to the increasingly simple, concrete, or “close to home.” Where
potential political objects are concerned, this progression tends to be from abstract, ‘ideological’
principles to the more obviously recognizable social groupings or charismatic leaders and finally
to such objects of immediate experience as family, job, and immediate associates.
The present paper posits that more “concrete” and “close to home” views of liberal economic
policy will be more likely to consider such policy in terms of its effects on personal economic
well-being, and thus with respect to its social protective functions. Thus, rather than limiting the
relationship between dispositions and economic preferences as argued by recent work (e.g.
Federico et al. 2011; Jost, Federico and Napier 2009), political engagement should change the
qualitative nature of the relationship between needs and preferences; more specifically, by
changing the way in which the domain is understood (in terms of institutional stability or social
protection), and thus the direction of the relationship between certainty and security needs and
economic conservatism. The key hypotheses of the present paper can thus be stated as follows:
(1) For politically unengaged citizens, increases in needs for certainty and security will be
associated with increases in economic liberalism.
(2) For politically engaged citizens, increases in needs for certainty and security will be
associated with increases in economic conservatism.
One of the most interesting aspects of these two hypotheses is that they offer a potential
explanation for the weak to null findings of recent research which has explicitly examined the
economic domain in the context of dispositional influences. Specifically, as illustrated in Figure
1, a failure to account for heterogeneity across levels of political engagement, to the extent that
the effects are of opposite sign for the engaged and unengaged, would result in marginal effects
of zero as a result of such aggregation. This hypothesis thus offers an elegant solution to the
problem of null and minimal findings in recent work within the economic domain: needs for
certainty and security are actually highly influential factors influencing the economic preferences
of citizens, but in unique ways conditional on how these issues are understood, and thus on
political engagement. I test these hypotheses across three data sets in the following pages. I then
turn to an examination of the broader implications of these findings for American politics. The
final section considers the variation in empirical findings across recent studies within the context
of the present theory, and suggests directions for future research.
DATA AND METHODS
Data
To test my hypotheses I rely on data from the 2000, 2004 and 2008 American National
Elections Studies. In 2000, the ANES utilized both face-to-face and telephone interviews, the
former via a national probability sample, and the latter via random digit dialing. In 2004 and
2008, all interviews were conducted face-to-face. The 2008 ANES included an oversample of
Black and Latino Americans. Information regarding all aspects of these studies can be accessed
at the ANES public website. In the analyses to follow I rely only on self-identified White
Americans. Recent work suggests that personality may operate differently in the economic
domain across racial and ethnic groups (Gerber et al. 2010), and thus analyses should examine
such groups separately rather than make general inferences that do not hold across all
subsamples. In 2000 and 2004 the sample sizes for minority group members are simply too
small to accurately assess these hypotheses. In 2008, however, the ANES conducted an
oversample of Black and Latino Americans. I estimated identical models for these subsamples
and found no evidence of an influence of personality on preferences. This is consistent with the
findings of Gerber et al. (2010).
I nonetheless estimated the full sample models for both 2000 and 2004, and these are
included in Appendix A. The pattern of results is similar but the effect sizes are weaker. This is
consistent with a “dilution” of the influence of personality upon including non-White citizens.
For 2008, I include in this appendix the separate models for the Black and Latino subsamples.
Needless to say, all inferences throughout the remainder of the paper apply to White Americans
only. At this time, little theoretical or empirical work has sought to address racial and ethnic
heterogeneity in the influence of personality on political attitudes and behavior. This is clearly
an important topic for future research, but beyond the theoretical scope of the present paper.
Dependent Variables
I test the two hypotheses of the present study utilizing two distinct measures of economic
conservatism, one more ideological in content, and one more policy-oriented.
Preferences for Limited Government. I draw on three items in each survey to measure
general orientations toward government intervention in economic matters. Each survey
contained an identical set of items. For each, respondents were asked to choose from two
statements. The first item contrasted the statement, “The main reason government has become
bigger over the years is because it has gotten involved in things that people should do for
themselves,” with, “government has become bigger because the problems we face have become
bigger.” The second contrasted, “We need a strong government to handle today’s complex
economic problems,” with, “The free market can handle these problems without government
being involved.” Finally, the third contrasted, “The less government, the better,” with, “there are
more things the government should be doing.” I recoded these items so that each takes on a value
of one for the limited government response. I then created a summative index with higher values
indicating greater preferences for limited government.
Economic Policy Conservatism. I measure economic policy conservatism in each study
as a scale constructed from responses to specific policy items. In 2000 and 2004, I rely on three
items: preferences for more or less government spending and services, preferences for
government guaranteed jobs and income, and preferences over the government provision of
health insurance. In 2000, the respondents were randomly assigned to receive one of two item
formats. The first utilized the typical ANES seven-point scale format, while the second utilized
a branching format resulting in a five-point scale. The three middle options of the former were
combined to form a five-point scale, and the two samples were combined to maximize sample
size. In 2004, all respondents received the seven-point branching format. In 2008, I rely on only
two items: preferences over government spending and services and over the government
provision of health insurance. In this study only half of the sample received the government
guaranteed jobs and income item. In addition, half of the sample was randomly assigned to
receive new versions of the health insurance and spending items. Both of the new versions of
these items utilized slightly different wordings and a branching format. I again combine the two
subsamples to maximize sample size. The exact wordings for all items from all three studies are
contained in Appendix B. The reliability of the scales in 2000 and 2004 were .59 and .71,
respectively. In 2008, the correlation between the spending and insurance items was .42. Each
scale was recoded to range from zero to one prior to analysis.
Independent Variables
Needs for Certainty and Security. I operationalize the key psychological dispositions
highlighted by recent work in political psychology in two ways. In all three studies I rely on four
items utilized by recent work to measure the construct of “authoritarianism.” The scale was
constructed equivalently in each study. Each item asked respondents to choose which of two
values is more important for a child to possess. The pairwise comparisons include: “Independent
or Respect for Elders,” “Curiosity or Good Manners,” “Obedience or Self-Reliance,” and
“Considerate or Well-Behaved.” Some respondents volunteer “Both,” and these are coded at an
intermediate point. The four items were then summed to create a single measure. Much work in
political psychology suggests that authoritarianism is derivative of more basic needs for certainty
and security (see Jost et al. 2003 for much evidence in this regard; see also Duckitt 2001;
Feldman 2003; Lavine et al. 1999; Lavine, Lodge and Freitas 2005; Stenner 2005). More
recently, Hetherington and Weiler (2009) demonstrated the convergent validity of this
operationalization. Moreover, the items themselves have face validity with respect to their ability
to distinguish respondents on the basis of epistemic and existential needs. Each item considers
potential solutions to dealing with a dangerous or uncertain world, more specifically, by adhering
to established norms, rules and institutions, and respecting legitimate authorities. As Duckitt
(2001) argues, “A view of the world as dangerous, unpredictable, and threatening…would
activate the motivational goal of social control and security. This motivational goal would be
expressed in the collectivist sociocultural values of conformity and traditionalism and
in…authoritarian social attitudes” (p. 50).
In 2008 I rely on a second measure for replication purposes which tap the need for
nonspecific cognitive closure (Kruglanski 1989; Kruglanski and Webster1996). The need for
closure is highly related conceptually to citizens’ relative comfort with, or aversion to,
uncertainty generally, and thus to substantial changes in their social environment. As Jost et al.
(2003) argue, “contents that promise or support epistemic stability, clarity, order, and uniformity
should be preferred by high-need-for-closure persons over contents that promise their epistemic
opposites (i.e. instability, ambiguity, chaos, and diversity)” (p. 348). One item included in the
2008 ANES was modeled after related items from the dispositional need for closure scale
(Kruglanski, Webster and Klem 1993). This item read, “Of the situations when you see two
people disagreeing with one another, in how many of them can you see how both people could
be right?” Responses options ranged from zero (“All of them”) to four (“None of them”).3
Higher values imply a greater need for closure.
Political Engagement. I operationalize political engagement with two variables: political
sophistication and partisan strength. Political sophistication has commonly been utilized to
proxy for attention to elite discourse, and the propensity of citizens to seek out and assimilate
information from the elite space (e.g. Berinsky 2007; Zaller 1992). Partisan strength taps the
Related items from Kruglanski et al.’s (1993) need for nonspecific cognitive closure scale
include: “When considering most conflict situations, I can usually see how both sides could be
right,” and, “In most social conflicts, I can easily see which side is right and which is wrong.”
3
extent to which citizens have a developed political identity, and thus the extent to which they are
personally invested in elite level partisan politics. In this sense, partisan strength can be seen in
terms of the extent to which a given citizen feels a part of their political “team” (Green,
Palmquist and Schickler 2002), and should thus be related to political information seeking, and
the tendency to assimilate rhetoric from in-partisans. Thus, both political sophistication and
partisan strength are intended to tap citizens’ exposure to elite rhetoric.4
In 2000 and 2004, political sophistication was operationalized with six knowledge items.
At the time of this writing, the knowledge items for 2008 had yet to be released by the ANES. I
thus rely on two items: the citizen’s self-reported interest in politics, and the interviewer’s
subjective rating of the respondent’s knowledge of politics. These two items were highly
correlated (r=.42) and were combined to form a single scale. To check the validity of this
operationalization, I generated an identical variable in the 2004 ANES, and correlated this scale
with the political sophistication operationalization constructed from the six knowledge items in
that study. The correlation was large, indicating that the former provides a reasonable proxy for
this indicator (r=.57). Partisan strength was operationalized by folding the traditional ANES
seven-point partisan identification measure about its midpoint, resulting in a four-point scale
ranging from “pure” independent to strong partisan. The exact items used to construct the
sophistication scales can be found in Appendix B.
Controls. I additionally control for several variables which are exogenous to needs for
certainty and security, but which may be correlated with both the latter and the dependent
With respect to partisan strength, the “sports analogy” is a good one. Someone with a strong
attachment to a sports team (i.e. a highly developed identity) is more likely to seek out
information about their team, and about sports generally, than someone with little attachment to,
and thus investment in, a team.
4
variable. These include age, gender (1=male), education (six categories, entered as dummy
variables to allow for non-linearity), income, and employment status (1=unemployed).
All variables were recoded to range from zero to one prior to analysis.
ANALYSIS AND RESULTS
Results for Preferences over Limited Government
I first consider the results for preferences over limited government for all three years, and
then turn to the estimates for economic policy conservatism. In each year, I regressed
preferences for limited government on needs for certainty and security (hereafter NCS), the two
measures of political engagement, the interactions of these with NCS, and all controls. As
limited government is ordinal, I estimate the model with maximum likelihood using an ordered
probit link function. The results for these first analyses are shown in Table 1. As the dependent
variable is coded in the conservative direction, hypotheses (1) and (2) imply negative coefficients
on the NCS constituent terms (i.e. NCS increases support for a government role when
engagement is low) and large and positive coefficients for the NCS interactions with political
engagement (indicating a reversal at high levels of engagement).
Across the three models, I find strong support for both hypotheses. In all three cases the
constituent term on NCS is substantively large, in the expected negative direction, and
statistically significant, implying an increase in support for government intervention across NCS
when political engagement is at a minimum. Moreover, and in support of hypothesis (2), the
interaction term of NCS and political sophistication is positive, large and statistically significant
in all three cases. While all three coefficients for the interaction of NCS with partisan strength
are in the expected positive direction, none attain conventional levels of statistical significance.
Knowledge thus seems to be the better indicator of exposure to, and assimilation of, symbolic
elite discourse and the resulting abstract representation of the economic domain which drives the
reversal.
Ordered probit coefficients are not easily interpretable. I thus convert these estimates
into predicted probabilities of support for limited government, across levels of NCS, for low and
high levels of political engagement. I graph these estimates for each of the three models in
Panels A-C of Figure 1. The points represent the predicted probability of giving two or three
conservative responses to the three limited government items. Low engagement corresponds
with an independent citizen at the 5th percentile of sophistication (zero correct responses for 2000
and 2004), while high engagement represents a strong partisan at the 95th percentile of
sophistication (five correct responses in 2000 and six in 2004). These estimates can thus be
interpreted as the effective bounds on the influence of NCS in terms of absolute magnitude. All
controls were held at their central tendencies.
Looking first at the estimates for the 2000 ANES, a change from low to high on NCS for
unengaged citizens corresponds with an expected decrease in the probability of giving two
limited government responses of about .29 (95% CI = [-.53, -.04]), while for engaged citizens
this same change in NCS corresponds with an increase in the probability of .44 (95% CI = [.24,
.62]). These are both substantial effects with respect to magnitude, but as expected by the
theorizing above, are in exactly opposite directions substantively. The results for 2004 are
highly similar. For low engaged citizens the marginal effect of NCS is -.34 (95% CI = [-.60, .07]), while for the highly engaged it is .30 (95% CI = [.09, .51]). Finally, turning to the 2008
study, a change from low to high on NCS for low engaged citizens corresponds with a decrease
in limited government of 56 percentage points (95% CI = [-.74, -.33]), while the same change for
the highly engaged corresponds with an increase of 41 percentage points (95% CI = [.24, 57]).
Overall, these results strongly support theoretical expectations. There is, however, an
interesting nuance to the findings which is worth mentioning, and which reinforces the basic
theoretical mechanism thought to underlie the patterns observed. Specifically, consider the
asymmetry in the magnitude of the absolute marginal effect of authoritarianism across
engagement across the different years. While only so much can be gleaned from this small
number of (yearly) observations, the patterns across the years with respect to this asymmetry
make good sense in light of theory. In 2000, a culmination of the late-1990s economic boom, the
marginal impact of authoritarianism on liberalism for the unengaged is 15 percentage points
lower (in absolute terms) than its impact on conservatism for the engaged. In contrast, in 2008 in
the midst of the recent economic crisis, this pattern reverses, with the absolute effect for low
engaged citizens 15 percentage points larger than that of the highly engaged. If citizens at low
levels of engagement are indeed more likely to view the economic domain through the personal
lens of economic certainty and security, then this pattern is exactly what we should expect to see,
namely, an increase in the importance of basic needs for certainty and security on preferences for
government intervention when the economy itself suggests objective reasons to be uncertain and
insecure.
Results for Economic Policy Conservatism
I turn now to the results for economic policy conservatism, which are shown in in Table
2. For each year, I estimated identical models to those for limited government, except that I
utilized ordinary least squares regression with robust standard errors to correspond with the
pseudo-continuous dependent variable. When interpreting coefficients, recall that all variables
are coded from zero to one. The patterns in Table 2 once again strongly support theoretical
expectations, and largely replicate the findings above with a distinct measure of economic
preferences. In 2004 and 2008 the constituent term for NCS is negative, large, and statistically
significant, while in 2000 it is marginally significant (p<.10). Moreover, the interaction term of
NCS and political sophistication is once again strong, positive, and significant in all three
models. As with the analyses for limited government above, despite being in the correct
direction, I find no significant moderating effect of partisan strength in 2000, but I do find a large
moderating effect in both 2004 and 2008 which is consistent with my more general expectations
regarding political engagement. Overall, looking across all six models, it appears that political
sophistication matters more than strength of political identity, although there is some evidence
that such strength matters given the consistently positive moderating influence of the variable,
and its significance in two of the models.
Once again, I facilitate interpretation of these dynamics by translating model estimates
into predicted values of the dependent variable. These are graphed in Panels A-C of Figure 2.
Looking first at 2000, the marginal effect of NCS for low engaged citizens is -.11 (95% CI = [.25, .03]), while for the highly engaged it is .17 (95% CI = [.05, .29]). Thus, NCS decreases
conservatism (albeit insignificantly) by about 11 percentage points for low engaged citizens, and
increases policy conservatism by about 17 percentage points for highly engaged citizens. The
results for 2004 are quite a bit stronger in terms of magnitude. At low levels of engagement, a
change from low to high on NCS corresponds with a decrease in policy conservatism of about 27
percentage points (95% CI = [-.42, -.10]). For the highly engaged, the effect of NCS is an
increase in conservatism of about 33 points (95% CI = [.21, .45]). Finally, in 2008 the estimates
for low and high engaged citizens were -.29 (95% CI = [-.47, -.10]) and .28 (95% CI = [.17,
.40]), respectively.
Taken as a whole, this second set of analyses provides converging evidence for
hypotheses (1) and (2). In contrast to the limited government analyses, however, the strong
asymmetry between 2000 and 2008 is not full replicated. While the marginal effect of
authoritarianism for low engaged citizens is indeed lower in 2000, the effects in 2004 and 2008
were comparable.
Replication with the Need for Nonspecific Cognitive Closure
I sought to replicate the findings above with a distinct measure of the key independent
variable; specifically, the need for nonspecific cognitive closure. I substituted this variable for
the “authoritarianism” operationalization in identical models of limited government preferences
and economic policy conservatism. These results are shown in Table 3 and Figures 4A-C. In
both cases, the results are weaker with respect to statistical significance, which is understandable
given the relatively weak operationalization of this construct. Nonetheless, the patterns show
support for theoretical expectations and hypotheses (1) and (2). In the model for limited
government preferences, the constituent term on the need for closure (NFC) is negative, although
fails to attain statistical significance (B=-.57, p<.25). The interaction term with political
sophistication is large, positive and marginally significant (B=1.32, p<.06). Thus, the basic
pattern of results is identical to the analyses above, although the estimates are rather inefficient.
In addition, the interaction of NFC with partisan strength in the limited government model is in
the incorrect direction and insignificant. I thus plot the predicted probabilities of giving two
limited government responses in a way identical to that above, as well as holding constant
partisan strength at its median value. These are panels A and B in Figure 4.
Turning to the results for economic policy conservatism, I find stronger support for
theoretical expectations, and a cleaner replication. The constituent term for NFC is substantively
large, negative, and statistically significant (B=-.25, p<.05). The interaction term with political
sophistication is large, positive, and marginally significant (B=.32, p<.06). The interaction term
with partisan strength is much smaller and not significant (B=.07, n.s.). The predicted values of
economic conservatism, across values of NFC, for low and high engagement are plotted in panel
C of Figure 4. The characteristic pattern is again evident. Thus, while weaker overall, I have
replicated the basic pattern of findings expected utilizing an entirely distinct measure of needs
for certainty.
BROADER IMPLICATIONS
These dynamics have broader implications for American politics, more specifically, for
the distribution of preferences for social welfare among low SES citizens, and thus potentially
for the overall level of social welfare provision in the United States. Needs for certainty and
security (NCS) are associated with both education and income among White Americans, and, as
demonstrated above, increases in political engagement “flip” the effect of this disposition with
respect to its association with economic conservatism. These two points interact to suggest that
citizens at the lower ends of the SES distribution may, perhaps surprisingly from the perspective
of pure individual self-interest, become more conservative as they become more politically
engaged and sophisticated.
I examine this dynamic empirically in two distinct ways using the economic policy
preferences model from the 2008 ANES.5 First, I generated expected values of economic policy
conservatism for citizens with a high school degree (but no additional education), who are at the
25th percentile of income, at the 95th percentile of NCS, across levels of political engagement. I
hold all other variables at their central tendencies. In effect, these estimates examine what
5
The 2008 study is particularly useful for purposes of demonstration as it represents a period
where we would most expect low SES citizens to support government provided social welfare.
happens to economic conservatism for high NCS, but low SES, citizens as they “become”
politically engaged. I plot these estimates in Panel A of Figure 5. The first bar shows the
expected value of conservatism for such citizens when they are independent and at the 5th
percentile of sophistication, and the second shows the expected value for strong partisans at the
95th percentile of sophistication. At low levels of engagement, these citizens are indeed quite
liberal, all else equal, with an expected value of the dependent variable of .25. In contrast, at
high levels of engagement, their expected value of conservatism increases to .60, a substantial
change, and one which places such citizens above the median value for all White Americans in
2008 (.50). In essence, as these citizens become more engaged politically, they go from below
average economic conservatism to above average conservatism.
A second, and perhaps more powerful demonstration of this dynamic relies not on
simulations which hold NCS at researcher-chosen levels, but rather on the existing association
between SES and NCS in the population. More specifically, I generated predicted values of
economic policy conservatism for each respondent using the estimated coefficients from the
model and the respondent’s own characteristics. I then use these predicted values to examine the
distribution of economic policy conservatism for different groups of citizens. As the data
become relatively sparse as they are broken down into relevant groups, I consider income and
education separately as indicators of SES. In Panel B of Figure 5, I plot these distributions for
two groups of citizens: independent or leaning partisans with below the median income who are
in the bottom 20% of sophistication, and weak or strong partisans with below the median income
who are in the upper 27% of sophistication. As seen, the distribution of predicted values
becomes much more conservative as we move from the first, relatively unengaged group to the
second, relatively engaged group. The median and mean predicted values for the former are .34
and .33 (SD=.10), while for the latter they are .49 and .49 (SD=.10). Again, it is important to
note that this change is a function of the characteristics of the respondents themselves in
interaction with the specified model. Panel C of Figure 5 is similar, but substitutes education for
income as a measure of SES. Specifically, I consider the same change for respondents with
some college (but no post-high school degrees) or less (i.e. the first three categories of
education). Once again, the change in the distribution is substantively meaningful, moving from
a median and mean of .36 and .35 (SD=.10) for the low engaged group to .52 and .55 (SD=.10)
for the high engaged group. The slightly greater change for the education operationalization of
SES in relation to the income operationalization is consistent with the greater correlation of NCS
to education.
Overall, these two examinations suggest that increasing attention to, and engagement
with, elite level politics among White citizens at the lower end of the SES distribution results in
substantially increased economic conservatism. This dynamic is particularly interesting given
recent debates regarding the importance of the so-called “culture war” to American politics.
While popular commentaries, such as Thomas Frank’s (2004) What’s the Matter with Kansas,
have argued that the rise of cultural conflict has displaced the economic domain, driving lowincome White citizens into the Republican camp, political scientists examining the issue
empirically have found that economic ideology and preferences remain the dominant predictors
of voting behavior and partisanship in the American mass public (Ansolabehere, Rodden and
Snyder 2006; Bartels 2006; 2008; Gelman 2008). As I have demonstrated above, however,
economic displacement by cultural conflict is not necessary to generate increased opposition to
the welfare state, because cultural conflict, in the guide of differential psychological profiles
related to epistemic and existential needs, in interaction with elite-level politics, shapes
economic ideology and preferences. Most importantly, there is a substantial increase in
conservatism as low SES citizens increase their attentiveness to elite politics. This suggests not
only that “cultural conflict” shapes economic orientations, but that low SES citizens who are
most likely to participate in politics (i.e. the politically engaged) are the most conservative,
relatively speaking, within the economic domain. Thus, culture war theorists may be right with
respect to the consequences of cultural conflict they highlight, but for the wrong reasons.
GENERAL CONCLUSIONS
The present paper has reconsidered both theoretically and empirically the relationship
between needs for certainty and security and economic conservatism. While past work has
theorized that such needs should be associated with opposition to government redistribution and
social welfare, empirical support for this expectation has been mixed at best. The present paper
seeks a theoretical model of the translation of such needs into economic preferences which can
account for the weak findings of previous work. More specifically, I have argued that there are at
least two reasonable paths by which needs for certainty and security may translate into economic
preferences. A first path considers social welfare and redistribution as social protection, and thus
as security and certainty enhancing. The second, consistent with recent theorizing, considers
economic intervention to be threatening to the institutional status quo, in terms of both
traditional values and extant socioeconomic hierarchies. I additionally argued that these two
representations make exactly opposite predictions regarding the relationship between certainty
and security needs and preferences, with the former increasing liberalism and the latter
increasing conservatism. Moreover, and consistent with the notion that ideological abstractions
are not a default representation of political issues, I argued that political engagement moderates
citizens’ representations of the economic domain, and thus the translation of needs into
preferences.
This hypothesis accounts for the mixed findings of past research. First, on the basis of
these considerations, we would expect null, unconditional associations between needs and
economic preferences in samples of the citizenry where the distribution of political engagement
is largely symmetric (because the signs of the effects of needs on preferences cancel across these
subgroups). Much of the work to date has found minimal associations, consistent with this
expectation. Second, we should expect that samples with disproportionately sophisticated
citizens should be more likely to show a relationship between needs and economic preferences.
One of the few studies to find evidence consistent with the unconditional hypothesis is Gerber et
al. (2010), who found that openness to experience and conscientiousness predicted economic
liberalism and conservatism, respectively, consistent with the broader theorizing of Jost et al.
(2003). The sample of this study consisted of registered voters, and thus citizens more politically
engaged than the average.
Third, we should expect that indicators of economic conservatism which are more
straightforwardly symbolic in content should be more likely to show unconditional relationships
than indicators with the potential for more concrete representations. In Jost et al.’s (2003) metaanalysis, they find evidence for unconditional relationships of needs with Social Dominance
Orientation (Sidanius and Pratto 2000), a highly symbolic indicator of the economic domain, but
more mixed evidence when preferences are operationalized in a more concrete fashion. In a
similar vein, Federico et al. (2011) find that political sophisticates show strong relationships of
authoritarianism to egalitarianism, an indicator of attitudes toward equality per se. The present
study has considered two operationalizations which are much more concrete, namely, beliefs
about the necessity of government intervention in economic matters, and specific preferences
over social welfare policies.
Fourth, Gerber et al. (2010) find that “neuroticism,” characterized by agreement that one
is typically “anxious” and “easily upset,” predicts economic liberalism. In past theorizing, such
sensitivity is typically theorized to be associated with conservatism (e.g. Thorisdottir and Jost
2011). However, if we control for the “Big Five” traits which are related to economic liberalism
and conservatism at a more symbolic level (i.e. openness to experience and conscientiousness,
respectively), then it is perhaps no surprise that we find significant associations of a trait which,
within the present theory, should be more straightforwardly relevant to the concrete aspects of
the economic domain.
More generally, the present paper suggests the importance of continuing to move beyond
the “left-right” continuum in studies of the psychological roots of political ideology, and, in
particular, to give greater attention to the economic domain as unique, with its own dynamics
which do not necessarily map cleanly onto expectations generates within other ideological
domains. It is my sense that we have only begun to unpack the nature of psychological constraint
within the economic domain.
Finally, the present paper contributes to the ongoing debates regarding the importance of
cultural conflict to politics in the United States. In the final empirical section of this paper, I
demonstrated that the present model predicts significant increases in conservatism among White
Americans as a function of increases in political engagement. This arises as a dual function of
the negative association between socioeconomic status and needs for certainty and security, and
the “flipping” of the relationship between needs and economic conservatism as engagement
increases. As these needs constitute the foundation for conflict over sociocultural values (see,
e.g., Hetherington and Weiler 2009), these results suggest that political scientists should
reconsider the extent to which the economic domain, despite its strong associations with voting
and partisanship, really lies at the core of American politics. If such preferences are endogenous
to more basic psychological dynamics related to culture, simple regressions showing the
importance of economic preferences to the vote miss the point. These preferences cannot be
taken as givens, and any interpretation of the meaning of such associations must be made in the
context of a broader model which considers from where these preferences arise in the first place.
APPENDIX A. SUPPLEMENTAL RESULTS
Table A1. Regressions of Preferences for Limited Government on Needs for Certainty and
Security (NCS) and Controls for Full Samples (2000 and 2004) and Blacks and Latinos
(2008 ANES)
______________________________________________________________________________
Variable
2000 ANES
2004 ANES
Blacks
Latinos
______________________________________________________________________________
NCS
-.78 (.30)*
-.86 (.35)*
-.92 (.79)
-.18 (.70)
NCS X Soph
1.48 (.46)*
1.06 (.46)*
.26 (1.08)
-1.09 (1.06)
NCS X Strength
.60 (.41)
.35 (.39)
.69 (.80)
1.41 (.75)^
Sophistication
Partisan Strength
Age
Male
HS Degree
Some College
Associate Degree
BA Degree
Professional Degree
Income
Unemployed
-.36 (.32)
-.58 (.29)*
.19 (.15)
.31 (.07)*
.17 (.12)
.31 (.13)*
.16 (.15)
.24 (.14)^
.10 (.16)
.98 (.25)*
-.14 (.16)
.30 (.32)
-.41 (.26)
.52 (.16)*
.16 (.08)*
-.07 (.15)
.15 (.16)
-.23 (.18)
.12 (.17)
-.33 (.19)^
.42 (.16)*
-.14 (.19)
-.22 (.88)
-.90 (.66)
.53 (.27)*
.34 (.11)*
.23 (.17)
.21 (.19)
.05 (.23)
.17 (.25)
-.59 (.41)
-.21 (.26)
-.26 (.18)
.86 (.83)
-1.16 (.57)*
.52 (.28)^
.38 (.12)*
.36 (.17)*
.32 (.19)^
.67 (.25)*
.67 (.23)*
.50 (.28)
.27 (.29)
-.19 (.21)
Cut 1
Cut 2
Cut 3
-.13 (.24)
.42 (.24)
.93 (.24)
.03 (.29)
.66 (.29)
1.22 (.29)
-.37 (.67)
.76 (.67)
1.60 (.68)
.80 (.58)
1.73 (.58)
2.53 (.59)
Pseudo R^2
.04
.05
.03
.05
N
1415
971
478
417
______________________________________________________________________________
Notes: Entries are maximum likelihood ordered probit estimates and standard errors
(in parentheses). *p<.05; ^p<.10.
Table A2. Regressions of Economic Policy Conservatism on Needs for Certainty and
Security (NCS) and Controls for Full Samples (2000 and 2004) and Blacks and Latinos
(2008 ANES)
______________________________________________________________________________
Variable
2000 ANES
2004 ANES
Blacks
Latinos
______________________________________________________________________________
NCS
-.08 (.06)
-.15 (.07)*
.21 (.17)
-.03 (.17)
NCS X Soph
.22 (.10)*
.22 (.09)*
-.16 (.23)
.09 (.25)
NCS X Strength
.08 (.09)
.15 (.08)^
-.25 (.16)^
.03 (.18)
Sophistication
Partisan Strength
Age
Male
HS Degree
Some College
Associate Degree
BA Degree
Professional Degree
Income
Unemployed
Constant
-.03 (.07)
-.06 (.06)
.06 (.03)*
.01 (.01)
.05 (.03)^
.09 (.03)*
.08 (.03)*
.07 (.03)*
.07 (.03)*
.20 (.05)*
-.08 (.03)*
.40 (.05)*
-.02 (.06)
-.07 (.05)
.10 (.03)*
.02 (.02)
.04 (.03)
.06 (.03)^
.02 (.04)
.05 (.03)
.00 (.04)
.18 (.03)*
-.08 (.04)*
.29 (.06)*
.10 (.18)
.13 (.13)
.12 (.06)*
.03 (.02)
.03 (.04)
-.01 (.04)
.01 (.05)
.04 (.06)
.00 (.07)
-.04 (.06)
-.04 (.03)
.15 (.14)
-.05 (.20)
-.08 (.15)
-.02 (.06)
.01 (.03)
.07 (.03)^
.10 (.04)*
.09 (.06)
.10 (.06)^
.06 (.07)
.16 (.06)*
.02 (.05)
.23 (.14)
R^2
.07
.12
.04
.06
N
1468
1003
476
416
______________________________________________________________________________
Notes: Entries are ordinary least squares estimates and robust standard errors (in parentheses).
*p<.05; ^p<.10.
APPENDIX B. QUESTION WORDINGS
Government Spending and Services
2000 (Face-to-Face; reduced to 5 points), 2004 and 2008 (Old)
Some people think the government should provide fewer services even in areas such as health
and education in order to reduce spending. Suppose these people are at one end of a scale, at
point 1. Other people feel it is important for the government to provide many more services even
if it means an increase in spending. Suppose these people are at the other end, at point 7. And, of
course, some other people have opinions somewhere in between, at points 2,3,4,5 or 6.
2000 (Phone)
Some people think the government should provide fewer services even in areas such as health
and education in order to reduce spending. Other people feel it is important for the government
to provide many more services even if it means an increase in spending. [Respondents
volunteering “stay the same” are coded at the midpoint] [If “Fewer” or “More” services, “should
the government reduce/increase services a great deal, or only some?”].
2008 (New)
Do you think the government should provide MORE services than it does now, FEWER services
than it does now, or ABOUT THE SAME NUMBER of services as it does now? [If “more” or
“fewer,” “do you think the government should provide A LOT more/fewer, SOMEWHAT
more/fewer, or SLIGHTLY more/fewer services than it does now?”]
Government Health Insurance
2000 (Face-to-Face; reduced to 5 points), 2004 and 2008 (Old)
Some people feel there should be a government insurance plan which would cover all medical
and hospital expenses for everyone. Suppose these people are at one end of a scale, at point 1.
Others feel that all medical expenses should be paid by individuals through private insurance
plans like Blue Cross or other company paid plans. Suppose these people are at the other end, at
point 7. And, of course, some other people have opinions somewhere in between, at points 2, 3,
4, 5, or 6.
2000 (Phone)
Some people feel there should be a government insurance plan which would cover all medical
and hospital expenses for everyone. Others feel that all medical expenses should be paid by
individuals through private insurance plans like Blue Cross or other company paid plans. Which
is closer to the way you feel or haven't you thought much about this? [Respondents could
volunteer “Should stay the same”] [If R chooses one of options, “Do you feel strongly or not
strongly?”]
2008 (New)
Do you FAVOR, OPPOSE, or NEITHER FAVOR NOR OPPOSE the U.S. government paying
for all necessary medical care for all Americans? [If “Favor” or “Oppose,” do you favor that A
GREAT DEAL, MODERATELY, or A LITTLE?”]
Government Jobs and Income
2000 (Face-to-Face; reduced to 5 points) and 2004
Some people feel the government in Washington should see to it that every person has a job and
a good standard of living. Suppose these people are at one end of a scale, at point 1. Others think
the government should just let each person get ahead on their own. Suppose these people are at
the other end, at point 7. And, of course, some other people have opinions somewhere in
between, at points 2,3,4,5, or 6.
2000 (Phone)
Some people feel the government in Washington should see to it that every person has a job and
a good standard of living. Others think the government should just let each person get ahead on
their own. Which is closer to the way you feel or haven't you thought much about this?
[Respondents could volunteer “Should stay the same”] [If R chooses one of options, “Do you
feel strongly or not strongly?”]
Political Sophistication
2000 and 2004
Do you happen to know which party had the most members in the House of Representatives in
Washington BEFORE the election (this/last) month?
Do you happen to know which party had the most members in the U.S. Senate BEFORE the
election (this/last) month?
Now we have a set of questions concerning various public figures. We want to see how much
information about them gets out to the public from television, newspapers and the like.
2000 Only
2004 Only
JANET RENO
TRENT LOTT
DENNIS HASTERT
DICK CHENEY
Both 2000 and 2004
WILLIAM REHNQUIST
TONY BLAIR
2008
For political interest, respondents were randomly assigned to one of two operationalizations:
A. Some people don't pay much attention to political campaigns. How about you? Would you
say that you have been VERY MUCH interested, SOMEWHAT interested or NOT MUCH
interested in the political campaigns so far this year?
B. How interested are you in information about what's going on in government and politics?
EXTREMELY INTERESTED, VERY INTERESTED, MODERATELY INTERESTED,
SLIGHTLY INTERESTED, or NOT INTERESTED AT ALL?
For political knowledge, the interviewer was asked the following: “R's general level of
information about politics and public affairs seemed: Very high, Fairly high, Average, Fairly
low, or Very low.”
Both items were recoded to range from zero to one and averaged to form a single scale.
REFERENCES
Adorno, T. W., Frenkel-Brunswick E., Levison, D.J., and Sanford, R.N. 1950. The Authoritarian
Personality. New York: Harper and Row.
Altemeyer, Bob. 1988. Enemies of Freedom: Understanding Right-Wing Authoritarianism. San
Francisco: Jossey-Bass Publishers.
Ansolabehere, Stephen, Jonathan Rodden, and James M. Snyder Jr. 2006. “Purple America.”
Journal of Economic Perspectives 20 (2): 97-118.
Bartels, Larry M. 2006. “What’s the Matter with What’s the Matter with Kansas?” Quarterly
Journal of Political Science 1: 201-226.
Bartels, Larry M. 2008. Unequal Democracy: The Political Economy of the New Gilded Age.
New Jersey: Princeton University Press.
Berinsky, Adam J. 2007. “Assuming the Costs of War: Events, Elites, and American Public
Support for Military Conflict.” The Journal of Politics 69 (4): 975-997.
Carmines, Edward G., and James A. Stimson. 1980. “The Two Faces of Issue Voting.” American
Political Science Review 74 (1): 78-91.
Carney, Dana R., John T. Jost, Samuel D. Gosling, and Jeff Potter. 2008. “The Secret Lives of
Liberals and Conservatives: Personality Profiles, Interaction Styles, and the Things They
Leave Behind.” Political Psychology 29 (6): 807-840.
Carver, Charles S., and Michael F. Scheier. 1998. On the Self-Regulation of Behavior. New
York: Cambridge University Press.
Cesario, Joseph, Heidi Grant, and Tory E. Higgins. 2004. “Regulatory Fit and Persuasion:
Transfer from ‘Feeling Right.’” Journal of Personality and Social Psychology 86 (3):
388-404.
Chirumbolo, Antonio, Alessandra Areni, and Gilda Sensales. “Need for Cognitive Closure and
Politics: Voting, Political Attitudes and Attributional Style.” International Journal of
Psychology 39 (4): 245-253.
Converse, Philip E. 1964. “The Nature of Belief Systems in Mass Publics.” In Ideology and
Discontent, ed. David Apter. New York: Free Press, pp. 206-261.
Duckitt, John, Claire Wagner, Ilouize du Plessis, and Ingrid Birum. 2002. “The Psychological
Bases of Ideology and Prejudice: testing a Dual Process Model.” Journal of Personality
and Social Psychology 83 (1): 75-93.
Ehrlich, Sean, and Cherie Maestas. 2010. “Risk Orientation, Risk Exposure, and Policy
Opinions: The Case of Free Trade.” Political Psychology 31 (5): 657-684.
Feldman, Stanley. 2003. “Enforcing Social Conformity: A Theory of Authoritarianism.”
Political Psychology 24 (1): 41-74.
Feldman, Stanley, and Christopher D. Johnston. 2009. “Understanding Political Ideology: The
Necessity of a Multidimensional Conceptualization.” Paper presented at the Annual
Meeting of the American Political Science Association.
Feldman, Stanley, and Karen Stenner. 1997. “Perceived Threat and Authoritarianism.” Political
Psychology 18 (4): 741-770.
Fiorina, Morris P., Samuel J. Abrams, and Jeremy C. Pope. 2005. Culture War? The Myth of a
Polarized America. New York: Pearson Longman.
Frank, Thomas. 2004. What’s the Matter with Kansas? How Conservatives Won the Heart of
America. New York: Henry Holt and Company, LLC.
Gelman, Andrew. 2008. Red State, Blue State, Rich State, Poor State: Why Americans Vote the
Way They Do. Princeton: Princeton University Press.
Gerber, Alan S., Gregory A. Huber, David Doherty, Conor M. Dowling, and Shang E. Ha. 2010.
“Personality and Political Attitudes: Relationships across Issue Domains and Political
Contexts.” American Political Science Review 104 (1): 111-133.
Green, Donald P., Bradley Palmquist, and Eric Schickler. 2002. Partisan Hearts and Minds:
Political Parties and the Social Identities of Voters. New Haven: Yale University Press.
Hetherington, Marc J. 2001. “Resurgent Mass Partisanship: The Role of Elite Polarization.”
American Political Science Review 95 (3): 619-631.
Hetherington, Marc J., and Elizabeth Suhay. 2011. “Authoritarianism, Threat, and Americans’
Support for the War on Terror.” American Journal of Political Science 55 (3): 546-560.
Hetherington, Marc J., and Jonathan Weiler. 2009. Authoritarianism and Polarization in
American Politics. New York: Cambridge University Press.
Higgins, E. Tory. 1998. “Promotion and Prevention: Regulatory Focus as a Motivational
Principle.” In Advances in Experimental Social Psychology, vol. 30, ed. Mark P. Zanna.
San Diego: Academic Press.
Iversen, Torben. 2005. Capitalism, Democracy, and Welfare. New York: Cambridge University
Press.
Iversen, Torben, and David Soskice. 2001. “An Asset Theory of Social Policy Preferences.”
American Political Science Review 95 (4): 875-893.
Jost, John T., Christopher M. Federico, and Jaime L. Napier. 2009. “Political Ideology: Its
Structure, Functions, and Elective Affinities.” Annual Review of Psychology 60: 307-337.
Jost, John T., Jack Glaser, Arie W. Kruglanski, and Frank J. Sulloway. 2003a. “Political
Conservatism as Motivated Social Cognition.” Psychological Bulletin 129 (3): 339-375.
Jost, John T., and Orsolya Hunyady. 2005. “Antecedents and Consequences of System-Justifying
Ideologies.” Current Directions in Psychological Science 14 (5): 260-265.
Jost, John T., Sally Blount, Jeffrey Pfeffer, and György Hunyady. 2003b. “Fair Market Ideology:
Its Cognitive-Motivational Underpinnings.” Research in Organizational Behavior 25: 5391.
Kahneman, Daniel, and Amos Tversky. 1979. “Prospect Theory: An Analysis of Decision Under
Risk.” Econometrica 47 (2): 263-292.
Kam, Cindy D., and Elizabeth N. Simas. 2010. “Risk Orientations and Policy Frames.” Journal
of Politics 72 (2): 381-396.
Lavine, Howard, Diana Burgess, Mark Snyder, John Transue, John L. Sullivan, Beth Haney, and
Stephen H. Wagner. 1999. “Threat, Authoritarianism, and Voting: An Investigation of
Personality and Persuasion.” Personality and Social Psychology Bulletin 25 (3): 337-347.
Mayda, Anna Maria, and Dani Rodrik. 2005. “Why Are Some People (and Countries) More
Protectionist Than Others?” European Economic Review: 49: 1393-1430.
McCarty, Nolan, Keith T. Poole, and Howard Rosenthal. 2008. Polarized America: The Dance
of Ideology and Unequal Riches. Cambridge: MIT Press.
McCrae, Robert R., and Paul T. Costa, Jr. 2003. Personality in Adulthood: A Five-Factor Theory
Perspective, 2nd ed. New York: The Guilford Press.
Meltzer, Allan H., and Scott F. Richard. 1981. “A Rational Theory of the Size of Government.”
The Journal of Political Economy 89 (5): 914-927.
Mischel, Walter, and Yuichi Shoda. 1995. “A Cognitive-Affective System Theory of
Personality: Reconceptualizing Situations, Dispositions, Dynamics, and Invariance in
Personality Structure.” Psychological Review 102 (2): 246-268.
Moene, Karl Ove, and Michael Wallerstein. 2001. “Inequality, Social Insurance, and
Redistribution.” American Political Science Review 95 (4): 859-874.
Mondak, Jeffery J. 2010. Personality and the Foundations of Political Behavior. New York:
Cambridge University Press.
Pollock, Philip H. III, Stuart A. Lilie, and Elliot Vittes. 1993. “Hard Issues, Core Values and
Vertical Constraint: The Case of Nuclear Power.” British Journal of Political Science 23
(1): 29-50.
Rehm, Philipp. 2009. “Risks and Redistribution: An Individual-Level Analysis.” Comparative
Political Studies 42: 855-881.
Sargent, Greg. “Mitt Romney Blasts Obama for Being ‘Nuanced’.” The Plum Line. The
Washington Post, 22 Mar 2011. Web. 17 Jul 2011.
Scheve, Kenneth, and Matthew J. Slaughter. 2001. “What Determines Individual Trade-Policy
Preferences?” Journal of International Economics 54: 267-292.
Shoda, Yuichi. 1999. “Behavioral Expressions of a Personality System: Generation and
Perception of Behavioral Signatures.” In The Coherence of Personality, eds. Daniel
Cervone and Yuichi Shoda. New York: The Guilford Press.
Sinn, Hans-Werner. 1995. “A Theory of the Welfare State.” The Scandinavian Journal of
Economics 97 (4): 495-526.
Smith, Mark A. 2009. The Right Talk: How Conservatives Transformed the Great Society into
the Economic Society. Princeton: Princeton University Press.
Stenner, Karen. 2005. The Authoritarian Dynamic. New York: Cambridge University Press.
Thorisdottir, Hulda, John T. Jost, Ido Liviatan, and Patrick E. Shrout. 2007. “Psychological
Needs and Values Underlying Left-Right Political Orientation: Cross-National Evidence
From Eastern and Western Europe.” Public Opinion Quarterly 71 (2): 175-203.
Van Hiel, Alain, Mario Pandelaere, and Bart Duriez. 2004. “The Impact of Need for Closure on
Conservative Beliefs and Racism: Differential Mediation by Authoritarian Submission
and Authoritarian Dominance.” Personality and Social Psychology Bulletin 30: 824-837.
Van Hiel, Alain, and Ivan Mervielde. 2004. “Openness to Experience and Boundaries in the
Mind: Relationships with Cultural and Economic Conservative Beliefs.” Journal of
Personality 72 (4): 659-686.
Varian, Hal R. 1980. “Redistributive Taxation as Social Insurance.” Journal of Public
Economics 14 (1): 49-68.
Zaller, John R. 1992. The Nature and Origins of Mass Opinion. New York: Cambridge
University Press.
Figure 1. A Theoretical Relationship between Needs for Certainty and Security and
Economic Conservatism
Low Political Engagement
High Political Engagement
Average Marginal Effect
Figure 2. The Influence of Needs for Certainty and Security on Preferences for Limited Government
Panel A. 2000 ANES
Low Engagement
High Engagement
High Engagement
.6
.5
.4
.3
.1
.1
.2
.2
.3
.4
.5
.6
Pr(Restrict Government)
.7
.7
.8
.8
Low Engagement
Panel B. 2004 ANES
0
.2
.4
.6
.8
1
0
Needs for Certainty and Security
.2
.4
.6
.8
1
Needs for Certainty and Security
Panel C. 2008 ANES
Low Engagement
High Engagement
.1
.2
.3
.4
.5
.6
.7
.8
Notes: Plots represent predicted probabilities of
giving two or three (out of three) limited
government responses. Estimates were generated
holding all controls at their central tendencies.
“Low Engagement” represents independent
citizens at the 5th percentile of political
sophistication, while “High Engagement”
represents strong partisans at the 95th percentile of
political sophistication.
0
.2
.4
.6
Needs for Certainty and Security
.8
1
Figure 3. The Influence of Needs for Certainty and Security on Economic Policy Conservatism
Panel A. 2000 ANES
Low Engagement
High Engagement
High Engagement
.7
.6
.5
.4
.3
.2
.2
.3
.4
.5
.6
.7
Predicted Economic Conservatism
.8
.8
Low Engagement
Panel B. 2004 ANES
0
.2
.4
.6
.8
0
1
.2
.4
.6
.8
1
Needs for Certainty and Security
Needs for Certainty and Security
Panel C. 2008 ANES
Low Engagement
High Engagement
.2
.3
.4
.5
.6
.7
.8
Notes: Plots represent predicted values of
economic policy conservatism. Estimates were
generated holding all controls at their central
tendencies. “Low Engagement” represents
independent citizens at the 5th percentile of
political sophistication, while “High Engagement”
represents strong partisans at the 95th percentile of
political sophistication.
0
.2
.4
.6
Needs for Certainty and Security
.8
1
Figure 4. The Influence of Needs for Nonspecific Closure on Economic Policy Conservatism, 2008 ANES
Panel A. Limited Govt.
Low Engagement
Panel B. Limited Govt. (Strength Constant)
High Engagement
High Soph
.6
.4
.2
0
0
.2
.4
Pr(Restrict Government)
.6
.8
.8
Low Soph
0
.2
.4
.6
.8
1
0
Need for Closure
.2
.4
.6
.8
1
Need for Closure
Panel C. Economic Policy Conservatism
Notes: Plots represent predicted probabilities of
limited government and values of economic
policy conservatism. Estimates were generated
holding all controls at their central tendencies.
“Low Engagement” represents independent
citizens at the 5th percentile of political
sophistication, while “High Engagement”
represents strong partisans at the 95th percentile of
political sophistication. Panel B holds partisan
strength constant at its median value.
High Engagement
.2
.3
.4
.5
Low Engagement
0
.2
.4
.6
Need for Closure
.8
1
Figure 5. Simulations and Predicted Value Distributions of Economic Policy Conservatism for Low SES Respondents
Panel B. Predicted Values for Low Income
.7
Panel A. Simulations
High Engagement
0
0
.1
1
.2
2
.3
Density
3
.4
.5
4
.6
5
Low Engagement
.1
Low Engagement
High Engagement
.2
.3
.4
.5
Predicted Economic Policy Conservatism
.6
Panel C. Predicted Values for Low Education
High Engagement
2
1
0
Density
3
4
Low Engagement
Notes: 2008 ANES. Panel A shows the expected values
of economic policy conservatism for high
authoritarianism citizens with a high school degree at
the 25th percentile of income across levels of political
engagement. Extended lines represent 95% confidence
bounds. Panel B shows the distributions of predicted
economic conservatism for respondents with below the
median income across levels of political engagement.
Panel C shows these same distributions for citizens
with some college education or less.
0
.2
.4
.6
Predicted Economic Policy Conservatism
.8
Table 1. Regressions of Preferences for Limited Government on Needs for Certainty and
Security (NCS) and Controls
___________________________________________________________________________
Variable
2000 ANES
2004 ANES
2008 ANES
___________________________________________________________________________
NCS
-.76 (.34)*
-1.03 (.42)*
-1.94 (.41)*
NCS X Soph
1.51 (.51)*
1.22 (.55)*
2.49 (.54)*
NCS X Strength
.69 (.46)
.61 (.47)
.54 (.40)
Sophistication
Partisan Strength
Age
Male
HS Degree
Some College
Associate Degree
BA Degree
Professional Degree
Income
Unemployed
-.48 (.35)
-.49 (.32)
.08 (.17)
.39 (.07)*
.18 (.15)
.35 (.15)*
.18 (.17)
.25 (.16)
.10 (.18)
.85 (.27)*
-.14 (.20)
.24 (.36)
-.48 (.29)^
.45 (.19)*
.24 (.09)*
.00 (.19)
.14 (.20)
-.16 (.23)
.19 (.21)
-.36 (.23)
.38 (.19)*
-.16 (.24)
-.97 (.38)*
-.17 (.26)
.68 (.16)*
.27 (.07)*
.14 (.15)
.31 (.17)^
.29 (.17)^
.40 (.17)*
.37 (.20)^
.45 (.16)*
-.06 (.16)
Cut 1
Cut 2
Cut 3
-.15 (.27)
.38 (.27)
.89 (.27)
-.09 (.34)
.52 (.34)
1.09 (.35)
-.31 (.32)
.29 (.32)
.93 (.32)
Pseudo R^2
.04
.05
.05
N
1127
699
975
___________________________________________________________________________
Notes: Entries are maximum likelihood ordered probit estimates and standard errors
(in parentheses). Estimates are for self-identified Whites only. *p<.05; ^p<.10.
Table 2. Regressions of Economic Policy Conservatism on Needs for Certainty and Security
(NCS) and Controls
___________________________________________________________________________
Variable
2000 ANES
2004 ANES
2008 ANES
___________________________________________________________________________
NCS
-.11 (.07)^
-.27 (.08)*
-.34 (.10)*
NCS X Soph
.23 (.11)*
.26 (.11)*
.39 (.13)*
NCS X Strength
.09 (.09)
.34 (.10)*
.22 (.10)*
Sophistication
Partisan Strength
Age
Male
HS Degree
Some College
Associate Degree
BA Degree
Professional Degree
Income
Unemployed
Constant
-.09 (.07)
-.02 (.06)
.05 (.03)
.03 (.02)^
.05 (.03)
.11 (.03)*
.07 (.04)*
.08 (.03)*
.07 (.04)*
.19 (.05)*
-.06 (.04)
.43 (.06)*
-.06 (.07)
-.11 (.06)*
.08 (.04)*
.03 (.02)^
.02 (.04)
.03 (.04)
.04 (.05)
.04 (.04)
.00 (.05)
.21 (.04)*
-.03 (.04)
.35 (.07)*
-.16 (.09)^
-.08 (.07)
.18 (.04)*
.09 (.02)*
.05 (.04)
.08 (.04)*
.07 (.04)^
.07 (.04)^
.02 (.05)
.24 (.04)*
.03 (.04)
.29 (.08)*
R^2
.07
.15
.14
N
1182
729
1005
___________________________________________________________________________
Notes: Entries are ordinary least squares estimates and robust standard errors (in parentheses).
Estimates are for self-identified Whites only. *p<.05; ^p<.10.
Table 3. Regressions of Preferences for Limited Government (LG) and Economic Policy
Conservatism (EPC) on Needs for Nonspecific Cognitive Closure (NFC) and Controls
______________________________________________________________________________
Variable
LG
EPC
______________________________________________________________________________
NFC
-.57 (.49)
-.25 (.11)*
NFC X Soph
1.32 (.70)^
.32 (.17)^
NFC X Strength
-.46 (.52)
.07 (.13)
Sophistication
Partisan Strength
Age
Male
HS Degree
Some College
Associate Degree
BA Degree
Professional Degree
Income
Unemployed
Constant
-.07 (.38)
.37 (.27)
.69 (.16)*
.26 (.07)*
.18 (.15)
.35 (.16)*
.32 (.17)^
.43 (.17)*
.34 (.19)^
.44 (.16)*
-.07 (.16)
-
-.08 (.09)
.02 (.07)
.19 (.04)*
.10 (.02)*
.04 (.04)
.07 (.04)^
.06 (.04)
.06 (.04)
.00 (.05)
.23 (.04)*
.03 (.04)
.21 (.07)*
Cut 1
Cut 2
Cut 3
.67 (.31)
1.25 (.31)
1.89 (.32)
-
Pseudo R^2/R^2
.04
.13
N
979
1008
______________________________________________________________________________
Notes: Entries are maximum likelihood ordered probit estimates and standard errors (in
parentheses) for LG, and ordinary least squares estimates and robust standard errors for EPC.
Estimates are for self-identified Whites only. *p<.05; ^p<.10.