Better or Different? How Political Ideology Shapes Preferences for Differentiation in the Social Hierarchy

Abstract

As consumers’ political opinions become more divided and more central to their identities, it is important to understand how political ideology shapes consumers’ attempts to differentiate from others in the marketplace. Seven studies demonstrate that political ideology systematically influences consumers’ preferences for differentiation. Conservative ideology leads consumers to differentiate from others vertically in the social hierarchy through products that signal that they are better than others, and liberal ideology leads consumers to differentiate from others horizontally in the social hierarchy through products that signal that they are unique from others. This happens because conservatism endorses, and liberalism opposes, the belief that the dominance-based hierarchical social structure is a legitimate mechanism to distinguish individual qualities. The effect is robust across measured and manipulated ideology, hypothetical and real product choices, and online searches in conservative and liberal U.S. states. Manipulating consumers’ differentiation goals and perceptions of hierarchy legitimacy mitigates the effect. The findings advance existing research on political ideology, social hierarchy, and consumer divergence, and they contribute to marketing practice.

Full text

!
!
1
Better or Different?
How Political Ideology Shapes Preferences for Differentiation in the Social Hierarchy
NAILYA ORDABAYEVA
DANIEL FERNANDES
Forthcoming, Journal of Consumer Research
January 19, 2018
Nailya Ordabayeva is the corresponding author and assistant professor of marketing at the Carroll
School of Management, Boston College, 140 Commonwealth Avenue, Chestnut Hill, MA 02467,
email: nailya.ordabayeva@bc.edu, tel.: 617-552-2928. Daniel Fernandes is assistant professor of
marketing at Católica-Lisbon School of Business and Economics, Catholic University of
Portugal, Palma de Cima, 1649-023 Lisboa, Portugal, email: daniel.fernandes@ucp.pt, tel.: +351-
21-7270250. The authors acknowledge the generous financial support of the Carroll School of
Management, Boston College, and the Católica-Lisbon School of Business and Economics. The
authors thank the Editor, the Associate Editor, the three anonymous reviewers, as well as Pierre
Chandon, Anat Keinan, Gergana Nenkov, Stefano Puntoni, Stijn van Osselaer, and Min Zhao for
their valuable feedback. Supplementary materials are included in the web appendix
accompanying the online version of this article.

2
!
Abstract
As consumers’ political opinions become more divided and more central to their
identities, it is important to understand how political ideology shapes consumers’ attempts to
differentiate from others in the marketplace. Seven studies demonstrate that political ideology
systematically influences consumers’ preferences for differentiation. Conservative ideology leads
consumers to differentiate from others vertically in the social hierarchy through products that
signal that they are better than others, and liberal ideology leads consumers to differentiate from
others horizontally in the social hierarchy through products that signal that they are unique from
others. This happens because conservatism endorses, and liberalism opposes, the belief that the
dominance-based hierarchical social structure is a legitimate mechanism to distinguish individual
qualities. The effect is robust across measured and manipulated ideology, hypothetical and real
product choices, and online searches in conservative and liberal U.S. states. Manipulating
consumers’ differentiation goals and perceptions of hierarchy legitimacy mitigates the effect. The
findings advance existing research on political ideology, social hierarchy, and consumer
divergence, and they contribute to marketing practice.
Keywords: Political ideology, consumer differentiation, divergence, social hierarchy, inequality.

3
!
As the divide between conservative and liberal opinions continues to grow (Abramowitz
and Saunders 2008), consumers find themselves on either end of the political debate, and very
few remain on the sidelines (Jost 2006). The distinction between conservatism and liberalism is
the most prevalent and parsimonious classification of political ideology in the West (Altemeyer
1998; Graham, Haidt, and Nosek 2009). Upwards of 65 percent of adults self-identify as either
conservative or liberal (Gallup 2012), and political identification is a central part of individual
identity in the increasingly partisan environment (Iyengar and Westwood 2014).
Noting the increasing significance of political ideology for individuals, researchers have
begun to explore how political ideology may shape individuals’ thoughts, feelings, and behaviors
(Fraley et al. 2012; Jost et al. 2003). Recent studies suggest that ideology is important for
marketers, as conservatives and liberals express distinct preferences in the contexts of
supermarket purchases (Khan, Misra, and Singh 2013), entertainment (Roos and Shachar 2013),
recycling (Kidwell, Farmer, and Hardesty 2013), donations (Winterich, Zhang, and Mittal 2012),
and variety-seeking (Fernandes and Mandel 2014). Media pundits also stress the marketing
implications of political ideology, noting its potential link to brand preferences. For example,
conservatives reportedly drink Coke and shop at Hallmark stores, whereas liberals drink Pepsi
and shop at Trader Joe’s (Experian Simmons 2011). However, although political ideology is a
key facet of consumers’ identities, little is known about how it may affect consumers’ attempts to
differentiate their identities from others in the marketplace – a powerful driver of consumer
behavior and market outcomes (Berger and Heath 2008; Snyder and Fromkin 1991; White and
Argo 2009; White and Dahl 2006). The present research addresses this question.
We draw on a key difference between conservatism and liberalism’s attitudes toward the
social hierarchy (Jost et al. 2003). The social hierarchy refers to the differences that exist among
individuals in society based on the amount of resources that they own compared to others: those

4
!
who own more resources than others hold dominant positions, and those who own few resources
compared to others hold inferior positions (Pratto et al. 1994; Pratto and Cathey 2002). While
both conservatives and liberals agree that a dominance-based hierarchical structure exists in
society (Norton and Ariely 2000), conservatives and liberals hold opposing beliefs about the
legitimacy of this hierarchical structure (Jost et al. 2003, 2009). Conservatives endorse the
dominance-based hierarchical structure because they believe that it legitimately reflects
individual differences in effort and ability (i.e. winners deserve their dominant positions because
they work harder than losers) (Graham et al. 2009; Jost et al. 2003). In contrast, liberals oppose
the dominance-based hierarchical structure and view it as illegitimate because they believe that
everyone works hard, but some attain higher positions than others due to luck or social
connections (Jost, Federico, and Napier 2009; Kluegel and Smith 1981).
Prior work shows that conservatives’ and liberals’ beliefs about hierarchy legitimacy
translate to their distinct attitudes toward policies such as taxes and actions such as donations that
redistribute resources in the hierarchy (Crawford, Mallinas, and Furman 2015; Graham et al.
2009; Ordabayeva and Fernandes 2017). Notably, prior studies rely on correlational data to
explore the link between political ideology and hierarchy beliefs (Altemeyer 1998; Jost et al.
2009), and they shed little light on the boundary conditions of this link (Fraley et al. 2012). Past
research thus leaves three key questions open: (1) how hierarchy beliefs affect conservatives’ and
liberals’ preferences beyond redistribution; (2) whether there is a causal, not just a correlational,
link between ideology and hierarchy beliefs, and (3) what boundary conditions can mitigate this
link. Addressing these questions, we propose that conservative and liberal ideologies lead
consumers to pursue distinct forms of differentiation in the marketplace.
Consumers have a ubiquitous desire to stand out from the crowd in order to signal who
they are to others, and prior studies separately examine two forms of consumer differentiation

5
!
(Bellezza, Gino, and Keinan 2013; Berger and Heath 2008; Dommer, Swaminathan, and
Ahluwalia 2013; White and Argo 2009; White and Dahl 2006). On the one hand, several studies
report that consumers may seek to differentiate themselves vertically in the social hierarchy to
show their superior position, role, or success compared to others (Belk 1988; Ordabayeva and
Chandon 2011; Vignoles, Chryssochoou, and Breakwell 2000). To achieve the goal to
differentiate vertically in the hierarchy, consumers buy products that signal that they are better
than others (Dommer et al. 2013). On the other hand, a separate group of studies reports that
consumers may seek to differentiate themselves horizontally in the social hierarchy to express
their unique traits, personality, and values compared to others (Bellezza et al. 2013; Chan, Berger,
and Van Boven 2012). To achieve the goal to differentiate horizontally in the hierarchy,
consumers buy products that signal that they are unique from others (Chan et al. 2012; Dommer
et al. 2013; Tian, Bearden, and Hunter 2001). While prior studies examine when and why
consumers diverge, and how divergence is perceived by others (Bellezza et al. 2013; Berger and
Heath 2008; White and Dahl 2006), little is known about: (1) how consumers trade off between
different forms of differentiation, (2) what drives consumers’ distinct differentiation preferences,
and (3) when such preferences emerge or disappear. Addressing these open questions on
consumer divergence, we propose that conservative and liberal ideologies lead consumers to have
distinct preferences for vertical and horizontal differentiation in the social hierarchy.
We propose that conservatism leads consumers to pursue vertical differentiation in the
hierarchy through products that signal that they are better than others, and liberalism leads
consumers to pursue horizontal differentiation in the hierarchy through products that signal that
they are unique from others. The effect of ideology on differentiation preferences emerges
because conservative and liberal ideologies endorse distinct beliefs about the legitimacy of the

6
!
dominance-based hierarchy, but manipulating the goal to differentiate vertically or horizontally in
the hierarchy, or changing perceptions of hierarchy legitimacy, mitigates this effect.
Our research contributes to the literature linking political ideology and hierarchy beliefs:
(1) by showing that political ideology affects consumers’ preferences for differentiation, not just
their preferences for redistribution; (2) by showing that ideology has a causal effect, not just a
correlational influence, on hierarchy beliefs and preferences; and (3) by identifying interventions
that can bridge ideological differences in hierarchy beliefs and preferences. Furthermore, our
work contributes to the literature on consumer divergence: (1) by establishing political ideology
as a new driver of consumers’ preferences between different types of differentiation; (2) by
identifying hierarchy beliefs as a novel process underlying these preferences; and (3) by showing
how the divide in differentiation preferences can be bridged. Our work thereby integrates two
streams of research that previously seemed unrelated – political ideology and consumer
divergence – by establishing hierarchy beliefs as a key link between them. More broadly, our
work supports an emerging view that political ideology presents a type of cultural lens that can be
used to study consumer identity and behavior. Next, we outline the theory behind our hypotheses.
THEORETICAL FRAMEWORK
Political ideology is an important facet of individual identity for several reasons. First,
political values are among the six core values that shape personality (Vernon and Allport 1931).
Second, individuals’ ideological inclinations are detected as early as childhood, and they shape
individuals’ self-concepts through early adulthood (Fraley et al. 2012). In the result, the majority
of adults identify themselves as conservative or liberal (Gallup 2012). Finally, like other facets of
social identity, individuals’ political ideology defines a set of norms that shapes their thoughts,
feelings, and actions (Jost et al. 2003; Nail et al. 2009). Accordingly, ideological differences have

7
!
recently been reported in certain consumer behaviors including variety seeking (Fernandes and
Mandel 2014), brand choices in supermarkets (Khan et al. 2013), recycling (Kidwell et al. 2013),
donation (Winterich et al. 2012), and movie preferences (Roos and Shachar 2013).
Importantly for the present research, conservative and liberal ideologies endorse opposing
beliefs about the legitimacy of the dominance-based social hierarchy. Next, we outline these
beliefs and explain how they impact consumers’ pursuit of differentiation in the marketplace.
Ideology and Hierarchy Legitimacy Beliefs
Conservative and liberal ideologies differ in their attitudes toward inequality: conservative
ideology endorses inequality, whereas liberal ideology opposes inequality (Jost et al. 2003).
These differences emerge because conservatism and liberalism hold opposing assumptions about
the legitimacy of the dominance-based hierarchical structure that exists in society (Graham et al.
2009). Conservatism assumes that the hierarchical structure is legitimate because it reflects
individual differences in hard work and ability (Crawford et al. 2015). Liberalism assumes that
the hierarchical structure is illegitimate because everyone works hard, but some attain high ranks
due to luck or social connections (Kluegel and Smith 1981). These differences are reflected in
conservatives’ (vs. liberals’) higher social dominance orientation – defined as the perceived
legitimacy of the social hierarchy and of dominant individuals in it (Pratto et al. 1994).
Using correlational data, prior work has shown that, because of their stronger belief in the
legitimacy of the dominance-based hierarchical structure, conservatives show weaker support
than liberals for taxes and donations which redistribute wealth in society (Crawford et al. 2015;
Ordabayeva and Fernandes 2017). Expanding beyond correlations and redistributive preferences,
we propose that opposing hierarchy legitimacy beliefs lead conservative and liberal ideologies to
value distinct forms of consumer differentiation. Furthermore, we propose that this happens

8
!
because people adopt hierarchy beliefs that fit and validate their ideology. Hence, the causal
effect extends from political ideology to hierarchy beliefs and then to differentiation preferences.
We propose this causal direction because political identification forms very early in life as
a result of family socialization and, to some extent, heritability (Alford, Funk, and Hibbing 2005;
Lyons 2017). Early exposure to parents’ political opinions and family values (such as the
emphasis placed by parents on authority and obedience) as well as genetic factors (such as
inherent extraversion and certain gene variants) significantly shape individuals’ political leanings
from a young age, leading children’s political identification to reflect that of their parents (Alford
et al. 2005; Fraley et al. 2012; Lyons 2017). In adulthood, political ideology becomes a salient,
internalized, and visible facet of social identity (Vernon and Allport 1931). Just like other key
identity facets such as gender and race, political identity triggers, quickly and implicitly, the
enactment of consistent beliefs in various domains including social norms, out-group members,
and public policies (Iyengar and Westwood 2014; Nail et al. 2009). For example, when a
conservative (vs. liberal) ideology is made salient, individuals become more averse to constraints
on their freedom imposed by small product assortments (Fernandes and Mandel 2014). In
contrast, beliefs about hierarchy legitimacy require some understanding of how society functions
and what roles different groups play in it (Pratto et al. 1994). In adults, hierarchy beliefs are not
particularly salient, visible, or constructive of social identity, and they are just one part of a
complex framework of mental models that comprise political ideology (Jost et al. 2009). This
makes the reversed causal direction from hierarchy legitimacy beliefs to political ideology less
likely. Next, we describe how the hierarchy legitimacy beliefs endorsed by conservative and
liberal ideologies translate to preferences for two distinct forms of consumer differentiation.
Two Forms of Consumer Differentiation

9
!
The desire to differentiate from others is a powerful driver of consumer behavior (Brewer
1991; Snyder and Fromkin 1980). While differentiation has a number of intrinsic psychological
benefits such as boosting self-esteem and positive emotions (Rubin and Hewstone 1998),
individuals often differentiate through consumption choices to signal their distinct identity to
others (Brewer 1991; White and Dahl 2006).
Previous research separately examines two forms of consumer differentiation. On the one
hand, consumers may differentiate vertically in the social hierarchy to distinguish themselves on
the basis of their superior role, position, or success compared to others (Belk 1988; Vignoles et al.
2000). On the other hand, consumers may differentiate horizontally in the hierarchy to distinguish
themselves on the basis of their unique traits and values (Brewer 1991; Vignoles et al. 2000). To
achieve vertical differentiation in the hierarchy, consumers buy products that signal that they are
better than others (Dommer et al. 2013; Han et al. 2010). For example, they may buy products
from higher status brands compared to what other people have (Berger and Heath 2008;
Ordabayeva and Chandon 2011). To achieve horizontal differentiation in the hierarchy,
consumers buy products that signal that they are unique from others (Tian et al. 2001). For
example, they may buy products with unusual features such as a bold color (Bellezza et al. 2013;
Chan et al. 2012).
Importantly, prior studies of consumer divergence each focus on one type of
differentiation (vertical or horizontal) by prompting participants to choose between group
assimilation and one type of differentiation. In an exception, Dommer and colleagues (2013)
show that low self-esteem individuals secure their group belonging by engaging in horizontal
differentiation when socially excluded and in vertical differentiation when socially included.
Furthermore, Chan and colleagues (2012) show that people balance their need to belong with
their need to stand out by assimilating to their group vertically (brand) and by diverging from

10
!
their group horizontally (color). Extending these findings, we propose that political ideology is a
new driver of consumers’ preferences between vertical and horizontal differentiation.
We predict that conservatism leads consumers to pursue vertical differentiation in the
social hierarchy through products that signal that they are better than others. This is because
conservatism views the dominance-based hierarchical structure as a legitimate method to
distinguish individual qualities, which implies that vertical differentiation should enable
individual distinction by showing one’s superiority over others. We predict that liberalism leads
consumers to pursue horizontal differentiation in the social hierarchy through products that signal
that they are unique from others. This is because liberalism views the hierarchical structure as an
illegitimate method to distinguish individual qualities, which implies that horizontal
differentiation should enable distinction by showing one’s uniqueness from others in other ways.
H1: As consumers’ political conservatism (liberalism) increases, so does their preference for
vertically- (horizontally-) differentiating products in the social hierarchy.
H2: The relationship between consumers’ political ideology and differentiation preferences is
mediated by their beliefs about the legitimacy of the dominance-based hierarchical structure.
Next, we discuss the boundary conditions that mitigate the effect of political ideology on
differentiation preferences. We build on prior reports that the influence of political ideology on
attitudes and behaviors depends on factors in the environment (Iyengar and Westwood 2014),
suggesting that the effects of ideology may be susceptible to situational influences and overrides.
Moderating Effects of Manipulated Differentiation Goals and Manipulated Hierarchy Beliefs
We propose that the hypothesized pattern emerges because of conservatives’ and liberals’
goal to differentiate themselves vertically or horizontally in the hierarchy, rather than alternative
motives such as a need for high product quality or uniqueness. Hence, to test our first boundary
condition, we manipulate individuals’ vertical or horizontal differentiation goals in the hierarchy.

11
!
Prior literature on goals establishes that externally manipulating individuals’ goals brings
their behaviors in line with these goals (Gollwitzer 1999). For example, prompting an
assimilation goal leads individuals to exhibit behaviors that assimilate them to their social
environment (Pickett, Silver, and Brewer 2002). Extending this logic to the effect of political
ideology on consumers’ differentiation goals in the hierarchy, we expect that the effect of
ideology on consumers’ distinct differentiation preferences will disappear when vertical or
horizontal differentiation goals in the hierarchy are externally prompted. Specifically, prompting
a goal to differentiate vertically in the hierarchy should motivate individuals to pursue vertical
differentiation in the hierarchy (Han et al. 2010) – as we predict only conservatives do without
the goal prompt. Furthermore, prompting a goal to differentiate horizontally in the hierarchy
should motivate individuals to pursue horizontal differentiation in the hierarchy (Chan et al.
2012) – as we predict only liberals do without the goal prompt.
H3: Externally prompting a goal to differentiate vertically vs. horizontally in the social hierarchy
moderates the relationship between political ideology and differentiation preferences:
a. Externally prompting a vertical differentiation goal increases consumers’ general desire to
differentiate vertically in the hierarchy, as only conservatives do without a goal prompt.
b. Externally prompting a horizontal differentiation goal increases consumers’ general desire
to differentiate horizontally in the hierarchy, as only liberals do without a goal prompt.
Finally, since opposing beliefs about hierarchy legitimacy mediate the effect of ideology
on differentiation preferences, manipulating hierarchy beliefs by externally framing the hierarchy
as legitimate – that is, as resulting from hard work, or illegitimate – as resulting from luck, should
moderate the effect of ideology and help us test the psychological process through moderation
(Spencer, Zanna, and Fong 2005). Specifically, we expect external manipulation of hierarchy
legitimacy to change the differentiation preferences of conservatives, but not those of liberals.
We expect this asymmetric effect because conservatives have a high need to justify the
social system – that is, they have a strong desire to rationalize the existing social arrangement. In

12
!
contrast, liberals have a low need to justify the system (Jost et al. 2003; Jung et al. 2017). These
differences in system justification tendencies make conservatives and liberals asymmetrically
susceptible to external manipulations of the social system: whereas conservatives respond to
manipulations of the social system by shifting their views to accept the present system, liberals
are not susceptible to manipulations of the social system (Carraro, Castelli, and Macchiella 2011).
For example, externally changing justifications of inequality by asking people to consider the
reasons behind rich people’s advantage changes the attitudes of conservatives, but not liberals,
toward redistribution (Chow and Galak 2012). Similarly, externally framing luck as an individual
quality rather than a random occurrence changes conservatives’, not liberals’, view of the role of
luck in determining individuals’ hierarchical positions (Gromet, Hartson, and Sherman 2015).
This suggests that conservatives’ views of the hierarchy can be changed by externally shifting the
framing of the hierarchy, but liberals’ views of the hierarchy are not changed by such prompts.
Building from these insights, we predict that experimentally framing the dominance-based
hierarchical social structure as legitimate will lead conservatives to align their views with the
hierarchical structure’s legitimacy, resulting in their endorsement of the hierarchy and preference
for vertical over horizontal differentiation. Framing the dominance-based hierarchical structure as
illegitimate will lead conservatives to accept the hierarchical structure’s illegitimacy, resulting in
their opposition to the hierarchy and preference for horizontal over vertical differentiation. In
contrast, we predict that liberals will be unsusceptible to external manipulations of hierarchy
legitimacy: liberals will view the hierarchical structure as illegitimate and will prefer horizontal
to vertical differentiation regardless of the external framing of the legitimacy of the hierarchy. In
effect, the divide in conservatives’ and liberals’ differentiation preferences will emerge in the
legitimate hierarchy condition, but not in the illegitimate hierarchy condition.

13
!
H4: Externally framing the dominance-based hierarchical structure as legitimate vs. illegitimate
moderates the relationship between political ideology and differentiation preferences:
a. The relationship between political ideology and differentiation preferences emerges when
the hierarchy is externally framed as legitimate, in which case political conservatism
(liberalism) results in preference for vertically- (horizontally-) differentiating products.
b. The relationship between political ideology and differentiation preferences disappears when
the hierarchy is externally framed as illegitimate, in which case both political conservatism
and liberalism result in preference for horizontally-differentiating products.
Empirical Plan
Figure 1 outlines our theoretical model and the seven studies that test this model.
--- Insert Figure 1 about here ---
Studies 1A-1C test the influence of inherent (Studies 1A-B) and manipulated (Study 1C)
political ideology on preferences for vertically- vs. horizontally-differentiating products (H1), and
they test the mediating role of hierarchy legitimacy beliefs (H2). Study 2 demonstrates the effect
of ideology on the individual appeal of vertically- and horizontally-differentiating products across
a wide range of incomes. Studies 3 and 4 examine the moderating effects of manipulated
differentiation goals (Study 3: H3) and manipulated hierarchy legitimacy (Study 4: H4). Finally,
Study 5 tests the theory using secondary data of online searches in conservative and liberal U.S.
states. Across studies, we use established measures as well as a manipulation of political
ideology, real and hypothetical product choices, diverse participant samples, and immediate and
delayed measures of ideology to establish the robustness and generalizability of the effects.
STUDY 1A: INFLUENCE OF MEASURED IDEOLOGY ON PREFERENCE FOR
VERTICAL VS. HORIZONTAL DIFFERENTIATION, AND UNDERLYING PROCESS
Study 1A had three goals. First, the study sought to establish the link between political
ideology and differentiation preferences (H1) using a validated measure of ideology from the
literature. Second, the study sought to examine the mediating role of hierarchy legitimacy beliefs

14
!
(H2). Third, the study sought to demonstrate these patterns with real product choice. We held the
product constant across differentiation types, and we varied the message displayed on the
product. This allowed us to rule out alternative explanations related to idiosyncratic differences
between vertically- and horizontally-differentiating products, and to rule out the role of
conservatives’ and liberals’ alternative motives such as their distinct desire for product quality.
Method
We recruited 169 students (Mage = 22, 59% female) on a university campus in exchange
for course credit.
Political ideology. Participants completed a single-item political ideology scale (Jost
2006; 1 = “extremely liberal” to 9 = “extremely conservative,” M = 4.74, SD = 1.82), which is
highly consistent with other measures of ideology (e.g., Jost et al. 2003; Nail et al. 1999) and is
highly predictive of voting patterns (Jost 2006).
Preference for vertical vs. horizontal differentiation. After completing the ideology scale,
participants were informed that 10 study participants would win a coffee mug, and that they could
choose between two mug designs. Both mug designs featured the university logo on one side, but
they featured two different engraved messages on the other side: vertically-differentiating “Just
Better” or horizontally-differentiating “Just Different.” In a pre-test, 68 students (Mage = 21, 37%
female) rated a mug with one of the two randomly-assigned messages (from 1 = “not at all” to 7
= “very much”) on vertical differentiation (“to what extent does the mug message signal that one
is different from others by enhancing their social standing – that is, by highlighting their
superiority and status among other people”) and on horizontal differentiation (“to what extent
does the mug message signal that one is different from others based on who they are as an
individual – that is, the qualities, opinions, and traits that make one different from other people”).

15
!
Table 1 shows the products used in Studies 1A through 4, and their vertical and horizontal
differentiation ratings obtained using the same scales across pre-tests. Table 1 shows that the
vertically-differentiating mug more strongly enabled vertical (vs. horizontal) differentiation, and
that the horizontally-differentiating mug more strongly enabled horizontal (vs. vertical)
differentiation. Another pre-test with 66 students (Mage = 22, 52% female) verified that the two
mugs were equally generally differentiating (“to what extent would the mug allow one to
differentiate from other people in the social environment”; 1 = “not at all”, 7 = “very much”; M =
4.06 and 3.85 for the vertically- and the horizontally-differentiating mug respectively, t(62) = .49,
p = .63). All pre-tests in this paper sampled the relevant population for each study.
--- Insert Table 1 about here ---
Participants chose between the two mug designs using a dichotomous scale. The
presentation order of each mug at each anchor was randomized, and it is included in the analyses.
Process. To examine the mediating role of hierarchy beliefs, we administered the Social
Dominance Orientation scale (Pratto et al. 1994; 16 items; α = .92; M = 2.55, SD = 0.98), which
captures the belief of individuals in the legitimacy of the dominance-based hierarchical structure
(e.g., “some groups of people are simply inferior to others,” 1 = “not at all” to 7 = “very much”).
Demographics. Since prior research links political ideology to the age and socioeconomic
status (SES) of individuals (Ghitza and Gelman 2014; McCarty, Poole, and Rosenthal 2006), we
measured age and SES (from 1 = “high class” to 5 = “low class”; M = 3.60, SD = 0.66) at the end
of the survey. We did not measure income in this study, because most students do not earn an
income. We measured income in subsequent studies that involved non-student adult participants.
Results
Effect of ideology on differentiation preference. A binary logistic regression on the
dichotomous choice of the vertically- (vs. horizontally-) differentiating mug with mean-centered

16
!
ideology as a predictor revealed a significant positive coefficient of ideology (b = .32, Wald =
10.67, p = .001), indicating that conservatives were more likely to choose the vertically-
differentiating mug over the horizontally-differentiating mug than were liberals.
Adding the mugs’ presentation order (.5 when the vertically-differentiating mug was
presented on the left, -.5 when it was on the right), participants’ age and SES (both mean-centered
around 0), and their interactions with ideology to the model did not change the results. Ideology
still significantly predicted differentiation choice (ideology: b = .62, Wald = 4.24, p = .04;
presentation order: b = .66, Wald = 2.70, p = .10; age: b = .54, Wald = 2.81, p = .09; SES: b =
.57, Wald = 3.09, p = .08; all interactions: b’s < .28, Wald’s < 1.36, p’s > .24). Table 2 shows the
coefficient estimates obtained in the analyses that included demographics in Studies 1A-4.
---Insert Table 2 and Figure 2 about here---
Process. In order to test the mediating role of hierarchy beliefs, we ran a mediation model
4 in SPSS (Hayes 2013) with 10,000 bootstrap samples and 95% confidence intervals. The results
revealed a significant indirect effect of ideology on differentiation preferences through hierarchy
beliefs (a = .1841, SE = .0668, 95% CI = [.0724, .3343]). The reversed mediation model (model 4
with 10,000 bootstrap samples and 95% confidence intervals) testing if ideology mediated the
effect of hierarchy beliefs yielded a non-significant indirect effect (a = .1421, SE = .1080, 95% CI
= [-.0486, .3808]). These results confirmed the hypothesized order of the variables in the causal
chain (H2). Figure 2 shows the mediation results from Studies 1A-C.
Discussion
Study 1A demonstrated the link between political ideology and differentiation preferences
(H1) using a well-established measure of ideology and real product choice. The study showed that
this link is independent from individuals’ age and SES, and that it is explained by the diverging
beliefs of conservatives and liberals in the dominance-based hierarchy’s legitimacy (H2). While

17
!
Study 1A controlled for potential alternative accounts by keeping product features such as quality
constant, one could argue that in contexts where product features naturally vary, conservatives
may prefer vertically-differentiating products because of their high need for quality, and liberals
may prefer horizontally-differentiating products because of their high need for uniqueness. To
test this possibility, Study 1B features real vertically- and horizontally-differentiating brands and
tests alternative accounts such as need for quality and uniqueness through competing mediation.
STUDY 1B: ALTERNATIVE EXPLANATIONS
Study 1B had three objectives. First, the study sought to generalize the effect of ideology
to a multi-item scale of political ideology (Kidwell et al. 2013). Second, the study sought to
generalize the effect of ideology to a real choice between brands that actually promote vertical or
horizontal consumer differentiation in the marketplace, in order to test whether the results of
Study 1A found with internally-valid stimuli (mugs) would also hold with externally-valid stimuli
(brands). Importantly, the study examined alternative explanations including consumers’ need for
quality and uniqueness by measuring these constructs and by checking their mediating effects.
Method
Three hundred and thirty-three (Mage = 34, 63% female) adult participants completed the
study online for payment on Amazon Mechanical Turk.
Political ideology. Participants completed, in random order, the single-item scale of
ideology used in Study 1A (Jost 2006; M = 4.61, SD = 2.10) and a multi-item scale of ideology
from the literature (Kidwell et al. 2013; M = 4.55, SD = 1.79; from 1 = “strongly oppose” to 9 =
“strongly support” capital punishment, abortion*, gun control*, socialized healthcare*, same-sex
marriage*, illegal immigration*, and democrats*, where * marks reversed items; α = .83).

18
!
Preference for vertical vs. horizontal differentiation. Participants were informed that they
could win a $100 gift card from one of two clothing brands, and they chose, using a dichotomous
scale, between Ralph Lauren, which is positioned as promoting vertical consumer differentiation,
and Urban Outfitters, which is positioned as promoting horizontal differentiation in the
marketplace. Table 1 shows that in a pre-test involving adults online, the two brands signaled
vertical or horizontal differentiation as intended. Another pre-test with 81 adults (Mage = 34, 65%
female) confirmed that Ralph Lauren and Urban Outfitters were equally generally differentiating
(M = 4.45 and 4.61, t(79) = -.45, p = .65, on the same scale as in the Study 1A pre-test).
Process. To examine the process, participants completed multiple scales in random order.
To test the hypothesized mediating role of hierarchy legitimacy beliefs, participants completed
the Social Dominance Orientation scale from Study 1A (Pratto et al. 1994; M = 2.29, SD = 1.12;
α = .94). To test the role of consumers’ need for quality as an alternative explanation, participants
completed three items measuring desire for quality (M = 5.68, SD = 1.12; 1 = “not at all” to 7 =
“very much”: “product quality is very important to me,” “I pay a lot of attention to product
quality,” “I seek out products of the highest quality”; α = .91), as well as two established Quality
Consciousness scales from the literature (the first scale was adopted from Sproles and Kendall
1986; M = 5.43, SD = 1.11; 7 items such as, “Getting very good quality is very important to me”;
α = .89; and the second scale was adopted from Ailawadi, Pauwels, and Steenkamp 2008, M =
4.66, SD = 1.11; 3 items such as, “Quality is decisive for me while buying a product”; α = .65).
To test the alternative role of consumers’ general need for uniqueness, participants completed the
Need for Uniqueness scale (Tian et al. 2001; M = 3.54, SD = 1.12; 12 items such as, “I actively
seek to develop my personal uniqueness by buying special products or brands”; α = .91).
Demographics. At the end of the survey, participants indicated their age, income (from 1
= “under $20,000” to 15 = “$150,000 or more”; the scale progressed in $10,000 increments; M =

19
!
5.94, SD = 4.09), and their SES on the MacArthur scale featuring a socioeconomic ladder (Adler
et al. 2000; from 1 = “lowest social tier” to 10 = “highest social tier”; M = 5.17, SD = 1.60).
Results
Effect of ideology on differentiation preference. A binary logistic regression on choice of
a gift card from a vertically- (vs. horizontally-) differentiating brand with a mean-centered multi-
item ideology scale as a predictor revealed a significant positive coefficient of ideology (b = .19,
Wald = 8.51, p = .004), confirming that conservatives were more likely to prefer the vertically-
differentiating brand over the horizontally-differentiating brand than were liberals. (All the results
were identical when the single-item ideology scale was used in the analyses: b = .16, Wald =
8.81, p = .003. The results featuring the single-item ideology scale are detailed in Web Appendix
A and are therefore no longer discussed.) Table 2 shows that this effect of ideology remained
significant when demographics were added to the analysis (b = .19, Wald = 7.91, p = .005).
Process. To test the process underlying this effect, we ran a mediation model 4 in SPSS
(Hayes 2013) with 10,000 bootstrap samples and 95% confidence intervals with hierarchy beliefs,
desire for quality, the two quality consciousness scales, and need for uniqueness as competing
mediators. The results revealed that the indirect effect of ideology on differentiation preferences
through hierarchy beliefs was significant (a = .0844, SE = .0397, 95% CI = [.0109, .1741]).
Importantly, the indirect effect of ideology through desire for quality (a = -.0053, SE = .0108,
95% CI = [-.0321, .0121]), the two quality consciousness scales (a = .0055, SE = .0115, 95% CI
= [-.0147, .0335] for the first scale; a = .0011, SE = .0075, 95% CI = [-.0118, .0194] for the
second scale) and need for uniqueness (a = .0018, SE = .0059, 95% CI = [-.0089, .0160]) was not
significant. As in Study 1A, the reversed mediation model (model 4 with 10,000 bootstrap
samples and 95% confidence intervals) that used hierarchy beliefs as the independent variable

20
!
and ideology as the mediator was not significant (a = .0891, SE = .0619, 95% CI = [-.0331,
.2113]), once again confirming the hypothesized order of variables in the theoretical model.
Discussion
While Study 1A controlled for potential alternative explanations by holding various
product characteristics constant, Study 1B measured alternative explanations directly and found
that hierarchy beliefs, rather than alternative constructs, mediate the influence of consumers’
ideology on differentiation preferences. Building on Studies 1A and 1B that examined the
influence of measured ideology on differentiation preferences, Study 1C tests the causal effect of
ideology using a manipulation of ideology, and it once again examines the underlying process.
STUDY 1C: EFFECT OF MANIPULATED IDEOLOGY ON PREFERENCE FOR
VERTICAL VS. HORIZONTAL DIFFERENTIATION, AND UNDERLYING PROCESS
Study 1C tested the causal effect of ideology on differentiation preferences by
manipulating ideology (H1), and it examined the role of hierarchy beliefs in driving this causal
effect (H2). In order to control for confounds such as product quality and for alternative
explanations such as desire for quality as we did in Study 1A, we kept the product, brand, price,
and quality ratings constant across differentiation types. We manipulated the vertically- or the
horizontally-differentiating positioning of the product by varying the product description, in line
with how products often attain their positioning through descriptions used in advertising.
Method
One hundred and eighty-one (Mage = 33, 67% female) adult participants completed the
study online for payment on Amazon Mechanical Turk.
Political ideology. To manipulate ideology, we had participants perform a recall task. In
the conservative (vs. liberal) condition, participants read the following instructions: “Please

21
!
remember a time when you were talking to or interacting with someone who was obviously more
liberal (vs. conservative) than you, and you felt that you had a more conservative (vs. liberal)
position. Please take some time to think about the situation, and please describe it in detail,
including whom you were with, what you did, what you and/or they said, how you felt, etc.” Our
manipulation was inspired by recent work which suggests that social context influences
individuals’ perceptions of their political identity: a moderate individual may self-identify as a
conservative in a strongly liberal geographic region, but as a liberal in a strongly conservative
region (Feinberg et al. 2017). After the recall task, participants indicated their ideology on the
single-item scale from Studies 1A-B (Jost 2006; M = 4.16, SD = 2.11) as a manipulation check.
A pre-test with 84 adults online (Mage = 34; 71% female) verified that the conservative
recall task led to a more conservative ideology than the liberal recall task on a single-item scale
(M = 5.08 vs. 3.56 respectively, F(1, 82) = 13.49, p < .001), a multi-item scale (M = 4.92 vs. 4.15
respectively, F(1, 82) = 4.45, p = .038; α = .80; Kidwell et al. 2013), and a dichotomous scale
(conservative vs. liberal; 53.85% vs. 24.44% respectively identified as conservative, χ2(1) = 7.66,
p = .007; Tetlock et al. 1984). The pre-test also verified that the ideology manipulation did not
influence self-esteem (p = .37, 10 items, α = .88, Rosenberg 1965), mood (positive: p = .24, 10
items, α = .83; negative: p = .54, 10 items, α = .87; Watson, Clark, and Tellegen 1988), subjective
status (p = .46, MacArthur ladder, 1 = “lowest social tier” to 10 = “highest social tier”; Adler et
al. 2000), or desire for quality (p = .11; three items from Study 1B; α = .80).
Preference for vertical vs. horizontal differentiation. After the manipulation check,
participants imagined that they were considering purchasing athletic shoes by a fictitious brand,
Setia, and they were choosing between the brand’s two product lines – a vertically-differentiating
line and a horizontally-differentiating line. They read a description of each line, which was
adopted from Warren and Campbell (2014). The vertically-differentiating line was “challenging

22
!
the industry standards by offering athletic shoes that are luxurious and classy, [and that] unlike
standard athletic shoes, are posh, elegant, and sometimes extravagant.” The horizontally-
differentiating line was “breaking with the industry standards by offering shoes that are edgy and
unique, [and that] unlike standard athletic shoes, are irreverent, hip, and sometimes eccentric.” To
control for potential confounds related to product quality, the price of both lines was $100, and
both lines were described as having received a rating of 94 (excellent) from Consumer Reports.
Table 1 shows the stimuli and the pre-test results confirming that the two shoe lines were
vertically- or horizontally-differentiating as intended. A separate pre-test with 58 adults (Mage =
33; 54% female) verified that the two lines were equally generally differentiating (M = 5.23 vs.
5.50 for the vertically- and horizontally-differentiating line, t(56) = -.75, p = .46, on the same
scale as in prior pre-tests). Since the stimuli were specifically designed to keep product quality
constant, the second pre-test also verified that the two shoe lines were perceived to have similar
quality (M = 5.77 vs. 5.50 for the vertically- and horizontally-differentiating line, t(56) = .85, p =
.40; from 1 = “not at all” to 7 = “very much”: “to what extent do you think the shoes are high
quality” and “how well do you think the shoes would serve their purpose”; r = .67).
Participants indicated which line was more appealing and attractive to them (1 =
horizontally-differentiating line, 7 = vertically-differentiating line; r = .92, M = 3.88, SD = 2.04).
Process. We administered the Social Dominance Orientation scale used in Studies 1A-B
(Pratto et al. 1994; α = .93; M = 2.13, SD = 1.03) to test the mediating role of hierarchy beliefs.
Demographics. At the end of the study, participants indicated their age, income (1 =
“under $20,000” to 15 = “$150,000 or more”; the scale progressed in $10,000 increments; M =
5.64, SD = 3.58), and SES on the MacArthur scale that features a socioeconomic ladder (Adler et
al. 2000; 1 = “lowest social tier” to 10 = “highest social tier”; M = 5.12, SD = 1.53).
Results

23
!
The manipulation check confirmed that the conservative recall task led to more
conservative ideology (M = 4.75) than the liberal task (M = 3.59, F(1, 179) = 14.90, p < .001).
Effect of ideology on differentiation preference. An ANOVA on preference for the
vertically-differentiating shoe line over the horizontally-differentiating line with the ideology
manipulation as a fixed factor revealed a significant effect: the preference for vertical (vs.
horizontal) differentiation was significantly higher in the conservative (M = 4.20) than in the
liberal condition (M = 3.58, F(1, 179) = 4.34, p = .039). Table 2 shows that this effect of ideology
remained significant when demographics were added to the model (b = .71, t = 2.36, p = .019).
Process. A mediation model 4 with 10,000 bootstrap samples and 95% confidence
intervals revealed a significant indirect effect of the ideology manipulation on differentiation
preferences through hierarchy beliefs (a = .1145, SE = .0719, 95% CI = [.0114, .3108]).
Discussion
The results of Study 1C complemented the findings of Studies 1A and 1B by establishing
the causal effect of ideology on differentiation preferences (H1), by confirming the role of
hierarchy legitimacy beliefs as the psychological process behind this effect (H2), and by
controlling for alternative possibilities. Study 2 tests how ideology impacts the individual appeal
of vertical and horizontal differentiation by manipulating differentiation type between-subjects,
instead of forcing participants to directly trade off between vertical and horizontal differentiation.
Study 2 also tests the role of income and SES more completely.
STUDY 2: INFLUENCE OF IDEOLOGY ON INDIVIDUAL APPEAL OF VERTICAL
AND HORIZONTAL DIFFERENTIATION ACROSS INCOME BRACKETS
Study 2 had three goals. First, it examined how ideology impacts the individual appeal of
vertical and horizontal differentiation. Second, whereas Studies 1A-C showed that ideology

24
!
effects are independent from SES and income among student participants who have no income
and M Turk participants who have relatively low income, Study 2 sought to further disentangle
the role of ideology from SES and income by using a sample of participants with a wide range of
incomes. Third, Study 2 sought to generalize the effect of ideology by using a different measure
of ideology and a new product. In order to control for confounds and alternative explanations, we
held the brand and the color of the product constant. In line with the operationalization of
consumer differentiation through product color in prior work (Bellezza et al. 2013; Warren and
Campbell 2014), we varied the differentiation framing of product color in a magazine article.
Method
Study 2 had a 2 × 2 design with ideology (conservative vs. liberal) measured and
differentiation type (vertical vs. horizontal) manipulated between-subjects. Three hundred and
fifty-three Qualtrics panelists (Mage = 48, 46% female) completed the study for payment. The
sampling method required a similar representation of participants from four brackets of annual
income: (1) $0 to $49,999, (2) $50,000 to $99,999, (3) $100,000 to $149,999, and (4) $150,000 or
above. In the result, 24%, 23%, 23%, and 30% of participants were in each respective bracket.
Political ideology. Participants indicated their political ideology on a dichotomous scale
(Tetlock et al. 1984; 54.96% identified as conservative and 45.04% identified as liberal).
Evaluation of vertical and horizontal differentiation. Participants imagined that they were
choosing an outfit to wear to a professional networking event. Participants read that most
attendees would probably wear black and that they were considering wearing a red outfit from J
Crew. We used the J Crew brand, because in a separate pre-test it was rated to be equally
vertically- and horizontally-differentiating (N = 60 adults, Mvertical = 4.82 vs. Mhorizontal = 4.58,
t(59) = 1.06, p = .29, on the same scales as in prior pre-tests). To manipulate differentiation type,
we varied between-subjects the framing of the color red as vertically or horizontally

25
!
differentiating. Participants read that they came across an article that described red as “the color
of luxury” and as being increasingly used to signal “one’s success, accomplishments, and
prosperity” in the vertical differentiation condition, or as “the color of nonconformity” and as
being increasingly used to signal “one’s originality, rebelliousness, and edge” in the horizontal
differentiation condition. Table 1 shows the stimuli and the pre-test results which confirmed that
the red outfit signaled vertical or horizontal differentiation in the two differentiation conditions as
intended. A separate pre-test with 57 adults (Mage = 33, 54% female) showed that the outfit was
similarly generally differentiating across the vertical and horizontal conditions (M = 5.38 vs. 5.96
respectively, t(54) = 1.52, p = .14, on the same scale as in prior pre-tests). Since the stimuli were
specifically designed to keep product quality constant, this pre-test also confirmed that the outfit
had similar perceived quality across conditions (M = 5.09 vs. 5.43 for vertical and horizontal
differentiation respectively, t(55) = -1.18, p = .24, on the same scale as in the Study 1C pre-test).
Participants indicated the extent to which wearing the red outfit would be appealing and
desirable for them (1 = “not at all” to 7 = “very much”; r = .93, M = 4.01, SD = 1.84).
Demographics. At the end, participants indicated their age, income (1 = “below $20,000”
to 20 = “$200,000 or above”; the scale progressed in $10,000 increments; M = 10.08, SD = 6.34),
and SES (M = 5.82, SD = 1.64 on the 10-point MacArthur ladder scale, Adler et al. 2000).
Results
We regressed the appeal of the differentiating outfit on ideology (.5 for conservative, -.5
for liberal), differentiation type (.5 for vertical, -.5 for horizontal), and their interaction. The
results revealed non-significant coefficients of ideology (b = -.04, t = -.20, p = .85) and
differentiation type (b = .26, t = 1.33, p = .18), but a significant ideology × differentiation
interaction (b = 1.17, t = 3.02, p = .003). Table 2 shows that this interaction remained significant
when demographics were added to the model (b = 1.20, t = 3.02, p = .003), whereas the effects of

26
!
income (|b|s < .05, |t|s < 1.10, p’s > .30) and SES (|b|s < .37, |t|s < 1.30, p’s > .20) were non-
significant. This once again confirmed that the effect of ideology is independent of demographics,
including income and SES, even among participants with wide-ranging incomes. As expected,
conservative ideology significantly increased the appeal of differentiation over liberal ideology in
the vertical differentiation condition (M = 4.43 for conservatives vs. 3.88 for liberals, b = .55, t =
2.20, p = .029), and it significantly decreased the appeal of differentiation in the horizontal
condition (M = 3.58 for conservatives vs. 4.21 and for liberals, b = -.63, t = -2.10, p = .037).
Whereas conservatives found the outfit to be significantly more appealing in the vertical (vs.
horizontal) condition (t = 3.22p = .002), liberals found the outfit to be directionally, albeit non-
significantly, less appealing in the vertical (vs. horizontal) condition (t = -1.15, p = .25).
Replication with Other Ideology Measures
We replicated the findings of Study 2 with a group of 320 M Turk participants (Mage = 35,
57% female), who completed, in random order, the single-item (Jost 2006; M = 4.34, SD = 2.03)
and multi-item (Kidwell et al. 2013; M = 4.33, SD = 1.82) ideology scales and then indicated
their interest in the red J Crew outfit using the measure of product appeal from Study 2. Analyses
revealed a significant ideology × differentiation interaction (b = .43, t = 4.32, p < .001 for the
single-item scale; b = .42, t = 3.67, p < .001 for the multi-item scale; |t|s < .64, p’s > .52 for the
remaining coefficients). Once again, conservative (vs. liberal) ideology significantly increased
the outfit’s appeal in the vertical condition (b = .19, t = 2.76, p = .006 for the single-item scale; b
= .17, t = 2.24, p = .027 for the multi-item scale), and it significantly decreased the outfit’s appeal
in the horizontal condition (b = -.24, t = -3.38, p = .001 for the single-item scale; b = -.24, t = -
2.97, p = .003 for the multi-item scale). Furthermore, conservative individuals (who scored 5.32
or higher on the single-item 1-to-9 ideology scale, and 5.39 or higher on the multi-item 1-to-9
scale) found the outfit to be significantly more appealing in the vertical (vs. horizontal) condition

27
!
(t = 1.97, p = .05 for both scales). Liberal individuals (who scored 3.24 or lower on the single-
item ideology scale, and 3.12 or lower on the multi-item scale) found the outfit to be significantly
less appealing in the vertical (vs. horizontal) condition (t = -1.98, p = .05 for both scales).
Discussion
Study 2 showed that political ideology influences the individual appeal of vertical and
horizontal differentiation independent from consumers’ income and SES. Conservative ideology
increases the appeal of vertically-differentiating products, and liberal ideology increases the
appeal of horizontally-differentiating products. Study 3 tests the role of consumers’ vertical and
horizontal differentiation goals in giving rise to such distinct preferences by explicitly prompting
the goal to differentiate vertically or horizontally in the hierarchy (H3).
STUDY 3: MANIPULATING DIFFERENTIATION GOALS
Study 3 had three objectives. First, Study 3 sought to verify that the hypothesized pattern
emerges due to effects of conservative and liberal ideologies on consumers’ vertical and
horizontal differentiation goals in the social hierarchy by manipulating a goal to differentiate
vertically in the hierarchy, a goal to differentiate horizontally in the hierarchy, or a neutral goal.
We predicted (H1) that conservative (vs. liberal) ideology would boost the preference for vertical
(vs. horizontal) differentiation under the neutral prompt. Furthermore, we predicted (H3) that
explicitly prompting a vertical or a horizontal differentiation goal would eliminate this effect of
ideology by increasing people’s general desire to differentiate vertically – as conservatives do
under the neutral prompt, or to differentiate horizontally – as liberals do under the neutral prompt.
Second, since our previous studies measured or manipulated the political ideology of
participants shortly before they made their product choices, one may question whether or not
ideology needs to be explicitly salient for the hypothesized effects to emerge. Third, since Studies

28
!
1A-C established the psychological process by measuring participants’ hierarchy beliefs shortly
after their product preferences, one may question whether or not the process evidence obtained in
those studies may have been inflated. Since the measurement of the mediator and the dependent
variable in the same survey may augment their correlation, introducing a temporal separation
between them may provide a more cautious test of the psychological process (Podsakoff,
Mackenzie, and Podsakoff 2012). To address these two issues, Study 3 measured ideology and
hierarchy beliefs one week after the measure of differentiation preferences.
Method
Two hundred and eighty adults (Mage = 34, 58% female) completed two surveys online for
a small payment through the TurkPrime application (Litman, Robinson, and Aberbock 2016).
Differentiation goal. In the first survey, participants completed a recall task that prompted
a vertical differentiation goal, a horizontal differentiation goal, or a neutral goal. In the vertical
(vs. horizontal) differentiation goal condition, participants read: “Please take a moment and think
of times when you felt very average (vs. conventional) compared to other people. In other words,
think of times and situations where you felt that you were so average (vs. conventional) compared
to other people that you lacked your own sense of accomplishment and success (vs. uniqueness
and identity).” These instructions were adapted from widely used goal induction methods (e.g.,
Pickett et al. 2002). To ensure that the task, mood, and self-esteem were similar across
conditions, in the neutral goal –control condition participants recalled times when they felt
uncomfortable around other people. All participants described two memories that fit the prompt.
A pre-test with 113 adults (Mage = 31, 52% female) verified that the goal manipulation did
not impact ideology (single-item scale: p = .83; multi-item scale: p = .75; dichotomous scale: p =
.36), self-esteem (p = .72), mood (negative: p = .38; positive: p = .32), or subjective status (p =
.32). Importantly, it verified that the goal manipulation significantly impacted the desire (from 1

29
!
= “not at all” to 7 = “very much”) to differentiate vertically (F(2, 111) = 5.17, p = .007; “to what
extent do you want to be considered as holding high prestige”) and the desire to differentiate
horizontally (F(2, 111) = 5.77, p = 004; “to what extent do you want to be considered as being
nonconformist”). The desire to differentiate vertically was higher in the vertical (M = 4.84) than
in the horizontal (M = 3.47, t(71) = 3.06, p = .003) and neutral (M = 3.71, t(76) = 2.56, p = .012)
goal conditions, and it did not differ between the latter two conditions (t(75) = -.51, p = .61). The
desire to differentiate horizontally was higher in the horizontal (M = 4.39) than in the vertical (M
= 3.08, t(71) = 3.35, p = .001) and neutral (M = 3.32, t(75) = 2.62, p = .011) goal conditions, and
it did not differ between the latter two conditions (t(76) = -.58, p = .56).
Preference for vertical vs. horizontal differentiation. After the recall task, participants
were informed, as in Study 1A, that 10 participants would win a coffee mug, and they chose
between two mug designs – one that featured a vertically-differentiating message “Just Better”
and one that featured a horizontally-differentiating message “Just Different.” Whereas in Study
1A the other side of the mug featured the logo of student participants’ university, to make the
mug personally relevant to M Turk participants in the present study, we informed them that the
other side of the mug would feature a name of their choosing (e.g., the name of their organization
or alma mater). Table 1 shows that a pre-test confirmed that the two mugs signaled vertical or
horizontal differentiation as intended. A separate pre-test with 53 adults (Mage = 34, 62% female)
confirmed that the vertically- and horizontally-differentiating mugs were equally generally
differentiating (M = 3.74 and 4.04, t(51) = .49, p = .63, on the same scale as in prior pre-tests).
Political ideology and process. One week after the first survey, participants were re-
contacted through TurkPrime for the second unrelated survey, and they completed, in random
order, the measures of ideology on a dichotomous scale (Tetlock et al. 1984; 37.50% were

30
!
conservative, 62.50% were liberal) and hierarchy beliefs using the Social Dominance Orientation
scale from Studies 1A-C (α = .93, M = 2.30, SD = 1.10).
Demographics. At the end of second survey, we measured age, annual income (1 = “under
$20,000” to 15 = “$150,000 or more”; the scale progressed in $10,000 increments; M = 5.70, SD
= 3.77), and SES (1 = “high class” to 5 = “low class”; M = 3.37, SD = 0.80).
Results
Effects of differentiation goal and ideology on differentiation preferences. Since we
expected a significant effect of ideology only in the neutral goal condition – that is, in one out of
the three goal conditions – following prior work (Naylor, Lamberton, and Norton 2013), we
created two dummy variables to capture the three goal conditions: dummy 1 was equal to .66 in
the neutral condition and -.33 in the vertical and horizontal goal conditions; and dummy 2 was
equal to .5 in the vertical goal condition, -.5 in the horizontal goal condition, and 0 in the neutral
condition. We then conducted a logistic regression on choice of the vertically- (vs. horizontally-)
differentiating mug with ideology (.5 for conservative; -.5 for liberal), the two dummy variables,
and the interaction of each dummy variable with ideology as predictors.
---Insert Figure 3 about here---
The results revealed significant positive coefficients of ideology (b = .75, Wald = 8.29, p
= .004) and dummy 2 (vertical vs. horizontal differentiation: b = .83, Wald = 7.21, p = .007), with
preference for vertical over horizontal differentiation significantly higher among conservatives
(53.33%) than liberals (34.29%) and when the vertical differentiation goal was prompted (52.0%)
than when the horizontal goal was prompted (31.87%). Importantly, the key interaction of
ideology with dummy 1 (neutral vs. vertical and horizontal differentiation goals) was marginally
significant (b = 1.01, Wald = 3.09, p = .079). The remaining coefficients were not significant (|b|s
< .34, Wald’s < .30, p’s > .58). Table 2 shows that these results remained unchanged when

31
!
demographics were included in the model (ideology: b = .77, Wald = 7.03, p = .008; dummy 2: b
= .70, Wald = 3.96, p = .047; ideology × dummy 1 interaction: b = 1.06, Wald = 2.84, p = .092).
Breaking down the key interaction of ideology with dummy 1 (neutral goal vs. vertical
and horizontal differentiation goals), as predicted (H3) and shown in Figure 3, conservative (vs.
liberal) ideology led to a significantly higher preference for vertical over horizontal
differentiation in the neutral condition (62.07% vs. 28.33% respectively, b = 1.42, Wald = 8.83, p
= .003), but it did not impact differentiation preferences when the vertical (b = .59, Wald = 2.08,
p = .15) or horizontal differentiation goal (b = .25, Wald = .28, p = .60) was explicitly prompted.
Compared to conservatives’ high baseline preference for vertical over horizontal differentiation in
the neutral condition (62.07%), conservatives’ choice of vertical differentiation did not change in
the vertical differentiation goal condition (60.0%, b = .09, Wald = .03, p = .86), but it
significantly decreased in the horizontal goal condition (35.48%, b = -1.09, Wald = 4.14, p =
.042). Compared to liberals’ low baseline preference for vertical over horizontal differentiation in
the neutral condition (28.33%), liberals’ choice of vertical differentiation did not change in the
horizontal goal condition (30.0%, b = .08, Wald = .04, p = .84), but it marginally significantly
increased in the vertical goal condition (45.45%, b = .75, Wald = 3.58, p = .059). In other words,
as expected, conservatives displayed similar preferences for vertical differentiation in the vertical
goal condition as they did in the neutral condition, and liberals displayed similar preferences for
horizontal differentiation in the horizontal goal condition as they did in the neutral condition.
Process. We tested whether hierarchy legitimacy beliefs mediated the main effect of
ideology on differentiation preferences. Since we expected the vertical and horizontal
differentiation goal prompts to boost consumers’ desire for vertical and horizontal differentiation
directly, without influencing hierarchy legitimacy beliefs, we ran a mediation model 5 (Hayes
2013), which states that the moderator (goal prompt) influences the dependent variable (mug

32
!
choice) without influencing the link between the independent variable (ideology) and the
mediator (hierarchy beliefs). The analysis with 10,000 bootstrap samples and 95% confidence
intervals revealed a significant indirect effect of ideology on choice through hierarchy beliefs (a =
.3767, SE = .1393, 95% CI = [.1420, .6908]). (A model 4 analysis also yielded a significant
indirect effect of ideology through hierarchy beliefs: a = .3805, SE = .1365, 95% CI = [.1529,
.6900].) Hierarchy beliefs thus mediated the main effect of ideology on differentiation
preferences even after a time delay. This minimizes the role of demand in facilitating the process.
Discussion
Study 3 verifies that the effect of ideology on distinct differentiation preferences is driven
by the desire of conservatives and liberals to differentiate vertically or horizontally in the
hierarchy. The study thereby adds to the results of Studies 1A-C and 2 in ruling out alternative
explanations related to the idiosyncratic preferences of conservatives and liberals for certain
products (e.g., high-quality products). Study 3 also shows that ideology effects emerge even if
ideology and hierarchy legitimacy beliefs are measured after a significant time delay. This
minimizes the possibility that demand effects drive our findings. Importantly, Study 3 proposes
the first set of conditions that mitigate the ideological divide in consumption preferences –
activating a vertical versus a horizontal differentiation goal in the hierarchy (H3). Study 4 tests the
second set of boundary conditions related to manipulated hierarchy legitimacy (H4).
STUDY 4: MANIPULATING PERCEIVED HIERARCHY LEGITIMACY
Study 4 sought to corroborate the role of hierarchy legitimacy beliefs in driving the effect
of ideology by manipulating, instead of measuring, hierarchy legitimacy perceptions. The study
thus used an established method to test the process through moderation (Spencer et al. 2005). We
predicted (H4) that prompting people to consider that the dominance-based hierarchy may be

33
!
illegitimate (vs. legitimate) would eliminate the effect of ideology on differentiation preferences,
and that this would result due to the shift in the preferences of conservatives, not liberals.
Method
Two hundred and seventy-two adults (Mage = 32, 65% female) completed the study on M
Turk for a small payment.
Political ideology. Participants indicated their ideology on a dichotomous scale (Tetlock
et al. 1984; 43.45% identified as conservative, 56.55% identified as liberal).
Hierarchy legitimacy. Participants then engaged in a recall task, which manipulated
perceptions of hierarchy legitimacy by stressing the relative importance of either hard work or
good luck in attaining high hierarchical positions. In the legitimate (vs. illegitimate) hierarchy
condition, participants were instructed as follows: “It takes a combination of good luck and hard
work for good things to happen to you. Please tell us the story of something good that happened
to you. In particular, we would like to know about the role that hard work, self-discipline, and
wise decisions (vs. chance, opportunity, and help from others) played in helping you get there.”
Two pre-tests, reported in the next section, verified the effectiveness of this manipulation.
Preference for vertical vs. horizontal differentiation. As in Study 1B, participants were
informed that they could win a $100 gift card from Ralph Lauren (vertical differentiation) or
Urban Outfitters (horizontal differentiation), and they made their choice on a dichotomous scale.
Demographics. At the end of the study, participants indicated their age, SES (from 1 =
“high class” to 5 = “low class”; M = 3.29, SD = 0.77), and income (from 1 = “under $20,000” to
15 = “$150,000 or more”; the scale progressed in $10,000 increments; M = 5.57, SD = 3.61).
Hierarchy Manipulation Pre-tests
Two pre-tests checked the effectiveness of the hierarchy manipulation. The first pre-test
included 80 adults (Mage = 31, 55% female) and verified that the hierarchy manipulation did not

34
!
impact ideology (single-item: p = .78; multi-item: p = .23; dichotomous: p = .28), self-esteem (p
= .87), mood (positive: p = .68; negative: p = .11), or subjective status (p = .74).
The second pre-test checked whether the hierarchy manipulation would effectively change
perceptions of hierarchy legitimacy, and if it would do so asymmetrically for conservatives and
liberals. We expected the hierarchy manipulation to shift the hierarchy beliefs of conservatives,
but not those of liberals, because conservatives have a high, and liberals have a low, need to
justify and accept the existing social system. A group of 153 adults (Mage = 32, 62% female)
completed the hierarchy manipulation and then indicated their belief in the legitimacy of the
dominance-based hierarchy (“to what extent do you think that people deserve the outcomes that
they get,” 1 = “not at all” to 7 = “very much”; M = 4.95, SD = 1.10). They also indicated their
political ideology on the dichotomous scale used in the main study (Tetlock et al. 1984; 34.95%
were conservative and 64.05% were liberal). A UNIANOVA revealed significant main effects of
the hierarchy manipulation (F(1, 149) = 4.89, p = .029; perceived hierarchy legitimacy was
higher in the legitimate condition: M = 5.09 than in the illegitimate condition: M = 4.81) and of
ideology (F(1, 149) = 4.34, p = .039; perceived hierarchy legitimacy was higher among
conservatives: M = 5.20 than liberals: M = 4.82). Importantly, as predicted, the hierarchy ×
ideology interaction was significant (F(1, 149) = 5.74, p = .018) such that the hierarchy
manipulation significantly boosted perceived hierarchy legitimacy among conservatives
(Mlegitimate = 5.61 vs. Millegitimate = 4.78; F(1, 149) = 8.29, p = .005), but not among liberals
(Mlegitimate = 4.80 vs. Millegitimate = 4.83; F(1, 149) = .02, p = .88). This ultimately resulted in
significantly different hierarchy legitimacy perceptions between conservatives and liberals in the
legitimate hierarchy condition (F(1, 149) = 10.25, p = .002), but not in the illegitimate hierarchy
condition (F(1, 149) = .05, p = .83).
Results

35
!
In the main study, we conducted a binary logistic regression on the choice of a gift card
from the vertically- (vs. horizontally-) differentiating brand with the hierarchy manipulation (.5
for legitimate, -.5 for illegitimate), ideology (.5 for conservative, -.5 for liberal), and their
interaction as predictors. The results revealed a significant coefficient of ideology (b = 1.03, Wald
= 15.46, p < .001) such that conservatives were more likely to choose the vertically- (vs.
horizontally-) differentiating brand (51.28%) than liberals (27.74%). The coefficient of the
hierarchy manipulation was positive, albeit non-significant (b = .37, Wald = .26, p = .15).
Importantly, the coefficient of the ideology × hierarchy manipulation was significant (b = 1.14,
Wald = 4.75, p = .029). These effects did not change when demographics were added to the
analysis (ideology: b = 1.12, Wald = 16.19, p < .001; ideology × hierarchy manipulation: b =
1.31, Wald = 5.55, p = .019; see Table 2 for details).
Breaking down the focal interaction, as predicted (H4) and shown in Figure 4, the effect of
conservative (vs. liberal) ideology was significant in the legitimate hierarchy condition (b = 1.60,
Wald = 17.55, p < .001), but it was not significant in the illegitimate hierarchy condition (b = .46,
Wald = 1.64, p = .20). The preference shift in response to the hierarchy manipulation emerged
because of the change in the preferences of conservatives, not liberals. Conservatives were less
likely to choose the vertically-differentiating brand when the hierarchy was cast as illegitimate
(40.0%) than when it was cast as legitimate (63.16%, b = .94, Wald = 6.16, p = .01). Liberals
were equally unlikely to choose the vertically-differentiating brand regardless of whether the
hierarchy was cast as legitimate (25.68%) or illegitimate (29.63%, b = -.20, Wald = .30, p = .58).
---Insert Figure 4 about here---
Discussion
Together, Studies 1A through 4 provide converging evidence of the effect of measured
and manipulated individual-level political ideology on consumers’ preferences for vertical and

36
!
horizontal differentiation in the hierarchy, and for hierarchy beliefs as the psychological process
behind this effect. Study 5 tests our theory using state-level political ideology and secondary data
from online searches in conservative and liberal U.S. states in order to boost external validity.
STUDY 5: IDEOLOGY AND ONLINE SEARCH
Study 5 tested whether the divide in differentiation preferences between conservatives and
liberals was prevalent in the online searches of conservative and liberal U.S. states. We used
Google search data to approximate the differentiation interests of individuals for two reasons.
First, Google searches are strongly related to the interests of individuals who perform the search,
and aggregating millions of searches gives a good proxy for an area’s interest in a topic (Ginsberg
et al. 2009; Goel et al. 2010). Second, Google search data can reliably predict actual behaviors in
a geographical area such as the consumption of movies, cars, homes, and video games (Goel et al.
2010), as well as broader outcomes such as the incidence of flu (Ginsberg et al. 2009) as long as
potential confounds are controlled for. This is because people search for a phenomenon (e.g., flu)
when it becomes prevalent in the area (when flu spreads) and for products before buying them. In
line with our theory, we predicted that the level of conservatism (vs. liberalism) of a state would
positively predict the state’s online search for vertical differentiation terms and negatively predict
the state’s online search for horizontal differentiation terms.
Google Search Terms
We selected 12 terms that represent vertical differentiation (e.g., “best”) and 12 terms that
represent horizontal differentiation (e.g., “distinct”). In a pre-test, 60 people (Mage = 35, 55%
female) rated the vertical differentiation terms, and 64 people (Mage = 37, 59% female) rated the
horizontal differentiation terms on the degree to which each term allowed vertical and horizontal
differentiation, using the descriptions of vertical and horizontal differentiation and 7-point scales

37
!
adopted from prior pre-tests. The terms that represent vertical differentiation were significantly
more vertically-differentiating (M = 5.86) than horizontally-differentiating (M = 3.32, t(59) =
10.93, p < .001), and the terms that represent horizontal differentiation were significantly more
horizontally-differentiating (M = 5.28) than vertically-differentiating (M = 3.17, t(63) = 8.30, p <
.001). Table 3 shows the terms, pre-test questions, and individual pre-test ratings of the terms.
--- Insert Table 3 about here ---
Factor analysis on the 12 vertical and 12 horizontal differentiation terms revealed that
most terms fit to their respective factors (most factor loadings were strongly significant, p < .01;
item-to-total correlations = .09 to .85 with an average of .49). We retained the terms (unorthodox,
counterculture, bohemian) that did not load strongly on their factor, because they positively
correlated with the other terms that pertained to the factor and yielded consistent pre-test ratings,
and because removing them from the analyses did not change the results. The 12 vertical and the
12 horizontal differentiation items formed reliable scales of vertical (M = 73.08, SD = 7.16, α =
0.90) and horizontal differentiation (M = 66.82, SD = 6.28, α = 0.71). Therefore, the scales of
vertical and horizontal differentiation were internally valid. We used general concepts as stimuli
instead of existing products or brands, because existing products or brands could introduce
confounds (e.g., product/brand perceptions). We replicated the results using searches for Ralph
Lauren and Urban Outfitters, which had been pre-tested and used as stimuli in Studies 1B and 4.
Method
Vertical and horizontal differentiation. As proxies for vertical and horizontal
differentiation search in each state, we examined Google searches for the selected 12 vertical and
12 horizontal differentiation terms in the period between January 2004 and November 2016. This
yielded more than 130 million searches that included the terms on desktop computers, laptops,
and mobile devices. For each term, we extracted a state-level search index from Google Insights

38
!
(www.google.com/trends/). The search index is the percentage (from 0 = no search volume to 100
= maximum search volume) of an area’s total Google searches that includes a given term relative
to other terms. The relative scaling is required to preclude regions with the highest search volume
overall from always ranking high, and it enables meaningful comparisons of indexes across
regions. This yielded 24 search indexes in each of 50 states. We averaged the 12 search indexes
of vertical differentiation and the 12 search indexes of horizontal differentiation into composite
measures of interest in vertical differentiation and interest in horizontal differentiation per state.
Political ideology. We collected two measures that have previously been used as proxies
of state political ideology (Jost et al. 2003; Khan et al. 2013; Roos and Shachar 2013): (1) the
percentage of people who self-identified as conservative in each state in the 2009-2015 Gallup
state of the states metrics (M = 38.77, SD = 5.59), and (2) the number of times that the state was
carried by a Republican candidate in the five presidential elections from 2000 to 2016 (M = 2.74,
SD = 2.24). The two measures were strongly correlated (r = .89, p < .01). We standardized and
then averaged these measures to form a single index of state conservatism. Separate analyses with
each individual ideology measure, reported in Web Appendix B, revealed consistent results.
Control measures. We controlled for the influence of each state’s population density (M =
164.48, SD = 204.54 per square mile) and median age (M = 37.40, SD = 2.32) obtained from the
U.S. Census Bureau (2010), and for each state’s gross domestic product (GDP; M = 310,864.80,
SD = 385,339.94 in millions of USD) and average income per capita in 2010-2015 (M =
43,263.99, SD = 6,555.64 in USD) obtained from the U.S. Bureau of Economic Analysis (2015).
These variables were used as proxies for state economic activity, income, and age.
Results
Table 4A reports correlations between variables. It shows that state conservatism was
negatively correlated with states’ income per capita, GDP, median age, and population density.

39
!
This is consistent with prior work showing that, compared to conservative individuals who are
more likely than liberals to reside in suburban and rural areas, liberal individuals are more likely
than conservatives to reside in urban, metropolitan areas which tend to report relatively higher
levels of economic activity, population density, infrastructure, and life expectancy (Gelman et al.
2007; Pew Research Center 2014). Key to our predictions, Table 4B summarizes the coefficient
estimates of variables predicting search for vertical and horizontal differentiation. There was no
multicollinearity: all variance inflation factors were below 3 and tolerance levels exceeded .50.
--- Insert Table 4 about here ---
Effect of ideology on online search for vertical and horizontal differentiation. We
regressed the composite measures of interest in vertical and horizontal differentiation in each
state on state conservatism, GDP, population density, income per capita, and median age. State
conservatism positively predicted search for vertical differentiation (b = .37, t = 2.47, p = .02) and
negatively predicted search for horizontal differentiation (b = -.49, t = -2.97, p < .01). Political
ideology explained a significant proportion of the variance – 6.27% for vertical differentiation
search and 11.47% for horizontal differentiation search. Similar results were found for search for
the vertically-differentiating Ralph Lauren brand (b = .33, t = 2.10, p = .04) and the horizontally-
differentiating Urban Outfitters brand (b = -.37, t = -3.22, p < .01).
There were also significant or marginally significant effects of GDP (b = .24, t = 2.05, p =
.05), population density (b = .84, t = 5.99, p < .01), and income per capita (b = -.27, t = -1.91, p =
.06) on search for vertical differentiation, and significant effects of GDP (b = .32, t = 2.50, p =
.02) and income per capita (b = -.33, t = -2.14, p = .04) on search for horizontal differentiation.
Search for vertical differentiation was more prevalent in states with greater economic activity,
population density, and lower per capita income, and search for horizontal differentiation was
more prevalent in states with greater economic activity and lower per capita income.

40
!
Robustness checks. We performed three robustness checks: (1) instrumental variable
analysis to account for potential omitted variables (Sargan 1958); (2) separate analyses with each
measure of state ideology; and (3) stimulus sampling to account for the role of the sampling of
specific search terms (Judd, Westfall, and Kenny 2012). The effect of ideology was significant
and consistent in all of these specifications. Web Appendix B details these robustness checks.
Discussion
Study 5 showed that conservatism and liberalism are linked to interest in different forms
of consumer differentiation using secondary data from all 50 U.S. states. Conservative states
searched more for terms related to vertical differentiation, and liberal states searched more for
terms related to horizontal differentiation. Conservative states also searched more for Ralph
Lauren – a vertically-differentiating brand, and liberal states searched more for Urban Outfitters –
a horizontally-differentiating brand. These results boost the external validity of our theory.
GENERAL DISCUSSION
This research shows that political ideology affects consumers’ preferences for vertical and
horizontal differentiation in the social hierarchy. Conservatism leads to vertical differentiation
and liberalism leads to horizontal differentiation in the hierarchy because conservatism endorses,
and liberalism opposes, the view that the dominance-based hierarchical structure reflects
legitimate differences in individual qualities. This effect is mitigated when vertical or horizontal
differentiation goals are primed, or when hierarchy’s legitimacy is called into question. The
results hold with established ideology measures as well as an ideology manipulation, with online
search, product evaluation, actual and hypothetical choice of well-controlled as well as real-world
product stimuli. The results also hold across different populations: students who were slightly
more liberal than conservative (4.74 on a single-item 1-to-9 ideology scale, Jost 2006), M Turk

41
!
participants who skewed liberal (4.41 on a single-item scale; 4.44 on a multi-item scale, Kidwell
et al. 2013; 40.43% were conservative, 59.57% were liberal on a dichotomous scale, Tetlock et al.
1984), and Qualtrics panelists who skewed conservative (54.96% were conservative, 45.04%
were liberal). Consistent results across settings affirm the generalizability of the findings.
Theoretical Implications
Our findings have important implications for research on political ideology, social
hierarchy, and consumer differentiation.
Our work contributes to research on political ideology and social hierarchy (Graham et al.
2009; Jost et al. 2003; Pratto and Cathey 2002) in three ways. First, it identifies the effects of
conservatism and liberalism and of their distinct hierarchy beliefs on consumers’ differentiation
strategies in the marketplace. Since the desire for distinction is an important driver of consumer
behavior (Berger and Heath 2008; Tian et al. 2001), this is an important addition to the literature.
Second, our work adds to research on ideology and hierarchy beliefs by establishing the
causal effect of ideology on hierarchy beliefs, which goes beyond correlational data used in prior
studies (Altemeyer 1998; Jost et al. 2009). We show that, since people’s political positions can be
temporarily shifted, ideological differences in hierarchical attitudes can also be bridged.
Importantly, our third contribution consists of identifying specific conditions that reduce
the ideological gap in hierarchy beliefs and consumption preferences. Specifically, our findings
show that ideological differences in preferences can be bridged if the legitimacy of the
dominance-based hierarchy is called into question (by stressing the importance of good luck vs.
hard work in attaining dominant positions). It will be important for future studies to test if this
intervention can mitigate ideological differences in other hierarchical attitudes such as support for
taxes, and whether it can impact broader outcomes such as motivation. Notably, our research
shows that manipulating hierarchy legitimacy does not change political ideology because the

42
!
causal effect extends from ideology. Still, it is possible that for some adults, political ideology is a
conscious result of hierarchy legitimacy beliefs. Future research can explore the more nuanced
effects that hierarchy legitimacy beliefs may have on ideological positions by focusing on
specific facets of political ideology. It is possible that manipulating hierarchy legitimacy changes
some people’s ideological positions along the economic dimension (which addresses economic
distribution and fairness), but not along the social dimension (which tackles social life and
values) (Everett 2013). Given the high importance of social beliefs in driving ideological
positions (Alesina and La Ferrara 2005), the insignificant effect of the hierarchy manipulation on
general ideology measures found in our research does not contradict this possibility.
Our work advances research on consumer divergence by identifying political ideology as
a new driver of consumers’ preferences between vertical and horizontal differentiation and by
establishing hierarchy beliefs as the new process behind them. This extends prior work, which
separately focused on one type of differentiation (Bellezza et al. 2013; Han et al. 2010;
Ordabayeva and Chandon 2011; White and Dahl 2006). Our findings suggest that, in consumers’
eyes, not all differentiation strategies are equal. Hence, consumers’ choices of differentiation
strategies and perceptions of other people’s differentiation attempts may vary as a function of
their political ideology. But manipulating differentiation goals or beliefs about hierarchy
legitimacy can mitigate these differences.
Finally, our findings contribute to emerging evidence on the role of political ideology in
marketing (Fernandes and Mandel 2014; Khan et al. 2013; Kidwell et al. 2013; Roos and Shachar
2013; Winterich et al. 2012), by extending the influence of ideology to consumers’ attempts to
differentiate their identity. Our work thereby suggests that identity is a useful framework for
studying the role of ideology in consumer behavior. Our findings further indicate that ideology is
a unique predictor of behavior that operates independently from SES and income. This supports

43
!
the importance of cultural, not just economic, interests in driving individuals’ political behavior,
and it adds to observations that low-SES people can be as conservative and as liberal as high-SES
people (Alesina and La Ferrara 2005). Notably, in our studies, SES did not have a consistent
effect on differentiation preferences (Table 2), nor did it consistently correlate with ideology (r
ranged between -.08 and .21). These findings are consistent with prior studies showing that
subjective status does not necessarily correlate with political ideology (Brown-Iannuzzi et al.
2014), and that SES may have conflicting – positive (Veblen 1899) or negative (Ordabayeva and
Chandon 2011) – effects on vertical differentiation. They are also consistent with prior reports
highlighting the limitations of measuring SES (e.g., Quon and McGrath 2014). Future research on
how different measures and manipulations of SES may impact the link between political ideology
and consumer differentiation will be useful.
More broadly, our work contributes to the emerging view that political ideology, with its
norms and beliefs, may constitute a type of cultural lens (Markus and Kitayama 1991) that can
explain individual behaviors. It will be interesting to study how such a political lens shapes other
aspects of individual identity and how different cultural dimensions, such as power distance (Gao,
Winterich, and Zhang 2016), interact with political ideology to influence behavior.
Practical Implications
Our findings have useful implications for marketers. First, they can inform marketers’
targeting decisions. Our results indicate that conservative consumers may be a better target
market for products that enable vertical differentiation in the hierarchy, and liberal consumers
may be a better target market for products that enable horizontal differentiation in the hierarchy.
This means that vertically-differentiating products may fare better in politically conservative
states and when advertised in conservative outlets such as The Wall Street Journal. Horizontally-
differentiating products may do better in liberal states and when advertised in liberal outlets such

44
!
as The New York Times. Provided the distinct media preferences of conservatives and liberals
(Experian Simmons 2011), our findings offer actionable opportunities to managers.
Second, our Studies 1C and 2 suggest that the same product may appeal differently to
conservatives and liberals depending on the product’s positioning. Hence, it may be profitable to
tailor the positioning of products in advertising campaigns to the political orientation of target
markets. However, when tailoring advertising campaigns to specific markets is infeasible, our
Study 2 results suggest that framing the product through tailored descriptions in companies’
public relations efforts, such as press interviews or magazine articles, may also be effective.
From the policy perspective, our findings suggest that strategies that shift individuals’
hierarchy legitimacy perceptions can bridge the divide between the opinions of conservatives and
liberals about issues informed by hierarchy beliefs. For example, it is possible that opinions on
issues such as wealth redistribution could be depolarized, at least temporarily, if public discourse
challenged individuals’ inherent views about the legitimacy of the dominance-based hierarchy.
In conclusion, the present research shows that political ideology significantly affects how
consumers differentiate from others in the marketplace. We hope that it will inspire future work
on the role of political ideology in shaping consumer identity and behavior.

45
!
REFERENCES
Abramowitz, Alan I. and Kyle L. Saunders (2008), “Is Polarization Really a Myth?” Journal of
Politics, 70 (2), 542-55.
Adler, Nancy E., Elissa S. Epel, Grace Castellazzo, and Jeannette R. Ickovics (2000),
“Relationship of Subjective and Objective Social Status with Psychological and
Physiological Functioning: Preliminary Data in Healthy, White Women,” Health
Psychology, 19 (6), 586-92.
Ailawadi, Kusum L., Koen Pauwels, and Jan-Benedict E.M. Steenkamp (2008), “Private-Label
Use and Store Loyalty,” Journal of Marketing, 72 (6), 19-30.
Alesina, Alberto and Eliana La Ferrara (2005), “Preferences for Redistribution in the Land of
Opportunities,” Journal of Public Economics, 89, 897-931.
Alford, John R., Carolyn L. Funk, and John R. Hibbing (2005), “Are Political Orientations
Genetically Transmitted?” American Political Science Review, 99 (2), 153-67.
Altemeyer, Robert A. (1998), “The Other ‘Authoritarian Personality’,” in Advances in
Experimental Social Psychology, ed. Mark P. Zanna, 47-91, New York: Academic Press.
Belk, Russel W. (1988), “Possessions and the Extended Self,” Journal of Consumer Research, 15
(September), 139-67.
Bellezza, Silvia, Francesca Gino, and Anat Keinan (2013), “The Red Sneakers Effect: Inferring
Status and Competence from Signals of Nonconformity,” Journal of Consumer Research,
41 (1), 35-54.
Berger, Jonah and Chip Heath (2008), “Who Drives Divergence? Identity-Signaling, Out- Group
Dissimilarity, and the Abandonment of Cultural Tastes,” Journal of Personality and
Social Psychology, 95 (3), 593–607.
Brewer, Marilynn B. (1991), “The Social Self: On Being the Same and Different at the Same

46
!
Time,” Personality and Social Psychology Bulletin, 17 (5), 475–82.
Brown-Iannuzzi, Jazmin L., Kristjen B. Lundberg, Aaron C. Kay, and B. Keith Payne (2014),
“Subjective Status Shapes Political Preferences,” Psychological Science, 26 (1), 15-26.
Carraro, Luciana, Luigi Castelli, and Claudia Macchiella (2011), “The Automatic Conservative:
Ideology-Based Attentional Asymmetries in the Processing of Valenced Information,”
PLoS ONE, 6 (11), 1-6.
Chan, Cindy, Jonah Berger, and Leaf Van Boven (2012), “Identifiable but not Identical:
Combining Social Identity and Uniqueness Motives in Choice,” Journal of Consumer
Research, 39 (October), 561-73.
Chow, Rosalind M. and Jeff Galak (2012), “The Effect of Inequality Frames on Redistributive
Income Policy Support,” Psychological Science, 23 (December), 1467-9.
Crawford, Jarret. T., Stephanie R. Mallinas, and Brian J. Furman (2015), “The Balanced
Ideological Antipathy Model: Explaining the Effects of Ideological Attitudes on Inter-
Group Antipathy across the Political Spectrum,” Personality and Social Psychology
Bulletin, 41, 1607-22.
Dommer, Sarah L., Vanitha Swaminathan, and Rohini Ahluwalia (2013), “Using Differentiated
Brands to Deflect Exclusion and Protect Inclusion: The Moderating Role of Self-Esteem
on Attachment to Differentiated Brands,” Journal of Consumer Research, 40 (4) 657-75.
Everett, Jim A. C. (2013), “The 12 Item Social and Economic Conservatism Scale,” PLOS One, 8
(12), e82131.
Experian Simmons (2011), Simmons Consumer Segmentations: PoliticalPersonasSM. Retrieved
from: http://www.experian.com/assets/simmons-research/white-papers/simmons-2011-
political-personas-report.pdf.
Federico, Cristopher M. and Jim Sidanius (2002), “Racism, Ideology, and Affirmative Action,

47
!
Revisited: The Antecedents and Consequences of ‘Principled Objections’ to Affirmative
Action,” Journal of Personality and Social Psychology, 82, 488-502.
Feinberg, Matthew, Alexa M. Tullett, Zachary Mensch, William Hart, and Sara Gottlieb (2017),
“The Political Reference Point: How Geography Shapes Political Identity,” PLoS ONE,
12 (2), e0171497.
Fernandes, Daniel and Naomi Mandel (2014), “Political Conservatism and Variety-Seeking,”
Journal of Consumer Psychology, 24 (1), 79-86.
Fraley, R. Chris, Brian N. Griffin, Jay Belsky, and Glenn I. Roisman (2012), “Developmental
Antecedents of Political Ideology: A Longitudinal Investigation From Birth to Age 18,”
Psychological Science, 23, 1425-31.
Gallup (2012). Gallup and USA Today/Gallup Polls 2011 Report. Retrieved from
http://www.gallup.com/poll/152021/conservatives-remain-largest-ideological-group.aspx.
Gao, Huachao, Karen Page Winterich, and Yinlong Zhang (2016), “All that Glitters Is Not Gold:
How Others’ Status Influences the Effect of Power Distance Belief on Status
Consumption,” Journal of Consumer Research, 43 (August), 265-81.
Gelman, Andrew, Boris Shor, Joseph Bafumi, and David Park (2007), “Rich State, Poor State,
Red State, Blue State: What’s the Matter with Connecticut?” Quarterly Journal of
Political Science, 2, 345-67.
Ghitza, Yair and Andrew Gelman (2014), “The Great Society, Reagan’s Revolution, and
Generations of Presidential Voting,” Columbia Working Paper.
Ginsberg, Jeremy, Matthew H. Mohebbi, Rajan S. Patel, Lynnette Brammer, Mark S. Smolinski,
and Larry Brilliant (2009), “Detecting Influenza Epidemics Using Search Engine Query
Data,” Nature, 457, 1012-4.

48
!
Goel, Sharad, Jake M. Hofman, Sébastien Lahaie, David M. Pennock, and Duncan J. Watts
(2010), “Predicting Consumer Behavior with Web Search,” Proceedings of the National
Academy of Sciences, 7 (41), 17486-490.
Gollwitzer, Peter M. (1999), “Implementation Intentions: Strong Effects of Simple Plans,”
American Psychologist, 54 (July), 493–503.
Graham, Jesse, Jonathan Haidt, and Brian A. Nosek (2009), “Liberals and Conservatives Rely on
Different Sets of Moral Foundations,” Journal of Personality and Social Psychology, 96
(5), 1029-46.
Gromet, Dena M., Kimberly A. Hartson, and David K. Sherman (2015), “The Politics of Luck:
Political Ideology and the Perceived Relationship between Luck and Success,” Journal of
Experimental Social Psychology, 59 (July), 40-6.
Han, Young Jee, Joseph C. Nunes, and Xavier Drèze (2010), “Signaling Status with Luxury
Goods: The Role of Brand Prominence,” Journal of Marketing, 74 (July), 15-30.
Hayes, Andrew (2013). Introduction to Mediation, Moderation, and Conditional Process
Analysis. New York: Guilford Press.
Iyengar, Shanto and Sean J. Westwood (2014), “Fear and Loathing across Party Lines: New
Evidence on Group Polarization,” American Journal of Political Science, 9, 690-707.
Jost, John T. (2006), “The End of the End of Ideology,” American Psychologist, 61 (7), 651-70.
Jost, John T., Christopher M. Federico, and Jaime L. Napier (2009), “Political Ideology: Its
Structure, Functions, and Elective Affinities,” Annual Review of Psychology, 60, 307-37.
Jost, John T., Jack Glaser, Arie W. Kruglanski, and Frank J. Sulloway (2003), “Political
Conservatism as a Motivated Social Cognition,” Psychological Bulletin, 129 (3), 339-75.

49
!
Judd, Charles M., Jacob Westfall, and David A. Kenny (2012), “Treating Stimuli as a Random
Factor in Social Psychology: A New and Comprehensive Solution to a Pervasive but
Largely Ignored Problem,” Journal of Personality and Social Psychology, 103 (1), 54-69.
Jung, Kiju, Ellen Garbarino, Donnel Briley, and Jesse Wynhausen (2017), “Blue and Red Voices:
Effects of Political Ideology on Consumers’ Complaining and Disputing
Behavior,” Journal of Consumer Research, 44 (3), 477-99.
Khan, Romana, Kanishka Misra, and Vishal Singh (2013), “Ideology and Brand Consumption,”
Psychological Science, 24 (3), 326-33.
Kidwell, Blair, Adam Farmer, and David M. Hardesty (2013), “Getting Liberals and
Conservatives to Go Green: Political Ideology and Congruent Appeals,” Journal of
Consumer Research, 40 (August), 350-67.
Kluegel, James R. and Eliot R. Smith (1981). Beliefs about Inequality: Americans’ Views of What
Is and What Ought To Be. Hawthorne, NY: Aldine de Gruyter.
Litman, Leib, Jonathan Robinson, and Tzvi Abberbock (2016), “TurkPrime.com: A Versatile
Crowdsourcing Data Acquisition Platform for the Behavioral Sciences,” Behavioral
Research Methods, 49, 433-42.
Lyons, Jeffrey (2017), “The Family and Partisan Socialization in Red and Blue America,”
Political Psychology, 38 (2), 297-312.
Markus, Hazel Rose and Shinobu Kitayama (1991), “Culture and the Self: Implications for
Cognition, Emotion, and Motivation,” Psychological Review, 98 (2), 224-53.
McCarty, Nolan, Keith T. Poole, and Howard Rosenthal (2006), Polarized America: The Dance
of Ideology and Unequal Riches, Cambridge, MA: MIT University Press.

50
!
Nail, Paul R., Ian McGregor, April E. Drinkwater, Garrett M. Steele, and Anthony W. Thompson
(2009), “Threat Causes Liberals To Think Like Conservatives,” Journal of Experimental
Social Psychology, 45 (4), 901-7.
Naylor Rebecca Walker, Cait Poynor Lamberon and David A. Norton (2011), “Seeing Ourselves
in Other: Reviewer Ambiguity, Egocentric Anchoring, and Persuasion,” Journal of
Marketing Research, 48 (June), 617-31.
Norton, Michael I. and Dan Ariely (2011), “Building a Better America—One Wealth Quintile at
a Time,” Perspectives on Psychological Science, 6 (1), 9-12.
Ordabayeva, Nailya and Pierre Chandon (2011), “Getting ahead of the Joneses: When Equality
Increases Conspicuous Consumption among Bottom-Tier Consumers,” Journal of
Consumer Research, 38 (June), 27-41.
Ordabayeva, Nailya and Daniel Fernandes (2017), “Similarity Focus and Support for
Redistribution,” Journal of Experimental Social Psychology, 72 (September), 67-74.
Pew Research Center (2014), Political Polarization in the American Public. Retrieved from:
http://www.people-press.org/2014/06/12/political-polarization-in-the-american-public/.
Pickett, Cynthia L., Michael D. Silver, and Marilynn B. Brewer (2002), “The Impact of
Assimilation and Differentiation Needs on Perceived Group Importance and Judgments of
Ingroup Size,” Personality and Social Psychology Bulletin, 28 (4), 546-68.
Podsakoff, Philip M., Scott B. MacKenzie, and Nathan P. Podsakoff (2012), “Sources of Method
Bias in Social Science Research and Recommendations on How To Control It,” Annual
Review of Psychology, 63, 539-69.
Pratto, Felicia and Christie Cathey (2002), “The Role of Social Ideologies in Legitimizing
Political Attitudes and Public Policy,” in The Social Psychology of Politics, eds. V. C.

51
!
Ottati, R. S. Tindale, J. Edwards, F. B. Bryant, L. Heath, D. C. O’Connell, Y. Suarez-
Balcazar, and E. J. Posavac, New York: Kluwer Academic, 135-55.
Pratto, Felicia, James Sidanius, Lisa M. Stallworth, and Bertram F. Malle (1994), “Social
Dominance Orientation: A Personality Variable Predicting Social and Political Attitudes,”
Journal of Personality and Social Psychology, 67 (4), 741-63.
Quon, Elizabeth C. and Jennifer J. McGrath (2014), “Subjective Socioeconomic Status and
Adolescent Health: A Meta-Analysis,” Health Psychology, 33 (5), 433-47.
Roos, Jason and Ron Shachar (2013), “When Kerry Met Sally: Politics and Perceptions in the
Demand for Movies,” Management Science, 60 (7), 1617-31.
Rubin, Mark and Miles Hewstone (1998), “Social Identity Theory’s Self-Esteem Hypothesis: A
Review and Some Suggestions for Clarification,” Personality and Social Psychology
Review, 2 (1), 40-62.
Sargan, John D. (1958) “The Estimation of Economic Relationships Using Instrumental
Variables,” Econometrica, 26 (3), 393-415.
Snyder, Charles R. and Harold L. Fromkin (1980), Uniqueness, New York: Plenum.
Spencer, Steven J., Mark P. Zanna, and Geoffrey T. Fong (2005), “Establishing a Causal Chain:
Why Experiments are Often More Effective than Mediational Analyses in Examining
Psychological Processes,” Journal of Personality and Social Psychology, 89 (6), 845–51.
Sproles, George B. and Elizabeth Kendall (1986), “A Methodology for Profiling Consumers’
Decision-Making Styles,” Journal of Consumer Affairs, 20 (2), 267-79.
Tajfel, Henry (1972), “Social Categorization,” in S. Moscovici (Ed.), Introduction a la
Psychologie Sociale, Vol. 1, 272-302. Paris: Larousse.
Tetlock, Philip E., Kristen A. Hannum, and Patrick M. Micheletti (1984), “Stability and Change
in the Complexity of Senatorial Debate: Testing the Cognitive versus Rhetorical Style

52
!
Hypotheses,” Journal of Personality and Social Psychology, 46 (5), 979-90.
Tian, Kelly T., William O. Bearden, and Gary Hunter (2001), “Consumers’ Need for Uniqueness:
Scale Development and Validation,” Journal of Consumer Research, 28 (June), 50-66.
U.S. Census Bureau (2010), Resident Population Data. Retrieved from:
http://www.census.gov/2010census/data/apportionment-dens-text.php.
U.S. Bureau of Economic Analysis (2015), Regional Economic Accounts. Retrieved from:
http://www.bea.gov/regional/downloadzip.cfm.
Veblen, Thorstein (1899), The Theory of the Leisure Class, New York: Macmillan.
Vernon, Philip E. and Gordon W. Allport (1931), “A Test for Personal Values,” Journal of
Abnormal and Social Psychology, 26 (3), 231-48.
Vignoles, Vivian L., Xenia Chryssochoou, and Glynis M. Breakwell (2000), “The Distinctiveness
Principle: Identity, Meaning, and the Bounds of Cultural Relativity,” Personality and
Social Psychology Review, 4 (4), 337-54.
Warren, Caleb and Margaret C. Campbell (2014), “What Makes Things Cool? How Autonomy
Influences Perceived Coolness,” Journal of Consumer Research, 41 (August), 543-63.
White, Katherine and Jennifer J. Argo (2009), “Social Identity Threat and Consumer
Preferences,” Journal of Consumer Psychology, 19 (3), 313-25.
White, Katherine and Darren W. Dahl (2006), “To Be or Not Be: The Influence of Dissociative
Reference Groups on Consumer Preferences,” Journal of Consumer Psychology, 16 (4),
404-13.
Winterich, Karen Page, Yinlong Zhang, and Vikas Mittal (2012), “How Political Identity and
Charity Positioning Increase Donations: Insights from Moral Foundations Theory,”
International Journal of Research in Marketing, 29 (4), 346-54.

53
!
TABLE 1. STUDIES 1A-4: VERTICAL AND HORIZONTAL DIFFERENTIATION STIMULI AND PRE-TEST RATINGS
Study
(Pre-test N)
Differentiation
Type
Product Stimuli
Vertical
Score
Horizontal
Score
t
p
Study 1A
(68 students)
Just Better
Coffee Mug
4.91
3.71
2.71
.010
Just Different
Coffee Mug
2.62
4.15
-5.27
< .001
Studies 1B
and 4
(82 adults)
Ralph Lauren
gift card
5.62
3.43
7.13
< .001
Urban Outfitters
gift card
3.77
5.43
-4.57
< .001
Study 1C
(78 adults)
Vertically-
differentiating
shoe line
Setia X shoes challenge the industry norms by offering athletic shoes that
are luxurious and classy. Unlike standard athletic shoes, these shoes are different by being posh,
elegant and sometimes extravagant. The shoes cost $100, and they received a rating of 94 (on a
scale from 0 and 100) from an independent product-rating magazine Consumer Reports.
5.68
3.59
5.26
< .001
Horizontally-
differentiating
shoe line
Setia Y shoes break with the industry standards by offering athletic shoes that
are edgy and unique. Unlike standard athletic shoes, these shoes are different by
being irreverent, hip and sometimes eccentric. The shoes cost $100, and they received a rating of
94 (on a scale from 0 and 100) from an independent product-rating magazine Consumer Reports
3.41
5.37
-4.83
< .001
Study 2
(91 adults)
Color red as
vertically-
differentiating
You recently came across a magazine article suggesting that red is the color of luxury in the 21st
century. The article explained that nowadays the color red is increasingly recognized as a
symbol of one’s success, prosperity, and accomplishments.
5.13
3.70
4.21
< .001
Color red as
horizontally-
differentiating
You recently came across a magazine article suggesting that red is the color of nonconformity in
the 21st century. The article explained that nowadays the color red is increasingly recognized as
a symbol of one’s originality, rebelliousness, and edge.
3.36
4.89
-4.16
< .001
Study 3
(51 adults)
Just Better
Coffee Mug
4.76
3.92
2.37
.026
Just Different
Coffee Mug
3.23
4.92
-2.81
.009
Notes:
• Each pre-test used a participant sample that was relevant for the main study and manipulated differentiation type between subjects.
• Vertical differentiation was measured from 1 = “not at all” to 7 = “very much”: “To what extent does … signal that one is different from others
by enhancing their social standing – that is, by highlighting their superiority and status among other people?”
• Horizontal differentiation was measured from 1 = “not at all” to 7 = “very much”: “To what extent does … signal that one is different from
others based on who they are as an individual – that is, the qualities, opinions, and traits that make one different from other people?”
vs.
vs.

54
!
TABLE 2. STUDIES 1A-4: REGRESSION RESULTS INCUDING DEMOGRAPHICS
(COEFFICIENT ESTIMATES AND STANDARD ERRORS)
Predictor
Study 1A
Study 1B
Study 1C
Study 2
Study 3
Study 4
IDEOLOGY
.62 (.30)**
.19 (.07)***
.71 (.30)**
.03 (.20)
.77 (.29)***
1.12 (.28)***
AGE
.54 (.32)*
.01 (.01)
.01 (.01)
-.01 (.10)
-.03 (.01)*
-.004 (.02)
SES
.57 (.32)*
.04 (.09)
-.29 (.12)**
.04 (.07)
.12 (.23)
-.23 (.24)
INCOME
---
-.03 (.4)
.09 (.05)*
.01 (.02)
.11 (.05)**
.05 (.05)
IDEOLOGY × AGE
.24 (.20)
-.01 (.01)
0 (.03)
.02 (.02)
.06 (.03)**
-.03 (.03)
IDEOLOGY × SES
.02 (.18)
.05 (.05)
.03 (.24)
.15 (.15)
-.34 (.47)
-.21 (.47)
IDEOLOGY × INCOME
---
-.03 (.02)
.01 (.10)
-.04 (.04)
-.06 (.10)
-.03 (.09)
FACTOR
---
---
---
.28 (.20)
.32 (.32)
.41 (.28)
IDEOLOGY × FACTOR
---
---
---
1.20 (.40)***
1.06 (.63)*
1.31 (.56)**
AGE × FACTOR
---
---
---
0 (.02)
0 (.03)
-.03 (.03)
SES × FACTOR
---
---
---
.19 (.15)
.40 (.52)
-.45 (.47)
INCOME × FACTOR
---
---
---
.01 (.04)
-.16 (.11)
.03 (.08)
IDEOLOGY × FACTOR
× AGE
---
---
---
0 (.03)
-.09 (.06)
-.01 (.06)
IDEOLOGY × FACTOR
× SES
---
---
---
.37 (.30)
1.27 (1.04)
-.85 (.95)
IDEOLOGY × FACTOR
× INCOME
---
---
---
-.04 (.08)
-.16 (.22)
-.05 (.19)
VERTICAL
---
---
---
---
.70 (.35)**
---
IDEOLOGY ×
VERTICAL
---
---
---
---
.42 (.70)
---
AGE × VERTICAL
---
---
---
---
-.03 (.04)
---
SES × VERTICAL
---
---
---
---
-.14 (.55)
---
INCOME × VERTICAL
---
---
---
---
-.06 (.12)
---
IDEOLOGY ×
VERTICAL × AGE
---
---
---
---
.03 (.07)
---
IDEOLOGY ×
VERTICAL × SES
---
---
---
---
-.27 (1.09)
---
IDEOLOGY ×
VERTICAL × INCOME
---
---
---
---
0 (.23)
---
CONSTANT
1.35 (.45)***
-.57 (.12)***
3.88 (.15)***
4.01 (.10)***
-.35 (.15)**
-.44 (.14)***

55
!
Notes:
• *** indicates a significant coefficient at p < .01, ** indicates p < .05, and * indicates p < .1.
• The dependent variable is the dichotomous choice between vertical and horizontal differentiation in
Studies 1A, 1B, 3, and 4. The dependent variable in Study 1C is the preference between vertical and
horizontal differentiation, measured on a scale from 1 = horizontally-differentiating product to 7 =
vertically-differentiating product, in response to questions, “which product is more appealing to you”
and “which product is more attractive to you.” The dependent variable in Study 2 is the individual
evaluation of vertical and horizontal differentiation, measured on a scale from 1 = “not at all” to 7 =
“very much” in response to questions, “how appealing is this product to you” and “how desirable is
this product to you.”
• IDEOLOGY is political ideology. It is a mean-centered single-item 9-point scale (from 1 = extremely
liberal to 9 = extremely conservative) in Study 1A. It is a mean-centered multi-item 9-point scale
(from 1 = extremely liberal to 9 = extremely conservative) in Study 1B. It is a manipulation, equal to
.5 in the conservative condition and -.5 in the liberal condition in Study 1C. It is a dichotomous
measure, equal to .5 for conservatives and -.5 for liberals in Studies 2, 3, and 4.
• AGE is the mean-centered age of the participant.
• SES is the mean-centered socioeconomic status of the participant. It is a 5-point scale measure, from 1
= “high class” to 5 = “low class” that is subsequently mean-centered around 0, in Studies 1A, 3, and 4.
It is a 10-point scale measure, from 1 = “lowest social tier” to 10 = “highest social tier” that is
subsequently mean-centered around 0, in Studies 1B, 1C, and 2.
• INCOME is the mean-centered income of the participant. It is a 15-point measure, from 1 = “under
$20,000” to 15 = “$150,000 or more” that is subsequently mean-centered around 0, in Studies 1B, 1C,
3, and 4. It is a 20-point measure, from 1 = “below $20,000” to 20 = “above $200,000” that is
subsequently mean-centered around 0, in Study 2.
• FACTOR is a manipulated factor in each study. It is the type of consumer differentiation, equal to .5
for vertical differentiation and -.5 for horizontal differentiation, in Study 2. It is the type of goal, equal
to .33 in the vertical differentiation and horizontal differentiation goal conditions and to -.66 in the
neutral goal condition, in Study 3. It is hierarchy legitimacy, equal to .5 in the legitimate hierarchy
condition and -.5 in the illegitimate hierarchy condition, in Study 4.
• VERTICAL is the additional dummy created in Study 3 to contrast the vertical and horizontal
differentiation goal conditions, equal to .5 in the vertical differentiation goal condition, -.5 in the
horizontal differentiation goal condition, and 0 in the neutral goal condition.

56
!
TABLE 3. STUDY 5: PRE-TEST RESULTS FOR GOOGLE SEARCH TERMS
Differentiation
Type
Search Terms
Vertical
Score
Horizontal
Score
t
p
Vertical
Better/best
5.12
4.12
2.37
.021
(N = 60)
Convertible
4.97
3.82
2.85
.006
Diamonds
6.32
2.59
11.91
< .001
Elegant
5.07
4.12
2.25
.028
Elite
6.03
3.24
6.51
< .001
Jewelry
5.73
3.64
6.06
< .001
Luxury
6.36
2.95
10.90
< .001
Mansion
6.33
2.87
10.60
< .001
Money
6.32
2.80
11.30
< .001
Prestige
5.95
3.10
7.26
< .001
Rich
6.47
2.57
11.79
< .001
Success
5.70
4.00
5.08
< .001
Average across
terms
5.86
3.32
10.93
< .001
Horizontal
Alternative
2.79
5.24
-6.28
< .001
(N = 64)
Antiestablishment
2.59
5.27
-7.56
< .001
Bohemian
3.05
5.30
-5.87
< .001
Counterculture
2.98
5.11
-5.25
< .001
Creative
3.25
5.49
-6.20
< .001
Distinct
4.05
5.08
-2.86
.006
Edgy
3.40
5.32
-5.26
< .001
Geek
3.05
5.22
-6.53
< .001
Hipster
3.38
5.38
5.61
< .001
Indie
3.17
5.41
-6.20
< .001
Unconventional
3.13
5.21
-5.44
< .001
Unorthodox
2.71
5.38
-7.75
< .001
Average across
terms
3.17
5.28
-8.30
< .001
Notes:
• For each differentiation type, the terms are listed in alphabetical order.
• Vertical differentiation was measured from 1 = “not at all” to 7 = “very much” using the following
item: “Vertical differentiation allows one to signal that they are different from others by enhancing
their social standing – that is, by highlighting their superiority and status among other people. Please
indicate the extent to which the concept allows vertical differentiation.”
• Horizontal differentiation was measured from 1 = “not at all” to 7 = “very much” using the following
item: “Horizontal differentiation allows one to signal that they are different from others based on who
they are as an individual – that is, by highlighting the qualities, opinions, and traits that make one
different from other people. Please indicate the extent to which the concept allows horizontal
differentiation.”
• For each term, the t-test compared the mean vertical differentiation score to the mean horizontal
differentiation score.

57
!
TABLE 4. STUDY 5: VARIABLE CORRELATIONS (A) AND REGRESSION MODEL
COEFFICIENTS (B)
A
Variable Correlations
1
2
3
4
5
6
7
1. Search for vertical
differentiation
1
2. Search for horizontal
differentiation
.35**
1
3. Income per capita
.06
.12
1
4. GDP
.29**
.46***
.22
1
5. Median age
-.02
-.03
.16
-.21
1
6. Population density
.54***
.34**
.57***
.25*
.26*
1
7. Conservatism
.34**
-.39***
-.57***
-.26*
-.39***
-.55***
1
B
Model 1 Coefficients (OLS with only the
control variables)
Model 2 Coefficients (OLS
adding conservatism)
Vertical
Search
Horizontal
Search
Vertical
Search
Horizontal
Search
Conservatism
-----
-----
.37**
-.49***
GDP
.17
.42***
.24**
.32**
Population
density
.75***
.33**
.84***
.22
Income per
capita
-.39***
-.16
-.27*
-.33**
Median age
-.11
.001
.01
-.15
R2
.39
.23
.45
.34
Notes:
• *** indicates a significant coefficient at p < .01, ** indicates p < .05, and * indicates p < .1.
• In Table 4A the correlations between conservatism and search for vertical and horizontal terms are the
partial r equivalent to the models reported in Table 4B.

58
!
FIGURE 1
THEORETICAL MODEL AND EMPIRICAL PLAN
Beliefs about Legitimacy of Dominance-Based Hierarchy
•
Process through mediation (H2): Studies 1A, 1B, 1C, 3
•Process through moderation (H4): Study 4
Note: Figure 1 summarizes the theoretical model and the empirical plan.
1. Studies 1A-C show the effect of inherent (1A, 1B) and manipulated (1C) ideology on preferences for vertical vs. horizontal differentiation and
test the process.
2. Study 2 shows the effect of ideology on individual appeal of vertical and horizontal differentiation across diverse income levels.
3. Study 3 shows the moderating effect of manipulated differentiation goal on preferences for vertical vs. horizontal differentiation. !
4. Study 4 shows the moderating effect of manipulated hierarchy legitimacy on preferences for vertical vs. horizontal differentiation.!
5. Study 5 validates the theory with secondary data of online searches for vertically- and horizontally-differentiating terms in conservative and
liberal U.S. states.
!
Political Ideology
(Conservative vs. Liberal) (H1)
•
Measured: Studies 1A, 1B, 2, 3, 4, 5
•Manipulated: Study 1C
Consumer Differentiation
(Vertical vs. Horizontal)
•Real preference: Studies 1A, 1B, 3, 4
•
Hypothetical preference: Studies 1C, 2
•
Online search: Study 5
Manipulated Differentiation Goal
(Vertical vs. Horizontal vs. Neutral) (H3)
•Study 3

59
!
FIGURE 2. STUDIES 1A-C:
HIERARCHY BELIEFS MEDIATE THE EFFECT OF MEASURED (1A, 1B) AND
MANIPULATED (1C) IDEOLOGY ON DIFFERENTIATION PREFERENCES
1A
• The predictor is measured ideology (1 = extremely liberal, 9 = extremely conservative).
• The outcome is the dichotomous choice of the vertically-differentiating mug over the horizontally-
differentiating mug (1 = vertically-differentiating mug, 0 = horizontally-differentiating mug).
• Indirect effect of ideology through hierarchy beliefs: a = .1841, SE = .0668, 95% CI = [.0724, .3343].
• Indirect effect of hierarchy beliefs through ideology: a = .1421, SE = .1080, 95% CI = [-.0486, .3808].
1B
• The predictor is measured ideology (1 = extremely liberal, 9 = extremely conservative).
• The outcome is the dichotomous choice of the vertically-differentiating brand over the horizontally-
differentiating brand (1 = vertically-differentiating brand, 0 = horizontally-differentiating brand).
• Indirect effect of ideology through hierarchy beliefs: a = .0844, SE = .0397, 95% CI = [.0109, .1741].
• Indirect effect of hierarchy beliefs through ideology: a = .0891, SE = .0619, 95% CI = [-.0331, .2113].
1C
• The predictor is manipulated ideology (-.5 = liberal condition, .5 = conservative condition).
• The outcome is preference for the athletic shoe line (1 = horizontally-differentiating line, 7 = vertically-
differentiating line).
• Indirect effect of the ideology manipulation through hierarchy beliefs: a = .1145, SE = .0719, 95% CI =
[.0114, .3108].
*** p < .01, ** p < .05, * p < .1

60
!
FIGURE 3
STUDY 3: EFFECTS OF IDEOLOGY AND MANIPULATED DIFFERENTIATON
GOAL ON CHOICE OF VERTICAL (VS. HORIZONTAL) DIFFERENTIATION
28.33%
45.45%
30.00%
62.07% 60.00%
35.48%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Neutral Vertical
differentiation
Horizontal
differentiation
Liberal
Conservative
Ideology measure
Goal condition

61
!
FIGURE 4
STUDY 4: EFFECTS OF IDEOLOGY AND MANIPULATED HIERARCHY
LEGITIMACY ON CHOICE OF VERTICAL (VS. HORIZONTAL) DIFFERENTIATION
25.68% 29.63%
63.16%
40.00%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Legitimate Illegitimate
Liberal
Conservative
Ideology measure
Hierarchy condition