Content Summary

Subjective Socioeconomic Status Causes Aggression: A Test of the Theory of Social Deprivation

Tobias Greitemeyer and Christina Sagioglou University of Innsbruck

Seven studies (overall N � 3690) addressed the relation between people’s subjective socioeconomic status (SES) and their aggression levels. Based on relative deprivation theory, we proposed that people low in subjective SES would feel at a disadvantage, which in turn would elicit aggressive responses. In 3 correlational studies, subjective SES was negatively related to trait aggression. Importantly, this relation held when controlling for measures that are related to 1 or both subjective SES and trait aggression, such as the dark tetrad and the Big Five. Four experimental studies then demonstrated that participants in a low status condition were more aggressive than were participants in a high status condition. Compared with a medium-SES condition, participants of low subjective SES were more aggressive rather than participants of high subjective SES being less aggressive. Moreover, low SES increased aggressive behavior toward targets that were the source for participants’ experience of disadvantage but also toward neutral targets. Sequential mediation analyses suggest that the experience of disadvantage underlies the effect of subjective SES on aggressive affect, whereas aggressive affect was the proximal determinant of aggressive behavior. Taken together, the present research found compre- hensive support for key predictions derived from the theory of relative deprivation of how the perception of low SES is related to the person’s judgments, emotional reactions, and actions.

Keywords: aggression, relative deprivation, social class, socioeconomic status

In most Western societies, wealth inequality is at its historic height. For example, in the United States, the richest 1% possesses more than 40% of the country’s wealth (Wolff, 2012). In Germany, the biggest economy in the European Union, the median household in the top 20% of the income class has 74 times more wealth than the bottom 20% (European Central Bank, 2013). Although there is widespread consensus among citizens that wealth inequality should be reduced (Kiatpongsan & Norton, 2014; Norton & Ari- ely, 2011), the wealth gap is actually increasing. For example, in the United States, in 2012 the top 0.1% (including 160,000 fam- ilies) owned 22% of the total household wealth, whereas in the late 1970s they owned “only” 7% (Saez & Zucman, 2014).

This wealth inequality yields differences in people’s relative social ranks that can be referred to by either social class or socioeconomic status (SES). In the present paper, we use the latter term. People of low SES have reduced access to material resources such as income, wealth, and educational attainment, whereas people of high SES have a high amount of access to these resources (Côté, 2011). Objective measures of SES are related to people’s subjective perceptions of their status (subjective SES), but only moderately so (Adler, Epel, Castel- lazzo, & Ickovics, 2000; Johnson & Krueger, 2006; Kraus, Piff, & Keltner, 2009). There has been abundant evidence documenting the link between people’s SES and their well-being and physical health,

with people of low SES being more likely to experience poor self- rated health and negative affect (Adler et al., 2000), cardiovascular diseases (Gallo & Matthews, 2003), and vulnerability to a flu virus (Cohen et al., 2008). Importantly, people’s subjective SES appears to be more important for their well-being than absolute levels of SES (Adler et al., 2000). Even after accounting for people’s objective social standing, higher subjective SES fosters well-being and health, whereas the opposite is true for low perceived SES. Overall, perceiv- ing oneself as having low SES has negative intrapersonal effects.

Recent research has begun to address the interpersonal effects of subjective SES. Interestingly, it appears that having high subjective SES has negative consequences for others. For example, individuals high in subjective SES are more likely to report psychological enti- tlement and narcissistic personality tendencies (Piff, 2014). Further- more, individuals high compared to low in subjective SES behave more unethically: They are more likely to break the law while driving, lie in negotiations, cheat to win a prize, and endorse unethical work behavior (Piff et al., 2012). Given that both narcissism and unethical behavior are related to aggressive inclinations (Buckley, Wiese, & Harvey, 1998; Bushman & Baumeister, 1998), individuals low in subjective SES could be less aggressive than those high in subjective SES. Nevertheless, relative deprivation theory (Smith, Pettigrew, Pip- pin, & Bialosiewicz, 2012; Stouffer, Suchman, DeVinney, Star, & Williams, 1949) predicts quite the opposite, that low subjective SES causes an increase in aggression. In the next section, we will elaborate on this prediction.

Theoretical Perspectives

Individuals often compare themselves with others, with these comparisons having important affective consequences. For exam-

This article was published Online First June 6, 2016. Tobias Greitemeyer and Christina Sagioglou, Institute for Psychology,

University of Innsbruck. Correspondence concerning this article should be addressed to Tobias

Greitemeyer, Institut für Psychologie, Universität Innsbruck, Innrain 52, 6020 Innsbruck, Austria. E-mail: [email protected]

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Journal of Personality and Social Psychology © 2016 American Psychological Association 2016, Vol. 111, No. 2, 178–194 0022-3514/16/$12.00 http://dx.doi.org/10.1037/pspi0000058

178

ple, according to social comparison theory (Festinger, 1954), in- dividuals have an innate drive to compare themselves with objec- tive standards but also with similar others to learn where they stand. These comparisons, in turn, can have an impact on the person’s subjective well-being. Wills (1981), among others, has theorized that comparisons with others who are worse off than oneself (downward comparison) increase well-being, whereas ex- posure to others who are more fortunate than or superior to oneself (upward comparison) has the opposite effect. In particular, if individuals come to the conclusion that their referent group is better off and that this outcome is undeserved, then they likely experience feelings of resentment and dissatisfaction (Crosby, 1976; Walker & Smith, 2002).

As first documented by Stouffer and colleagues (Stouffer et al., 1949), comparisons with those who are better off affect people’s judgments and feelings. The belief that one’s standing is relatively disadvantaged and that the disadvantage is undeserved may lead to the experience of personal relative deprivation, and this judgment invokes feelings of anger and resentment (Pettigrew, 2015; Smith & Huo, 2014). Importantly, subjective more than objective cir- cumstances shape people’s judgments (Smith & Pettigrew, 2014). As noted above, people’s perceptions of their social status are more strongly associated with their physical health than is their objective social status (Adler et al., 2000). Corroborating this finding, a recent meta-analysis (Smith et al., 2012) found that relative deprivation has a much stronger impact on people’s atti- tudes and behavior than do objective measures of deprivation.

Smith and colleagues (2012) outlined a model that describes the relative deprivation experience. An individual’s objective position in a social hierarchy evokes a comparison process, with the expe- rience of disadvantage stemming from an interpersonal compari- son between the individual and other persons that are better off. As a consequence, the individual responds with anger and resentment to the undeserved disadvantage. These emotional reactions, in turn, could evoke aggressive or violent activities. That is, this model proposes a direct link between the experience of disadvantage and hostile emotional reactions, whereas the experience of disadvan- tage is indirectly related (via hostile emotions) to aggressive be- havior.

In the present research, we provide a comprehensive test of how subjective SES, via the experience of relative deprivation, affects aggressive affect and behavior. Based on Smith and colleagues’ model, we anticipated to find a negative relation between subjec- tive SES and aggression. Moreover, we examined the underlying mechanisms. We expected that the experience of disadvantage would mediate the impact of subjective SES on aggressive affect, but would not directly mediate the impact of subjective SES on aggressive behavior. Instead, we proposed that aggressive affect would serve as the proximate mechanism evoking aggressive behavior. In sum, we anticipated a sequential process from sub- jective SES to the experience of disadvantage to aggressive affect to aggressive behavior.

Previous Empirical Evidence for a Negative Subjective SES–Aggression Link

There has been some sociological research addressing the link between relative deprivation and violent crime. For example, Ken- nedy, Kawachi, Prothrow-Stith, Lochner, and Gupta (1998) exam-

ined the relationship between state-level income inequality and homicide rates and found a strong positive correlation. This rela- tionship held when controlling for the impact of household pov- erty. Because relative deprivation was a stronger predictor of homicide rates than household poverty rates, the authors argued that violent crimes can often be explained by resource deprivation that leads to an individual’s personal frustration and hostility. In fact, feelings of personal relative deprivation have been shown to be associated with delinquency (Agnew, 2001; Crosby, 1976). Overall, it appears that relative inequality is a better predictor of violent crimes than absolute poverty.

Social-psychological evidence for the hypothesis that subjective SES is negatively related to aggression comes from research by Kraus and colleagues (Kraus, Horberg, Goetz, & Keltner, 2011). They found that lower-class individuals respond with greater hos- tile reactions to threat. For example, in one study, after perceptions of social class rank were manipulated, participants were presented with ambiguous story stems to be completed. Results revealed that lower-class-rank participants expected more aggressive responses from the characters in the stories than did upper-class-rank partic- ipants.

Finally, given that people’s perception of their subjective SES is related to objective measures of SES (Adler et al., 2000), the relation between objective SES and aggression may give some indication of the link between subjective SES and aggression. It is well-established that individuals with lower SES based on objec- tive components (having less money, less advanced education, and holding less prestigious employment than others) are exposed to more aggression and violence (for a review, Evans, 2004). Fur- thermore, a number of cross-sectional studies have examined the association between objective SES and hostility and found an inverse relationship between education, income, and hostile incli- nations (for a review, Gallo & Matthews, 2003). For example, low- compared with middle-income mothers were less likely to control hostile than nonhostile emotions in response to hypothetical stories of their child’s anger (Martini, Root, & Jenkins, 2004). In sum, a number of empirical findings directly and indirectly support the predictions based on the theory of relative deprivation that sub- jective SES is negatively related to aggression.

The Present Research

The present research was designed to provide an extensive test of hypotheses deduced from the theory of relative deprivation. Specifically, we addressed in what way subjective SES is related to aggression. To examine this issue, seven studies were carried out. Studies 1a, 1b, and 2 employed a cross-sectional design. Whereas the correlational design allowed for falsification of our hypothesis that low subjective SES causes an increase in aggres- sion (when the relation is either nonsignificant or positive), it does not address the direction of the effect. Hence, Studies 3 through 6 employed an experimental design where we manipulated partici- pants’ subjective SES. Studies 1a, 1b, and 2 examined the relation between subjective SES and trait hostility. Studies 3 through 6 investigated whether subjective SES would cause an increase in aggressive affect, whereas Studies 4 through 6 also assessed ag- gressive behavior. In Studies 2 through 6, participants’ experience of disadvantage was assessed, and it was examined whether this judgment would directly mediate the impact of subjective SES on

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

179SUBJECTIVE SES AND AGGRESSION

aggressive affect and indirectly mediate the impact of subjective SES on aggressive behavior (via aggressive affect). Study 4 further examined whether any effects of differences in SES on aggression would be driven by individuals low and/or high in subjective SES. Finally, in Studies 5 and 6, we tested whether low subjective SES would cause aggressive behavior only toward targets that are the source of their experience of disadvantage or whether neutral targets are also at risk of becoming a victim.

By testing these hypotheses, the present research expands pre- vious research in at least four important ways. First, whereas previous research has examined merely correlational associations, we provide a test of the causal relation between subjective SES and aggression. Second, our working hypothesis is that low subjective SES increases aggression, but we also examine whether, compared with a medium subjective SES condition, higher subjective SES will decrease aggression, lower subjective SES will increase ag- gression, or both. Third, based on relative deprivation theory (Smith et al., 2012), a comprehensive test of the sequential process from the experience of disadvantage to increased aggressive affect to increased aggressive behavior is provided. Fourth, we aim to elucidate whether low subjective SES people are only aggressive when their action has the potential to redress their deprivation or whether low subjective SES increases aggression even toward neutral parties. Overall, the present research thereby provides a comprehensive test of how subjective SES influences the experi- ence of relative deprivation and, in turn, aggression and related affect.

In the correlational studies, we also examined the association between an objective measure of SES (namely, education) and trait aggression and further tested whether subjective or objective SES was more strongly associated with a person’s level of aggression. In the experimental studies (with the exception of Study 3), we also examined whether the status manipulation depended on the participant’s actual SES (operationalized by participant’s educa- tional level and income; income was not assessed in Studies 1a through 3). Inasmuch as previous research has shown that subjec- tive rather than objective SES has a greater impact on one’s well-being (e.g., Adler et al., 2000), we expected that aggression would be more likely to be associated with participants’ subjective perceptions of their status compared to the actual status.

We employed two complimentary measures of behavioral ag- gression. Aggression is generally defined as intentional behavior to harm another person. This behavior has been categorized to take verbal or physical form (Bushman & Anderson, 1998). Verbal aggression can be measured via evaluations of another person that may cause harm. For example, participants get to rate an ostensible experimenter who is currently employed by the university. They are asked to judge his or her capability, likability, and to give a recommendation about extending the person’s contract. Subse- quently, their rating is sealed in an envelope and ostensibly for- warded to the responsible department. To damage another person’s chances of keeping a desirable job by giving a negative evaluation can be considered just as meaningful as physical harm (Twenge, Baumeister, Tice, & Stucke, 2001). Previous research (e.g., Bush- man, Bonacci, Pedersen, Vasquez, & Miller, 2005; Greitemeyer & McLatchie, 2011; Ohbuchi, Kameda, & Agarie, 1989; Sagioglou & Greitemeyer, 2014; Twenge et al., 2001) has relied on similar job-relevant evaluations to measure aggressive behavior (for re-

views, Baron & Richardson, 1994; Bushman & Anderson, 1998). We employed such a measure in our Studies 4 and 5.

The second measure we employed is based on the laws of sympathetic magic (Rozin, Millman, & Nemeroff, 1986). Rozin and colleagues showed that people tend to attribute magical prop- erties to certain objects. Based on the law of similarity stating that the image equals the object, DeWall and colleagues (2013) devel- oped a measure of physical aggression. In the online version of the study, they present an image of a doll to participants and tell them that this doll represents a certain person. The number of pins people stick into the doll (in online settings by typing in a number) constitutes the measure of aggressive behavior. In a series of studies, DeWall and colleagues found this measure to be reliable and valid as a measure of aggressive behavior. To rule out alter- native explanations of our findings and to vary our measure of aggression, we used the Voodoo Doll Task in Study 6.

In all studies, participants were citizens of the U.S. who took part on Amazon Mechanical Turk (MTurk) in exchange for a small payment, in accordance with standard online payment norms. Especially with regard to our SES-related hypothesis, we sought to recruit a sample diverse in educational background and other status-related characteristics. Thus, MTurk participants deemed the appropriate sample, as they are substantially more representa- tive of the U.S. population than are university samples (Buhrm- ester, Kwang, & Gosling, 2011; Paolacci & Chandler, 2014). All participants were run before any analyses were performed, and all data exclusions, experimental conditions, and variables assessed are reported. For debriefing, in all studies, participants were given an e-mail address where they could contact the present authors to get more information about the study.

Study 1a

In a cross-sectional correlational study, participants were asked to indicate their subjective SES as well as to respond to a stan- dardized scale measuring trait aggression. Moreover, participants responded to measures that are potentially related to both subjec- tive SES and trait aggression and hence could account for the link between subjective SES and trait aggression. Concretely, we em- ployed measures of the dark tetrad and the Big Five. The dark tetrad of personality consists of the dimensions narcissism, Ma- chiavellianism, and psychopathy (known as the Dark Triad, see Furnham, Richards, & Paulhus, 2013), as well as everyday sadism (Buckels, Jones, & Paulhus, 2013; Chabrol, Van Leeuwen, Rod- gers, & Séjourné, 2009). The dark tetrad dimensions share a common core but are conceptually distinct (Buckels et al., 2013; Furnham et al., 2013). Narcissists are defined by a grandiose sense of self-importance and a sense of superiority. Individuals high in Machiavellianism tend to manipulate and exploit others. Psychop- athy can be characterized by callous and unemotional behavior, and everyday sadists obtain pleasure from harming other people. Recent research established that narcissism is a direct positive correlate of social class (Piff, 2014) and that the dark tetrad dimensions are positively associated with trait aggression (Greit- emeyer, 2015). Therefore, we examined whether the relation be- tween subjective SES and trait aggression (if there is any) would remain significant when controlling for the influence of the dark tetrad. The fundamental personality dispositions of the Big Five consist of conscientiousness, neuroticism, openness, agreeable-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

180 GREITEMEYER AND SAGIOGLOU

ness, and extraversion. Agreeableness, in particular, has been shown to be associated with trait aggression, and thus we also controlled for the possible influence of the Big Five (as well as participant sex and age).

Method

The sample included 510 individuals. One participant provided implausible answers (i.e., his reported age was 452 years). This participant was excluded from the following analyses, leaving a final sample of 509 individuals (218 females, 291 males; mean age � 31.6 years, SD � 9.7). As noted above, MTurk participants are more representative of the U.S. population than are university samples, which is reflected in participants’ diverse educational backgrounds. Three participants reported to have less than a high school degree, 107 participants completed high school, 151 par- ticipants completed some college, 188 participants obtained a bachelor’s degree, 53 participants had a Master’s degree, and 8 participants had a Ph.D. degree.

At the onset, participants learned that they would participate in two separate studies, one about food preferences and one about personality. In the present context, the focus is on the personality part. First, trait aggression was assessed, using the short version of the Buss and Perry aggression questionnaire (Bryant & Smith, 2001). The scale comprises 12 items. Sample items are: “Given enough provocation, I may hit another person” and “I have threat- ened people I know.” The items were assessed on a scale from 1 to 5 (� � .88). To measure narcissism, Machiavellianism, and psychopathy, a 12-item measure of the Dark Triad, with four items per subscale, was used (Jonason & Webster, 2010). Sample items are: “I tend to want others to pay attention to me“ (narcissism, � � .87), “I have used deceit or lied to get my way“ (Machiavellianism, � � .86), and “I tend to lack remorse“ (psychopathy, � � .80). These items were assessed on a scale from 1 to 9. Sadism was assessed using the expanded version of the Comprehensive As- sessment of Sadistic Tendencies (Buckels et al., 2013). The scale comprises 18 items. Sample items are: “I enjoy physically hurting people“ and „When making fun of someone, it is especially amus- ing if they realize what I’m doing.” The items were assessed on a scale from 1 to 5 (� � .88). To measure the Big Five, a 10-item measure (Gosling, Rentfrow, & Swann, 2003) was employed. There are two items per dimension that were assessed on a scale from 1 to 7. Note that some scale reliabilities were relatively poor (Conscientiousness: � � .62, Neuroticism: � � .76, Openness: � � .49, Agreeableness: � � .50, Extraversion: � � .78), which is a typical psychometric cost of using short measures (cf. Gosling et al., 2003). For all scales, items were pooled, using the average.

After responding to some filler items, participants’ subjective SES was assessed using two different measures. First, participants com- pleted the MacArthur Ladder (Adler et al., 2000; Goodman et al., 2001). They were shown a picture of a 10-rung ladder and asked to imagine that this ladder represents where people stand in the United States. They were told that at the top of the ladder are the people who are the best off—those who have the most money, the most edu- cation, and the most respected jobs, whereas at the bottom are the people who are the worst off—who have the least money, the least education, and the least respected jobs or no jobs. Participants were asked to indicate where they think they stand at this time of their life relative to other people in the United States (M � 4.72,

SD � 1.64). Second, participants were asked to indicate their agreement with statements about their material resources. This scale was adapted from Griskevicius and colleagues (Griskevicius, Delton, Robertson, & Tybur, 2011). On a scale from 1 to 7, participants indicated their agreement with the following state- ments: “I have enough money to buy things I want,” “I don’t need to worry too much about paying my bills,” “My family usually had enough money for things when I was growing up,“ “I grew up in a relatively wealthy neighborhood,” and “I felt relatively wealthy compared to the other kids in my school.” The scale was reliable (� � .79) so items were collapsed using the mean (M � 3.42, SD � 1.34). For the sake of brevity and because both measures were significantly correlated, r � .48, these measures were stan- dardized and averaged into one overall measure of subjective SES. For all correlational studies, the pattern of findings was very similar when using the separate measures. For example, the cor- relation between the ladder measure and trait aggression was: r � �.13, p � .01; the correlation between the scale measure and trait aggression was: r � �.14, p � .01. The full analyses can be obtained from the authors upon request.

Finally, participants were thanked and asked what they thought this experiment was trying to study, but none of the participants noted the link between subjective SES and aggression.

Results and Discussion

Descriptive statistics and intercorrelations of all measures are shown in Table 1. Most importantly, subjective SES was nega- tively related to trait aggression. Moreover, subjective SES was positively related to narcissism, extraversion, and conscientious- ness, and negatively related to neuroticism.

To examine whether the measure of subjective SES is associated with trait aggression when controlling for the impact of the dark tetrad, the Big Five, participant age, and participant sex (coded 1 � male, 2 � female), a multiple regression was performed on the data. The overall regression was significant, F(12, 497) � 53.45, p � .001. Subjective SES was still significantly negatively related to trait aggression, � � �.07, p � .042. Moreover, Machiavellianism, � � .16, p � .001, Psychopathy, � � .10, p � .027, sadism, � � .14, p � .001, Agreeableness, � � �.26, p � .001, Neuroticism, � � .40, p � .001, and participant sex, � � �.09, p � .006, also significantly predicted trait aggression. In contrast, narcissism, � � .04, p � .334, extraversion, � � .05, p � .181, conscien- tiousness, � � .03, p � .361, openness, � � �.02, p � .469, and participant age, � � .03, p � .389, did not.

Education, as an objective measure of participant’s SES, was positively related to subjective SES, r(509) � .20, p � .001, and negatively related to trait aggression, r(509) � �.10, p � .032. In a multiple regression, when subjective SES and education are used to predict trait aggression, subjective SES was a significant pre- dictor, � � �.14, p � .003, whereas education was not, � � �.07, p � .130.

Study 1a suggests that there is negative link between subjective SES and trait aggression. Moreover, although trait aggression and narcissism were negatively related, we replicated previous re- search (Piff, 2014) in that higher subjective SES was related to increased narcissism. Hence, it appears that subjective SES has both positive (narcissism) and negative (trait aggression) links to antisocial forms of personality. It is important to notice that the

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

181SUBJECTIVE SES AND AGGRESSION

relation between subjective SES and trait aggression held when controlling for narcissism (and the other measures of the dark tetrad, as well as the Big Five and demographics). As in previous research (Adler et al., 2000; Johnson & Krueger, 2006; Kraus et al., 2009), subjective SES was moderately positively associated with an objective measure of SES (education). Just like subjective SES, objective SES was negatively related to trait aggression. Importantly, however, a multiple regression showed that subjec- tive SES more than objective SES predicted trait aggression (see Adler et al., 2000). Overall, it appears that subjective SES has a robust association with a person’s level of trait aggression.

Study 1b

Study 1b provides an almost exact replication of Study 1a. The main variables were assessed in the same way, but some filler items were different. Study 1b employed a data set that has been used in previous work (Sagioglou & Greitemeyer, 2016). The original study focused on the relation between food preferences and antisocial personality, whereas the present purpose is to ex- amine the relation between subjective SES and aggression.

Method

The sample included 508 individuals. One participant gave implausible answers (i.e., her reported standing on the social

ladder was 50, whereas the scale from 1 to 10). This participant was excluded from the following analyses, leaving a final sample of 507 individuals (237 females, 270 males; mean age � 36.8 years, SD � 11.3). One hundred thirty-two participants completed high school, 130 participants completed some college, 178 partic- ipants obtained a bachelor’s degree, 57 participants had a Master’s degree, and 10 participants had a Ph.D. degree.

The same questionnaires were employed as in Study 1a (trait aggression: � � .90; Narcissism: � � .89; Machiavellianism: � � .87; Psychopathy: � � .83; Sadism: � � .89; Conscientiousness: � � .60; Neuroticism: � � .73; Openness: � � .48; Agreeable- ness: � � .51; Extraversion: � � .76). The correlation between the MacArthur Ladder (M � 3.48, SD � 1.32) and the social class scale (� � .78; M � 4.88, SD � 1.73) was r � .48.

Results and Discussion

Descriptive statistics and intercorrelations of all measures are shown in Table 2. As in Study 1a, subjective SES was negatively related to trait aggression. Subjective SES was again positively related to narcissism, extraversion, and conscientiousness, and negatively related to neuroticism. In contrast to Study 1a, there was a reliable negative link between subjective SES and psychop- athy.

The same multiple regression as in Study 1a was performed next. The overall regression was significant, F(12, 494) � 60.05,

Table 1 Means, Standard Deviations, and Bivariate Correlations (Study 1a)

Variable M SD 1 2 3 4 5 6 7 8 9 10

1. Subjective SES .00 .78 2. Trait aggression 2.15 .74 �.15��

3. Narcissism 3.85 1.92 .09� .30���

4. Machiavellianism 3.27 1.79 .05 .47��� .49���

5. Psychopathy 3.02 1.70 �.02 .55��� .27��� .58���

6. Sadism 1.76 .57 .08 .44��� .29��� .49��� .54���

7. Extraversion 3.14 1.59 .14�� �.09� .20��� .05 �.16� .04 8. Agreeableness 5.07 1.27 .08 �.55��� �.16��� �.32��� �.59��� �.39��� .10�

9. Conscientiousness 5.13 1.27 .20��� �.34��� �.19��� �.27��� �.33��� �.21��� .14� .31���

10. Neuroticism 3.36 1.50 �.24��� .51��� .18��� .18��� .23��� .05 �.27��� �.27��� �.42���

11. Openness 5.03 1.22 �.05 �.16��� �.02 �.10� �.17��� �.21��� .23��� .20��� .15�� �.06

Note. SES � socioeconomic status. � p � .05. �� p � .01. ��� p � .001.

Table 2 Means, Standard Deviations, and Bivariate Correlations (Study 1b)

Variable M SD 1 2 3 4 5 6 7 8 9 10

1. Subjective SES .00 .75 2. Trait aggression 2.05 .80 �.21���

3. Narcissism 3.60 2.01 .10� .39���

4. Machiavellianism 3.11 1.83 �.01 .51��� .57���

5. Psychopathy 2.75 1.73 �.15�� .54��� .36��� .64���

6. Sadism 1.73 .60 �.05 .59��� .40��� .55��� .56���

7. Extraversion 3.53 1.66 .21��� �.06 .09� .13�� �.08 .10�

8. Agreeableness 5.18 1.32 .10� �.57��� �.27��� �.39��� �.61��� �.43��� .04 9. Conscientiousness 5.49 1.22 .19��� �.35��� �.22��� �.25��� �.27��� �.27��� .09� .34���

10. Neuroticism 3.01 1.46 �.19��� .47��� .18��� .18��� .23��� .12�� �.27��� �.36��� �.41���

11. Openness 4.97 1.27 .06 �.11� .05 �.08 �.18��� �.12�� .27��� .15�� .05 �.19���

Note. SES � socioeconomic status. � p � .05. �� p � .01. ��� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

182 GREITEMEYER AND SAGIOGLOU

p � .001. Subjective SES remained a significant predictor of trait aggression, � � �.12, p � .001. Moreover, Machiavellianism, � � .13, p � .005, Sadism, � � .36, p � .001, Agreeableness, � � �.23, p � .001, and Neuroticism, � � .30, p � .001, significantly predicted trait aggression. In contrast, Narcissism, � � .07, p � .064, Psychopathy, � � .02, p � .728, Extraversion, � � �.02, p � .615, Conscientiousness, � � .03, p � .451, Openness, � � .04, p � .210, participant age, � � �.01, p � .701, and participant sex, � � .01, p � .768, did not.

Education was positively related to subjective SES, r(507) � .31, p � .001. In contrast to Study 1a, the correlation between education and trait aggression was not significant, r(507) � �.07, p � .142. In a multiple regression, subjective SES significantly predicted trait aggression, � � �.21, p � .001, whereas education did not, � � �.00, p � .970.

Replicating Study 1a, Study 1b found a negative link between subjective SES and trait aggression. As in Study 1a, subjective SES was positively related to increased narcissism (cf. Piff, 2014). In contrast to Study 1a, subjective SES and psychopathy were negatively related. Importantly, the relation between subjective SES and trait aggression remained significant when controlling for narcissism, psychopathy as well as a number of other measures related to the constructs of interest. Taken together, Studies 1a and 1b provided correlational evidence for the hypothesis that subjec- tive SES is negatively associated with aggression. Moreover, both studies converge in that this relation cannot be accounted for by measures that are related to one or both subjective SES and trait aggression.

Study 2

Study 2 provides a further replication of Studies 1a and 1b. More importantly, we aimed to elucidate a possible underlying mechanism of the relation between subjective SES and trait ag- gression. As noted in the Introduction, we reasoned that low subjective SES should be associated with perceptions of disadvan- tage, which in turn may have an impact on participants’ level of aggression. Hence, we also assessed participants’ perceptions of disadvantage. We anticipated that perceptions of disadvantage would mediate the effect of subjective SES on trait aggression.

Method

The sample included 788 individuals. Three participants pro- vided implausible answers (i.e., their reported age was between 287 and 10543 years). These participants were excluded from the following analyses. We verified attentive participation with an item attention check (cf. Oppenheimer, Meyvis, & Davidenko, 2009), which was placed among the dependent measures: “Please leave this item blank (don’t select an answer), so we know you are reading the questionnaire properly.” Twenty-one individuals failed this item manipulation check, which further reduced the final sample to 764 individuals (390 females, 374 males; mean age � 32.6 years, SD � 10.9). Sixty-two participants completed high school, 308 participants completed some college, 298 participants obtained a bachelor’s degree, 78 participants had a Master’s de- gree, and 18 participants had a Ph.D. degree.

Participants’ perception of disadvantage was assessed with four items, created by us for the present purposes. However, as one item (“How much do you feel like you deserve your rank in society” (recoded) notably reduced scale reliability (� � .57), we dropped this item from the scale and did not include it in the subsequent studies. The remaining items (“How satisfied are you with your rank in society (recoded),” “To what extent do you feel like a winner compared to other Americans (recoded),” and “To what extent do you feel worse off than other Americans”) were com- bined using the average (� � .82). The remaining questionnaire was the same as in Studies 1a and 1b (although different filler items were employed). Reliabilities were as follows: trait aggres- sion, � � .88; Narcissism, � � .87; Machiavellianism, � � .85; Psychopathy, � � .81; Sadism, � � .87; Conscientiousness, � � .68; Neuroticism, � � .79; Openness, � � .54; Agreeableness, � � .55; Extraversion, � � .76. The correlation between the MacArthur Ladder (M � 4.90, SD � 1.60) and the social class scale (� � .76; M � 3.67, SD � 1.26) was r � .51.

Results and Discussion

Descriptive statistics and intercorrelations of all measures are shown in Table 3. Most importantly, subjective SES was again negatively related to trait aggression. Subjective SES was posi-

Table 3 Means, Standard Deviations, and Bivariate Correlations (Study 2)

Variable M SD 1 2 3 4 5 6 7 8 9 10 11

1. Subjective SES .00 .87 2. Perceived

disadvantage 3.83 1.31 �.69���

3. Trait aggression 2.04 .74 �.21��� .29���

4. Narcissism 3.82 1.96 .06 .06 .34���

5. Machiavellianism 3.10 1.74 .09�� �.00 .46��� .53���

6. Psychopathy 2.61 1.66 �.01 .06 .51��� .35��� .65���

7. Sadism 1.92 .55 .06 �.00 .51��� .37��� .59��� .56���

8. Extraversion 3.48 1.65 .16��� �.19��� �.08� .10�� .05 �.11�� .01 9. Agreeableness 5.21 1.28 .04 �.07 �.54��� �.24��� �.41��� �.60��� �.45��� .08�

10. Conscientiousness 5.33 1.33 .15��� �.23��� �.29��� �.17��� �.19��� �.23��� �.19��� .19��� .23���

11. Neuroticism 3.21 1.53 �.17��� .25��� .51��� .21��� .17��� .21��� .08� �.17��� �.35��� �.39���

12. Openness 5.11 1.25 .04 �.05 �.06 �.02 �.03 �.09� �.10�� .27��� .14��� .06 �.12��

Note. SES � socioeconomic status. � p � .05. �� p � .01. ��� p � .001.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

183SUBJECTIVE SES AND AGGRESSION

tively related to Machiavellianism, extraversion, and conscien- tiousness, and negatively related to neuroticism.

The same multiple regression as in Studies 1a and 1b was performed next. The overall regression was significant, F(12, 751) � 82.15, p � .001. Subjective SES remained a significant predictor of trait aggression, � � �.17, p � .001. Machiavellian- ism, � � .10, p � .009, Psychopathy, � � .09, p � .018, Sadism, � � .28, p � .001, Agreeableness, � � �.18, p � .001, Neurot- icism, � � .36, p � .001, and Openness, � � .05, p � .049, also significantly predicted trait aggression. In contrast, Narcissism, � � .04, p � .207, Extraversion, � � .01, p � .741, Conscien- tiousness, � � .02, p � .455, participant age, � � �.05, p � .050, and participant sex, � � �.01, p � .797, did not.

Perception of disadvantage was negatively related to subjective SES. To test whether perception of disadvantage mediates the effect of subjective SES on trait aggression, a bootstrapping anal- ysis based on 5,000 bootstraps was performed (Preacher & Hayes, 2004). This analysis showed that the confidence interval for the indirect effect did not include 0 (95% CI � �0.23, �0.10). The reverse mediation whether subjective SES mediates the relation between perception of disadvantage and trait aggression was not reliable (95% CI � �0.04, 0.05).

Education was positively related to subjective SES, r(764) � .23, p � .001, and was negatively related to trait aggression, r(764) � �.09, p � .016. In a multiple regression, subjective SES significantly predicted trait aggression, � � �.20, p � .001, whereas education did not, � � �.04, p � .251.

Study 2 further showed that subjective SES is negatively related to trait aggression. As in Studies 1a and 1b, this relation held when controlling for a host of other variables. Most importantly, Study 2 provided some initial evidence about the underlying mechanism. It appears that perceptions of disadvantage constitute a mediating path from differences in subjective SES to trait aggression. As a whole, the correlational studies suggest quite consistently that subjective SES is negatively related to a person’s level of aggres- sion. However, because of the correlational nature of these find- ings, no causal interpretations are appropriate. It may be that subjective SES increases aggression, but it is also conceivable that aggression has an impact on an individual’s subjective SES and/or that some third variable leads to both. To address this issue, Studies 3 through 6 employed an experimental design. Because age and participant gender had no significant effect in our corre- lational studies (and because we had no a priori predictions), we abstained from including these variables in the following analyses.

Study 3

Study 3 provides a first test of the causal hypothesis that differences in subjective SES lead to aggression (aggressive af- fect). Kraus and colleagues (e.g., C. Anderson, Kraus, Galinsky, & Keltner, 2012; Kraus, Côté, & Keltner, 2010) manipulated partic- ipants’ temporary perceptions of their status by employing a prim- ing technique where participants are asked to compare themselves with people of very high SES or very low SES. We adapted this procedure, but made some changes (see below). Participants’ ag- gressive affect was assessed via a well-established state hostility scale. Moreover, we tested whether perceptions of disadvantage would account for the effect of subjective SES on aggressive affect. In sum, we expected that low (relative to high) SES would

lead to increased aggressive affect. In addition, low (relative to high) SES should evoke the experience of disadvantage and this effect may account for low SES participants’ increased aggressive affect.

Method

Participants were 502 individuals (266 females, 235 males, one indicated other; mean age � 34.7 years, SD � 11.8). All partici- pants were shown three photos. In the high-status condition (n � 246), the photos showed an expensive sports car, a luxury house with an adjacent swimming pool, and a gourmet diner. In the low-status condition (n � 256), the photos showed a low-budget car, a prefab estate, and a hot dog stand. Participants were asked to think about the people who live a life such as the one presented in the photos and to briefly write about what a life like this would look like. As a manipulation check, participants were asked how they would rate their socioeconomic status compared to these people (see C. Anderson et al., 2012; Kraus et al., 2010). The scale ranged from 1 to 10. Afterward, participants responded to three items measuring perceptions of disadvantage (� � .77): “How satisfied are you with your socioeconomic status?” (recoded), “To what extent do you feel like a winner compared to others who matched your profile?” (recoded), and “To what extent do you feel worse off than others who matched your profile?”. Finally, partic- ipants completed the State Hostility Scale (C. A. Anderson, Deuser, & DeNeve, 1995). This scale consists of 35 mood state- ments (e.g., “I feel enraged” and “I feel angry”), and participants are asked to indicate the extent to which they agree or disagree, assessed on a 5-point Likert-type scale (1 � strongly disagree, 5 � strongly agree). Scale reliability was very good (� � .95).

Results

The manipulation check was successful. Participants in the high-status condition (M � 6.38, SD � 2.32) reported higher status than did participants in the low-status condition (M � 3.95, SD � 2.38), t(500) � 11.61, p � .001, d � 1.03. Most importantly, participants in the low-status condition (M � 2.11, SD � 0.62) reported higher levels of state hostility than did participants in the high-status condition (M � 1.96, SD � 0.58), t(500) � 2.73, p � .001, d � 0.25. Moreover, they perceived more disadvantage (M � 4.88, SD � 1.40) than did participants in the high-status condition (M � 2.87, SD � 1.14), t(500) � 17.65, p � .001, d � 1.57. Perception of disadvantage was positively related to state hostility, r(502) � .39, p � .001. Finally, a bootstrapping analysis based on 5,000 bootstraps showed that perception of disadvantage mediates the effect of status condition on state hostility: The confidence interval for the indirect effect did not include 0 (95% CI � �0.48, �0.29).

Discussion

Study 3 provides the first experimental evidence that low (rel- ative to high) subjective SES leads to increased aggression. It thereby mirrors Studies 1a, 1b, and 2 in that it showed a negative relation between subjective SES and aggression and extends our previous studies by showing that this relation is causal. In addition, Study 3 demonstrated that perceptions of disadvantage account for the effect of differences in subjective SES on aggressive affect.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

184 GREITEMEYER AND SAGIOGLOU

Overall, these findings were fully in line with our theoretical model. However, Study 3 had some major limitations that need to be addressed. First, although the results suggest that subjective SES causally affects aggressive affect, it is unclear whether this effect is driven by low and/or high subjective SES. It may be that low subjective SES increases aggressive affect, but it is also conceivable that high subjective SES decreases aggressive affect. Hence, Study 4 employed an additional experimental condition in which participants learned the (false) feedback that they had an SES similar to others. Second, Study 3 did not employ a measure of actual aggressive behavior. Thus, it is unknown whether sub- jective SES not only has an impact on people’s aggressive affect, but also influences their aggressive action. To address this limita- tion, Study 4 employed measures of both aggressive affect and aggressive behavior. If subjective SES indeed has an impact on participants’ aggressive behavior, it is a further question whether perceptions of disadvantage do not only account for differences in aggressive affect but also underlie the effect on aggressive behav- ior. However, according to the model of relative deprivation (Smith et al., 2012), affective aggression rather than perceptions of disadvantage should serve as the proximate mechanism evoking aggressive behavior. Basing our predictions on this model, we examined whether subjective SES would influence perceptions of disadvantage, which in turn would affect aggressive affect, which finally would have a direct impact on aggressive behavior.

Study 4

Study 3 showed that low subjective SES causally increases aggression. Although the manipulation suggests that we success- fully induced differences in the experience of social status, some readers may be concerned that our manipulation in Study 3 primes materialistic desires or aspirations. Hence, in Study 4, we em- ployed a different approach to manipulate subjective SES. Con- cretely, we adapted a procedure by Callan, Shead, and Olson (2011) (see also Brown-Iannuzzi, Lundberg, Kay, & Payne, 2015; Callan, Ellard, Shead, & Hodgins, 2008). After answering a series of questions, participants were randomly assigned to either learn that they have a higher SES than similar others, an SES in the average range of the socioeconomic status of similar others, or a lower SES than similar others. Afterward, aggressive affect and behavior were assessed.

Method

Participants, procedure, and materials. Participants were 452 individuals (242 females, 210 males; mean age � 35.9 years, SD � 12.4). One participant reported to have less than a high school degree, 53 participants completed high school, 194 partic- ipants completed some college, 126 participants obtained a bach- elor’s degree, 52 participants had a Master’s degree, and 26 participants had a Ph.D. degree. The mean average monthly in- come was: $3,737 (SD � 8579).

At the onset, participants learned that the aim of the study was to investigate people’s financial beliefs and behaviors. They re- ceived the following instructions to a bogus test ostensibly de- signed to calculate their SES (adapted from Callan et al., 2011).

In this survey, we would like to investigate people’s financial beliefs and behaviors. To do so, we ask you to complete a series of ques-

tionnaires that are all part of a current research project examining comparative trends in the socioeconomic status (SES) of MTurkers. At the end of the survey, you will be given feedback about your SES as it compares with the SES of people who match your particular profile. Your feedback is determined by powerful statistical proce- dures that compare your personal information with information in a large database provided by similar others.

Participants further learned that an SES score of $0�20 is in the average range of the socioeconomic status of similar others. A negative SES score of less than –$20 means that they have a lower socioeconomic status than similar others. A positive SES score of more than �$20 means they have a higher socioeconomic status than similar others.

In actuality, all participants received false feedback. In the high SES condition, participants (n � 175) received a positive score (� C$87). In the medium SES condition, participants received a score close to zero (� C$3). In the low SES condition, participants received a negative score (– C$523). The same scores were used by Callan et al. (2011). Participants were then asked to write down their SES score and why they thought they received this score.

The manipulation check, perceptions of disadvantage (� � .85), and aggressive affect (� � .95) were assessed as in Study 3. Afterward, participants received the following information about a doctoral-level student who was introduced as the developer of the SES program (emphasis in original).

Project employee evaluation. Please think back to the socio- economic status program you completed earlier. As mentioned in the instructions, it is part of an ongoing research project of our lab. In fact, the SES program was developed by one of the lab’s doctoral students, who is soon to complete his PhD. Regarding this doctoral-level student, we would like to ask you for your opinion. As is now custom in organizational contexts, employee evaluation is based not only on superiors’ evaluation, but also on customers’ and colleagues’ judgments. Therefore, we would appreciate if you evaluated the developer of the SES program on the dimensions given on the next page. Please note that the other parts of the survey were not developed by this student. We will use your and other parties’ feedback to decide whether to extend the student’s current contract of employment.

Participants then evaluated the candidate by answering four questions (“how competent would you rate the student,” “how likable would you rate the student,” “how would you evaluate the student’s technical proficiency,” and “would you recommend ex- tending the student’s contract agreement”). All items were as- sessed on a scale from 1 to 7. Because they were correlated (� � .95), they were combined, using the average, with lower values indicating more aggressive behavior.

Results

The manipulation check was successful, F(2, 449) � 153.29,

p � .001, 2 � .41. Post hoc tests (Tukey) revealed that partici- pants who had received low-SES feedback (M � 3.18, SD � 2.53) reported lower status than did participants who had received medium-SES feedback (M � 5.41, SD � 1.57), p � .001, and participants who had received high-SES feedback (M � 7.42,

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

185SUBJECTIVE SES AND AGGRESSION

SD � 2.05), p � .001. The medium-SES feedback and the high- SES feedback conditions also differed significantly, p � .001.

Influence of subjective SES on the perception of disadvan- tage, aggressive affect, and aggression. We first examined whether state hostility, aggressive behavior, and perceived disad- vantage differed as a function of the feedback manipulation. In- deed, a one-factorial ANOVA revealed that SES feedback had a significant effect on participants’ aggressive affect, F(2, 449) � 25.64, p � .001, 2 � .10 (see Figure 1). Post hoc tests (Tukey) revealed that participants who had received low-SES feedback (M � 2.34, SD � 0.73) reported higher levels of state hostility than did participants who had received medium-SES feedback (M � 1.91, SD � 0.53), p � .001, and participants who had received high-SES feedback (M � 1.86, SD � 0.56), p � .001. The medium-SES feedback and the high-SES feedback conditions did not differ significantly, p � .712.

We also examined whether participants’ objective SES moder- ated the impact of the feedback condition on aggressive affect. For this analysis, we first standardized both education level and aver- age monthly income and averaged these scores into an overall objective SES index. Because the level of aggressive affect in the medium-SES and high-SES conditions did not significantly differ, we contrasted the low SES condition with the medium and high SES conditions combined. Using the PROCESS macro for SPSS (Hayes, 2013), a bootstrapping analysis (with 5.000 iterations) showed that the contrast remained a significant predictor of ag- gressive affect (95% CI � �0.59, �0.33). In contrast, for both objective SES (95% CI � �0.31, 0.20) and the interaction (95% CI � �0.13, 0.17), the confidence interval included 0.

There was a significant effect of the feedback manipulation on aggressive behavior, F(2, 449) � 19.07, p � .001, 2 � .08 (see Figure 2). Post hoc tests (Tukey) revealed that participants who had received low-SES feedback (M � 4.64, SD � 1.28) judged the doctoral-level student less positively than did participants who had received medium-SES feedback (M � 5.34, SD � 1.08), p � .001, and participants who had received high-SES feedback (M � 5.43, SD � 1.09), p � .001. The medium-SES feedback and the high- SES feedback conditions did not differ significantly, p � .769. Aggressive behavior was related to state hostility, r(452) � �.35, p � .001, in that the more hostile people reported to feel, the more negatively they rated the doctoral-level student.

We then examined whether participants’ objective SES moder- ated the impact of the feedback condition on aggressive behavior. However, the interaction was not significant (95% CI � �0.23, 0.35). The main effect of the contrast remained significant (95% CI � 0.48, 0.97), whereas the main effect of objective SES was not significant (95% CI � �0.67, 0.30).

Finally, the feedback manipulation also influenced participants’ perception of disadvantage, F(2, 449) � 100.00, p � .001, 2 � .31. Post hoc tests (Tukey) revealed that participants who had received low-SES feedback (M � 5.38, SD � 1.54) perceived more disadvantage than did participants who had received medium-SES feedback (M � 3.80, SD � 1.28), p � .001, and participants who had received high-SES feedback (M � 3.09, SD � 1.33), p � .001. The medium-SES feedback and the high- SES feedback conditions also differed significantly, p � .001. Perception of disadvantage was related to state hostility, r(452) � .48, p � .001, and aggressive behavior, r(446) � �.29, p � .001.

Sequential mediation analysis. Next, we performed a two- step mediation analysis. Because the high-SES and medium-SES conditions did not differ in terms of the effect on aggressive behavior, these conditions were combined for the analysis. It is proposed that low (relative to high and medium) levels of SES elicit the experience of disadvantage, which promotes state hostil- ity, which then evokes aggressive behavior. To test the signifi- cance of the overall indirect effect of SES feedback (low-SES vs. medium/high-SES) on aggressive behavior via experience of dis- advantage and state hostility, a bootstrapping analysis (with 5.000 iterations) that models the mediators as operating in serial was conducted using the PROCESS macro for SPSS (Hayes, 2013). This analysis revealed that the indirect effect was significantly different from zero (95% CI � 0.08, 0.25). It thus appears that the experience of disadvantage is a distal determinant of aggressive behavior, whereas aggressive affect is the proximal determinant. The mediation model and the path coefficients are illustrated in Figure 3.

Discussion

As in Study 3, these results showed that low (relative to high) subjective SES is related to increased aggressive affect, and the

2

3

4

5

St at

e ho

st ili

ty

Figure 1. State hostility as a function of subjective SES (Study 4). Error bars depict two standard errors.

2

3

4

5

6

7

tneduts DhPfotne

mgduJ

Figure 2. Aggressive behavior as a function of subjective SES (Study 4). Higher scores indicate less aggression. Error bars depict two standard errors.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

186 GREITEMEYER AND SAGIOGLOU

experience of disadvantage operated as the mediating mechanism accounting for the impact of subjective SES on differential levels of affective hostility. Extending our previous findings, Study 4 revealed that differences in subjective SES do not only causally influence participants’ aggressive affect but also aggressive behav- ior. In addition, these effects appear to be mainly driven by low subjective SES participants being more aggressive, rather than high subjective SES participants being less aggressive. Whereas low subjective SES participants reported more state hostility and judged the doctoral-level student less favorably than did medium subjective SES participants, state hostility and aggressive behavior did not differ between medium and high subjective SES partici- pants. Finally, differences in the experience of disadvantage me- diated the effect of subjective SES on aggressive affect, but it had no direct impact on participants’ level of aggressive behavior. Our results thus empirically support theorizing by Smith and col- leagues (2012) that differences in aggressive behavior appear to be evoked directly by people’s experienced aggressive affect.

Note that even participants who had received low SES feedback did not evaluate the doctoral-level student very negatively. In fact, their ratings were above the midpoint and thus slightly positive. It appears that participants in the low SES condition were not neg- ative but rather less positive than participants in the high SES condition. Nevertheless, a less positive evaluation may result in the doctoral-level student not keeping the job (which, we believe, constitutes serious negative consequences). In fact, a letter of recommendation that is positive but not stellar often does more harm than good. Likewise, in the academic setting, many grant agencies require the evaluation to be excellent, while very good would not be good enough. So a less positive evaluation may involve serious harm.

It is also noteworthy that for both aggressive affect and behav- ior, objective SES did not moderate the impact of subjective SES, suggesting that both participants high or low in objective SES were similarly influenced by our feedback manipulation. Moreover, objective SES had no significant impact on aggressive affect and behavior. Overall, it appears that people’s subjective SES, more than their objective SES, is associated with their level of aggres- sive affect and behavior.

Study 5

Study 4 demonstrated that low subjective SES increases aggres- sive behavior. Please note that the target of the aggression was associated with the participants’ plight of having a low SES in that

he created the program that gave the (false) feedback. Therefore, it is yet unknown whether low SES feedback causes people to become more aggressive if the target is not the source of their feelings of deprivation. To address this issue, in Study 5, we implemented an experimental variation of the target’s responsibil- ity. After participants received either high or low SES feedback, some of them evaluated the doctoral-level student who allegedly created the SES program, whereas others evaluated a doctoral- level student who created other parts of the survey but did not develop the SES program. Previous research into displaced ag- gression has shown that individuals often react aggressively against a target that is innocent of any wrongdoing after the individual is provoked by another person (Marcus-Newhall, Ped- ersen, Carlson, & Miller, 2000). However, according to the model of relative deprivation (Smith et al., 2012), taking action should be directed toward redressing their deprivation. Intentionally harming a doctoral-level student who has no relation to the source of one’s anger would thus not be of any instrumental value to this goal. In light of these findings, both outcomes are conceivable—that low subjective SES causes an increase to- ward a completely uninvolved target or only toward the target responsible for the feelings of low SES.

Testing whether aggression is also directed toward neutral tar- gets allows us to address an alternative interpretation of the pro- posed relation between subjective SES and aggressive behavior. It is important to note that a relatively negative evaluation of the doctoral-level student who created the SES test does not necessar- ily constitute an intention to harm. Rather, the devaluation of the doctoral-level student could be employed to depreciate the validity of the conclusions that can be drawn from the test results, which ultimately helps to maintain high self-regard. That is, it is con- ceivable that participants who received low SES feedback were mainly concerned about discounting the feedback, rather than trying to damage the doctoral-level student’s chances of keeping a desirable job. In contrast, aggressing toward an innocent target does not reduce the validity of the feedback and thus is not a means to diminish the experienced disadvantage. Hence, if we find that low subjective SES also increases aggression toward a person who was not the source of participants’ disadvantage, then we can be more confident that any negative evaluation is indeed intended to cause harm.

Method

Three hundred ninety-eight individuals participated in this study (227 females, 171 males; mean age � 35.6 years, SD � 12.4). Five participants reported to have less than a high school degree, 35 participants completed high school, 189 participants completed some college, 107 participants obtained a bachelor’s degree, 52 participants had a Master’s degree, and 10 participants had a Ph.D. degree. The mean average monthly income was $3,818 (SD � 8088).

The procedure and materials were similar to Study 4, with the following modifications. After participants received either low- SES (n � 197) or high-SES feedback (n � 201), they completed the measures of perceived disadvantage (� � .88) and state hos- tility (� � .96). Aggressive behavior was once again assessed via evaluations of a doctoral-level student (� � .90). However, the doctoral-level student was either responsible for evoking feelings

Figure 3. Influence of subjective SES on aggressive behavior, sequen- tially mediated by perceived disadvantage and hostile affect (Study 4). Unstandardized coefficients are displayed. All paths are significant unless otherwise noted. (High � medium status coded 1, low status coded 0.)

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

187SUBJECTIVE SES AND AGGRESSION

of low SES or not. Concretely, participants either learned that the doctoral-level student created the SES program and that he did not develop other parts of the survey or they learned that the doctoral- level student developed the mood questionnaire (which was the label we used for the measure of state hostility) and that he did not develop other parts of the survey. At the end of the questionnaire, we checked whether participants were aware of whether the doctoral-level student created the SES program or the mood ques- tionnaire. Fifty-one participants failed this attention check. How- ever, the pattern of findings was very similar when those partici- pants were excluded or not. Hence, we kept them in all analyses.

Results

The manipulation check was successful. Participants in the high-status condition (M � 7.35, SD � 1.84) reported higher status than did participants in the low-status condition (M � 2.89, SD � 2.26), t(396) � 21.60, p � .001, d � 2.16.

Influence of subjective SES on the perception of disadvan- tage, aggressive affect, and aggression. As in Studies 3 and 4, participants who had received low-SES feedback (M � 2.27, SD � 0.70) reported higher levels of state hostility than did participants who had received high-SES feedback (M � 1.85, SD � 0.52), t(396) � 6.93, p � .001, d � 0.68. Moreover, participants who had received low-SES feedback (M � 5.11, SD � 1.58) perceived more disadvantage than did participants who had received high- SES feedback (M � 2.96, SD � 1.26), t(396) � 15.06, p � .001, d � 1.50. Because responsibility of the doctoral-level student was manipulated after the measurement of perceived disadvantage and state hostility, this factor was not included in these analyses. Perception of disadvantage was positively related to state hostility, r(398) � .49, p � .001.

As in Study 4, objective SES did not moderate the impact of the feedback condition on aggressive affect (95% CI � �0.19, 0.13). The main effect of objective SES was also not significant (95% CI � �0.14, 0.37), whereas SES feedback remained to be a significant predictor of aggressive affect (95% CI � �0.56, �0.31).

To test whether subjective SES influenced aggressive behavior, a 2 (SES feedback: high vs. low) 2 (responsibility of doctoral- level student: status vs. unrelated) ANOVA was performed. It revealed that participants who had received low-SES feedback (M � 5.04, SD � 1.19) judged the doctoral-level student less positively than did participants who had received high-SES feed- back (M � 5.32, SD � 1.04), F(1, 394) � 6.16, p � .013, 2 � .02. The interaction was not significant, F(1, 394) � 0.56, p � .453, 2 � .00 (see Figure 4). Aggressive behavior was related to state hostility, r(398) � �.19, p � .001, and perception of disad- vantage, r(398) � �.14, p � .007.

We also examined whether objective SES moderated the impact of the feedback condition on aggressive behavior, but the interac- tion was not significant (95% CI � �0.19, 0.39). The main effect of objective SES was also not significant (95% CI � �0.66, 0.27), whereas the main effect of SES feedback remained significant (95% CI � 0.07, 0.51).

Sequential mediation analysis. As in Study 4, we tested the sequential mediation process that low levels of SES elicit the experience of disadvantage, which promotes state hostility, which then evokes aggressive behavior (see Figure 5). In fact, the boot-

strapping analysis (with 5.000 iterations) revealed that the indirect effect was significantly different from zero (95% CI � 0.01, 0.19).

Discussion

The primary aim of Study 5 was to examine whether low subjective SES would cause an increase in aggression even toward targets that were not responsible for participants’ experience of disadvantage. It did. Participants who had received low SES feed- back were more aggressive toward the doctoral-level student who was responsible for evoking feelings of low SES than were high- SES feedback participants, but they were also more aggressive toward the doctoral-level student who was not responsible for the experience of having low SES. These results show that people of low subjective SES even direct aggressive behavior toward neutral targets. Moreover, Study 5 replicated some of the key findings from Studies 3 and 4. First, low subjective SES led to increased aggressive affect. Second, the experience of disadvantage medi- ated the effect of subjective SES on aggressive affect, but it only indirectly did so for aggressive behavior (via aggressive affect). This study thereby provides further evidence for the claim that emotional reactions are the proximal motivator for action (Smith et al., 2012).

Overall, Studies 4 and 5 showed that low SES led to increased aggressive behavior. In both studies, we relied on how favorably participants evaluated a doctoral-level student, with a less positive evaluation possibly resulting in the doctoral-level student not keeping his job. Although this is a standard procedure to assess aggressive behavior (for reviews, Baron & Richardson, 1994; Bushman & Anderson, 1998), it is conceivable that a relatively negative evaluation of the doctoral-level student was not mainly intended to cause harm, but, rather, to discount the validity of the negative feedback. However, low subjective SES increased ag- gression toward a neutral target where a negative evaluation could not discount the feedback given to them. This suggests that the negative evaluation of the doctoral-level student is not merely aimed at the restoration of one’s status. Nevertheless, in Study 6, we employed a different measure of behavioral aggression where

1

2

3

4

5

6

7

Responsible status Responsible status unrelated

tneduts Dh P eht fo tne

m gd uJ

PhD student

Low subjective SES

High subjective SES

Figure 4. Aggressive behavior as a function of subjective SES and responsibility of the doctoral-level student (Study 5). Higher scores indi- cate less aggression. Error bars depict two standard errors.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

188 GREITEMEYER AND SAGIOGLOU

it is unequivocally clear that the intention is to cause harm (even when the target is the source of participants’ experienced disad- vantage).

Study 6

Rozin et al. (1986) have shown that people tend to bestow certain objects with magical properties. Hence, DeWall and col- leagues (2013) reasoned that any harm toward an inanimate object that represents an actual person is akin to direct aggression toward the person. They presented their participants with the option to place needles in a doll to punish another person. A series of studies suggested the Voodoo Doll task to be a valid and reliable measure of aggression, in that it correlates positively with aggressive per- sonality and actual aggressive and violent behavior. Accordingly, we expected that participants who had received low SES feedback would be more likely to punish another person (i.e., place more needles in the doll) than participants who had received high SES feedback. We further examined whether aggression would be only increased toward the person who was the source for participants’ experienced disadvantage, or (as in Study 5) whether an unin- volved party would be at risk, too.

Method

Participants were 537 individuals (322 females, 215 males; mean age � 34.5 years, SD � 11.3). Six participants reported to have less than a high school degree, 44 participants completed high school, 229 participants completed some college, 174 participants obtained a bachelor’s degree, 70 participants had a Master’s de- gree, and 14 participants had a Ph.D. degree. The mean average monthly income was: $3,803 (SD � 8785).

The procedure and materials were similar to Study 5, with the following modifications. After participants received either low- SES (n � 258) or high-SES feedback (n � 279), they completed the measures of perceived disadvantage (� � .88) and state hos- tility (� � .95). To measure aggressive behavior, the Voodoo Doll task paradigm was employed (DeWall et al., 2013; see also Ches- ter, Merwin, & DeWall, in press). As in Chester et al. (in press), all participants viewed 20 images that show what a doll looks like with 0 to 19 pins placed into it. They learned that the doll was intended to represent another person and they were asked to choose how many needles (up to 51) they would like to put in the doll to punish the other person. Participants either learned that the other person was the creator of the SES score program or the creator

of the mood questionnaire. The number of pins inserted into the doll constitutes our measure of aggressive behavior. We also checked whether participants were aware of whether the doctoral-level student created the SES program or the mood questionnaire. Perhaps because this attention check appeared one page after the manipulation, not one participant failed this check.

Results

The manipulation check was successful. Participants in the high-status condition (M � 7.36, SD � 1.92) reported higher status than did participants in the low-status condition (M � 2.93, SD � 2.34), t(535) � 21.60, p � .001, d � 2.07.

Influence of subjective SES on the perception of disadvan- tage, aggressive affect, and aggression. As in the previous studies, participants who had received low-SES feedback reported higher levels of state hostility (M � 2.29, SD � 0.71) and per- ceived more disadvantage (M � 5.15, SD � 1.59) than did participants who had received high-SES feedback (M � 1.84, SD � 0.46; M � 3.06, SD � 1.41, respectively), t(535) � 8.81, p � .001, d � 0.75; t(535) � 16.09, p � .001, d � 1.39, respectively. Perception of disadvantage was positively related to state hostility, r(537) � .50, p � .001. As in Studies 4 and 5, participants’ objective SES did not moderate the impact of the feedback condition on aggressive affect (95% CI � �0.11, 0.14). The main effect of SES feedback remained significant (95% CI � �0.55, �0.35). The main effect of objective SES was not significant (95% CI � �0.20, 0.21).

To test whether subjective SES influenced aggressive behavior, a 2 (SES feedback: high vs. low) 2 (responsibility of doctoral- level student: status vs. unrelated) ANOVA was performed. It revealed that participants who had received low-SES feedback (M � 5.47, SD � 12.27) placed a greater number of pins in the doll than did participants who had received high-SES feedback (M � 1.55, SD � 5.95), F(1, 533) � 22.64, p � .001, 2 � .04. The interaction was not significant, F(1, 533) � 0.73, p � .392, 2 � .00 (see Figure 6). Aggressive behavior was related to state hos- tility, r(537) � .43, p � .001, and perception of disadvantage,

Figure 5. Influence of subjective SES on aggressive behavior, sequen- tially mediated by perceived disadvantage and hostile affect (Study 5). Unstandardized coefficients are displayed. All paths are significant unless otherwise noted. (High status coded 1, low status coded 0.)

0

1

2

3

4

5

6

7

8

9

10

Responsible status Responsible status unrelated

N um

be r

of n

ee dl

es

PhD student

Low subjective SES

High subjective SES

Figure 6. Aggressive behavior as a function of subjective SES and responsibility of the doctoral-level student (Study 6). Error bars depict two standard errors.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

189SUBJECTIVE SES AND AGGRESSION

r(537) � .15, p � .001. Participants’ objective SES did not moderate the impact of the feedback condition on aggressive behavior (95% CI � �3.07, 1.06). Whereas there was a significant main effect of SES feedback (95% CI � �5.53, 2.29), the main effect of objective SES was not reliable (95% CI � �1.19, 5.36).

Sequential mediation analysis. We also tested the sequential mediation process that low levels of SES elicit the experience of disadvantage, which promotes state hostility, which then evokes aggressive behavior (see Figure 7). In fact, the bootstrapping analysis (with 5.000 iterations) revealed that the indirect effect was significantly different from zero (95% CI � �3.42, �1.32).

Discussion

Study 6 provided a conceptual replication of Study 5, employing a different measure of aggressive behavior. As predicted, partici- pants in the low SES condition were more likely to punish another person (placing needles in the dolls) than were participants in the high SES condition. Importantly, and as in Study 5, low SES increased aggression toward the source of participants’ disadvan- tage, but also toward an innocent third party. Moreover, mediation analyses demonstrated that low subjective SES evoked the expe- rience of disadvantage, which elicited aggressive affect, which then had a direct impact on aggressive behavior. Overall, it appears that hostile affect translates into aggressive behavior and that people learning about having a low SES are not able to withhold their hostile feelings even when confronted with an uninvolved party.

General Discussion

Based on relative deprivation theory (Smith et al., 2012), we reasoned that people of low subjective SES should be more ag- gressive than people of higher SES. In fact, the results of the present studies provide consistent evidence for this notion. Studies 1a, 1b, and 2, employing a cross-sectional design, found a negative link between subjective SES and trait aggression. Importantly, this relation held when controlling for the dark tetrad, as well as the Big Five and demographics. Thus, although there are a number of third variables contributing to the link between subjective SES and trait aggression, a subjectively low status independently fosters aggressive behavioral proclivities. In addition, the relation be- tween subjective SES and trait aggression statistically held when controlling for the influence of the participants’ level of actual

SES, and, in all studies, subjective more so than objective SES predicted aggression.

Studies 3 through 6 tackled the question of cause and effect by employing an experimental design. All four studies converge in that subjective SES causally affects aggression. Notably, the find- ings were similar irrespective of whether participants learned about the (false) feedback that one has a lower (or higher) SES than others or were asked to compare themselves with people at the very top or the very bottom. In addition, in all studies, partic- ipants high or low in objective SES were similarly affected by the status manipulation, in that the level of aggression was higher in the low compared to the high status condition. Study 4 further revealed that low subjective SES causes an increase in aggression rather than high subjective SES causing a decrease in aggression. It is noteworthy that two different study designs were used and that both yielded converging findings. Moreover, similar effects of subjective SES were found on aggressive affect and behavior, which further enhances confidence in the validity of the conclu- sions drawn. Finally, subjective SES had an impact on both verbal and physical behavioral aggression.

Theoretical and Practical Implications

The present studies lend credence to some assumptions derived from the theory of relative deprivation of how the perception of low SES should be related to the person’s judgments, emotional reactions, and actions. In fact, all studies provided consistent evidence that low rather than high subjective SES is related to more aggression. The present research also addressed how low subjective SES causes an increase in aggression. As proposed by the model of relative deprivation (Smith et al., 2012), learning that one is worse off than others led to the perception of being disad- vantaged, which in turn resulted in negative emotional reactions. This pattern of findings has been both shown in our correlational (Study 2) and experimental studies (Studies 3–6). As is also proposed by the model, Studies 4 through 6 did not find that the experience of disadvantage directly mediated the effect of subjec- tive SES on aggressive behavior, but, rather, aggressive affect was the most proximal antecedent of aggressive action. Overall, our studies provide strong support for the model’s claim that one’s position in a social hierarchy evokes cognitive comparisons, which are followed by cognitive appraisals, affective responses, and finally, interpersonal behavior.

Our findings also shed some light on how increased aggres- sive affect and behavior by people of low SES could be re- duced. The mediation analyses suggest that the key is to tackle the person’s experience of disadvantage. In everyday life, peo- ple of low SES often experience classism that denotes negative attitudes and behaviors directed toward the poor by the non- poor. For example, people with low SES are dehumanized by others (Loughnan, Haslam, Sutton, & Spencer, 2014) and ex- perience interpersonal and institutional distancing and social exclusion (Lott, 2002). This highlights the dynamics of the problem at hand: The derogating behavior toward low SES individuals fosters the subjective experience of their SES and thereby also increases the likelihood of acting out aggressively. Naturally, interventions should aim at people no longer dero- gating others on the basis of their relative SES (and altogether, for that matter). And of course, the results lend substantive

Figure 7. Influence of subjective SES on aggressive behavior, sequen- tially mediated by perceived disadvantage and hostile affect (Study 6). Unstandardized coefficients are displayed. All paths are significant. (High status coded 1, low status coded 0).

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

190 GREITEMEYER AND SAGIOGLOU

empirical support to the effectiveness of public policies that achieve greater redistribution of wealth in society. By attenu- ating the actual wealth gap, such policies are likely to reduce feelings of personal relative deprivation by people of low SES and, as a consequence, would also reduce interpersonal hostil- ity.

On an applied level, the present results speak directly to the mixed-income housing efforts that were initiated in the United States and a number of European countries to create neighbor- hoods that reduce the social, economic, and spatial isolation of public housing (Kleit, 2005). First empirical investigations of the social consequences of such neighborhoods lend strong support to a critical view on these efforts and are in line with the present findings. For example, young boys from mixed-income neighbor- hoods showed more antisocial behavior than boys from mainly poor neighborhoods (Odgers, Donley, Caspi, Bates, & Moffitt, 2015). Moreover, children from mixed-income neighborhoods were reported to have more conduct problems, and poor families generally felt less safe and distressed in such areas (Flouri, Mid- ouhas, Joshi, & Sullivan, 2015). Our findings similarly suggest that fostering the salience of one’s relative poverty increases rather than decreases antisocial behavior.

Limitations and Future Research

In all studies, we collected a relatively large U.S. community sample. Given that college student samples are relatively homog- enous in terms of their SES, we used MTurk samples that are shown to be more representative of the U.S. population than undergraduates (Buhrmester et al., 2011; Paolacci & Chandler, 2014). Nevertheless, we have to admit that the typical MTurk sample may also be prone to selection biases (e.g., only Internet users can participate via MTurk). Future research on the impact of subjective SES on aggression may employ samples that are even more representative of the general population.

Likewise, future research could complement our Internet studies with samples that involve face to face interactions. We aimed for high statistical power to detect significant effects, but also to increase the level of confidence in the estimate of the real effect. Given our limited participants’ pool, a lab study would have had inadequate statistical power (in particular for the impact of sub- jective SES on aggressive behavior where the effect size was small to medium).

Based on the model of relative deprivation, we examined whether state hostility would mediate the impact of the experience of disadvantage on aggressive behavior. However, people may not only respond to experienced injustice with aggressive emotions. In fact, the model also considers the possibility that people experi- ence fear or sadness as a consequence of relative deprivation. Importantly, these discrete emotions serve different functions and thus lead to different behavioral responses. Whereas anger moti- vates people to attack, fear is assumed to result in escape and sadness should evoke withdrawal. Hence, future research may assess a wider range of discrete emotions, as well as their likely behavioral consequences.

In our Studies 5 and 6, low subjective SES caused an increase in aggressive behavior toward those that were responsible for evoking feelings of having low SES but also toward others who were not responsible. It thus appears that the experience of having

low SES is even displaced onto others that have no relation to the source of one’s disadvantage (cf. Marcus-Newhall et al., 2000). However, it should be noted that the model of relative deprivation (Smith et al., 2012) suggests that when people respond to the experience of disadvantage with anger, they should be more likely to react aggressively toward the source of their deprivation (com- pared to an innocent other party). One reason we did not find support for this assumption may be that in our studies the doctoral- level student who created the SES measure evoked feelings of having low SES, but did not cause the outcomes. Hence, it is well conceivable that future research indeed finds that having low subjective SES is more likely to cause aggression toward the actual sources of being disadvantaged than toward neutral targets. For example, people low in subjective SES might be particularly aggressive toward upper class individuals, because they are seen as defending the status quo.

In the present research, we focused on the experience of indi- vidual relative deprivation that is evoked by comparing the indi- vidual’s situation with others that are better off. However, the model of relative deprivation (Smith et al., 2012) includes a further form of relative deprivation, namely, group relative deprivation. Group relative deprivation occurs when the own group’s situation is compared to superior other groups (e.g., when the salience of group membership is increased). Importantly, whereas individual relative deprivation predicts individual-focused behavior, group relative deprivation predicts group-focused behavior. For example, faculty members that responded with angry resentment to pay cuts were more willing to voice their individual concerns to the admin- istration (Osborne, Smith, & Huo, 2012). In contrast, the experi- ence of group relative deprivation is a better predictor for collec- tive action, such as support for political protest (Walker & Mann, 1987). Future research may address the causal effects of group relative deprivation on group-level aggression, as well as their underlying mechanisms.

The negative relation between subjective SES and aggression (that was documented in Studies 1a, 1b, and 2) could suggest that either one can be antecedent or consequence. Our experimental studies showed that low SES causes an increase in aggression. However, hostile behavior may also feed back to SES in that these reactions are considered inadequate by others, which then impairs one’s ability to attain a higher social class and/or causes one to drift down the socioeconomic hierarchy. In fact, the relationship between subjective SES and aggression is likely bidirectional and can be mutually reinforcing. Further research is needed to examine the possible dynamic interplay between subjective SES and ag- gression.

We consistently found that the perception of low SES caused an increase in aggression. However, it is conceivable that under different circumstances people with high SES are more prone to aggression. As shown in previous research (Piff, 2014) and also in Studies 1a and 1b, high subjective SES was associated with increased narcissism. Narcissism has been shown to be posi- tively associated with aggression (Donnellan, Trzesniewski, Robins, Moffitt, & Caspi, 2005). Importantly, however, narcis- sists are more aggressive than others when they had been insulted, whereas aggression levels are similar as long as there was no insulting provocation (Bushman & Baumeister, 1998). In our studies, participants who received high SES feedback were not provoked, so it may well be that people of high SES

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

191SUBJECTIVE SES AND AGGRESSION

are the most aggressive when their favorable view of them- selves is threatened.

Concluding Remarks

The perception of having low subjective SES has a negative impact on the person’s well-being and health (Adler et al., 2000; Gallo & Matthews, 2003). Our findings suggest that the person’s social environment may also suffer. Results consis- tently show that the perception of having lower rank in a social hierarchy evokes hostile emotions and aggression toward those that are deemed responsible for the person’s experience of disadvantage but also toward innocent parties. In sum, the present findings strongly suggest that greater wealth equality would substantially contribute to improved well-being and physical health, as well as benefit social interactions. Wealth equality could thereby posit a key factor in launching an upward spiral movement, from less inequality to less perception of disadvantage to less interpersonal hostility.

References

Adler, N. E., Epel, E. S., Castellazzo, G., & Ickovics, J. R. (2000). Relationship of subjective and objective social status with psychological and physiological functioning: Preliminary data in healthy white wom- en. Health Psychology, 19, 586–592. http://dx.doi.org/10.1037/0278- 6133.19.6.586

Agnew, J. (2001). The new global economy: Time-space compression, geopolitics, and global uneven development. Journal of World-systems Research, 7, 133–154. http://dx.doi.org/10.5195/jwsr.2001.167

Anderson, C. A., Deuser, W. E., & DeNeve, K. (1995). Hot temperatures, hostile affect, hostile cognition, and arousal: Tests of a general model of affective aggression. Personality and Social Psychology Bulletin, 21, 434–448. http://dx.doi.org/10.1177/0146167295215002

Anderson, C., Kraus, M. W., Galinsky, A. D., & Keltner, D. (2012). The local- ladder effect: Social status and subjective well-being. Psychological Science, 23, 764–771. http://dx.doi.org/10.1177/0956797611434537

Baron, R. A., & Richardson, D. R. (1994). Human aggression (2nd ed.). New York, NY: Plenum Press.

Brown-Iannuzzi, J. L., Lundberg, K. B., Kay, A. C., & Payne, B. K. (2015). Subjective status shapes political preferences. Psychological Science, 26, 15–26. http://dx.doi.org/10.1177/0956797614553947

Bryant, F. B., & Smith, B. D. (2001). Refining the architecture of aggres- sion: A measurement model for the Buss-Perry aggression questionnaire. Journal of Research in Personality, 35, 138–167. http://dx.doi.org/10 .1006/jrpe.2000.2302

Buckels, E. E., Jones, D. N., & Paulhus, D. L. (2013). Behavioral confir- mation of everyday sadism. Psychological Science, 24, 2201–2209. http://dx.doi.org/10.1177/0956797613490749

Buckley, M. R., Wiese, D. S., & Harvey, M. G. (1998). An investigation into the dimensions of unethical behavior. Journal of Education for Business, 73, 284–290. http://dx.doi.org/10.1080/08832329809601646

Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechan- ical Turk: A new source of inexpensive, yet high-quality, data? Per- spectives on Psychological Science, 6, 3–5. http://dx.doi.org/10.1177/ 1745691610393980

Bushman, B. J., & Anderson, C. A. (1998). Methodology in the study of aggression: Integrating experimental and nonexperimental findings. In R. G. Geen & E. Donnerstein (Eds.), Human aggression: Theories, research, and implications for social policy (pp. 23–48). San Diego, CA: Academic Press. http://dx.doi.org/10.1016/B978-012278805-5/ 50003-1

Bushman, B. J., & Baumeister, R. F. (1998). Threatened egotism, narcis- sism, self-esteem, and direct and displaced aggression: Does self-love or

self-hate lead to violence? Journal of Personality and Social Psychol- ogy, 75, 219–229. http://dx.doi.org/10.1037/0022-3514.75.1.219

Bushman, B. J., Bonacci, A. M., Pedersen, W. C., Vasquez, E. A., & Miller, N. (2005). Chewing on it can chew you up: Effects of rumination on triggered displaced aggression. Journal of Personality and Social Psychology, 88, 969–983.

Callan, M. J., Ellard, J. H., Will Shead, N., & Hodgins, D. C. (2008). Gambling as a search for justice: Examining the role of personal relative deprivation in gambling urges and gambling behavior. Personality and Social Psychology Bulletin, 34, 1514–1529. http://dx.doi.org/10.1177/ 0146167208322956

Callan, M. J., Shead, N. W., & Olson, J. M. (2011). Personal relative deprivation, delay discounting, and gambling. Journal of Personality and Social Psychology, 101, 955–973. http://dx.doi.org/10.1037/ a0024778

Chabrol, H., Van Leeuwen, N., Rodgers, R., & Séjourné, N. (2009). Contributions of psychopathic, narcissistic, Machiavellian, and sadistic personality traits to juvenile delinquency. Personality and Individual Differences, 47, 734–739. http://dx.doi.org/10.1016/j.paid.2009.06.020

Chester, D. S., Merwin, L. M., & DeWall, C. N. (in press). Maladaptive perfectionism’s link to aggression and self-harm: Emotion regulation as a mechanism. Aggressive Behavior.

Cohen, S., Alper, C. M., Doyle, W. J., Adler, N., Treanor, J. J., & Turner, R. B. (2008). Objective and subjective socioeconomic status and sus- ceptibility to the common cold. Health Psychology, 27, 268–274. http:// dx.doi.org/10.1037/0278-6133.27.2.268

Côté, S. (2011). How social class shapes thoughts and actions in organi- zations. Research in Organizational Behavior, 31, 43–71. http://dx.doi .org/10.1016/j.riob.2011.09.004

Crosby, F. (1976). A model of egotistical relative deprivation. Psycholog- ical Review, 83, 85–113. http://dx.doi.org/10.1037/0033-295X.83.2.85

DeWall, C. N., Finkel, E. J., Lambert, N. M., Slotter, E. B., Bodenhausen, G. V., Pond, R. S., Jr., . . . Fincham, F. D. (2013). The voodoo doll task: Introducing and validating a novel method for studying aggressive inclinations. Aggressive Behavior, 39, 419–439.

Donnellan, M. B., Trzesniewski, K. H., Robins, R. W., Moffitt, T. E., & Caspi, A. (2005). Low self-esteem is related to aggression, antisocial behavior, and delinquency. Psychological Science, 16, 328–335. http:// dx.doi.org/10.1111/j.0956-7976.2005.01535.x

European Central Bank. (2013). The Eurosystem household finance and consumption survey, results from the first wave. Statistics Paper Series, 2, April.

Evans, G. W. (2004). The environment of childhood poverty. American Psychologist, 59, 77–92. http://dx.doi.org/10.1037/0003-066X.59.2.77

Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117–140. http://dx.doi.org/10.1177/001872675400700202

Flouri, E., Midouhas, E., Joshi, H., & Sullivan, A. (2015). Neighbourhood social fragmentation and the mental health of children in poverty. Health & Place, 31, 138–145. http://dx.doi.org/10.1016/j.healthplace.2014.11 .009

Furnham, A., Richards, S. C., & Paulhus, D. L. (2013). The Dark Triad of personality: A 10 year review. Social and Personality Psychology Com- pass, 7, 199–216. http://dx.doi.org/10.1111/spc3.12018

Gallo, L. C., & Matthews, K. A. (2003). Understanding the association between socioeconomic status and physical health: Do negative emo- tions play a role? Psychological Bulletin, 129, 10–51.

Goodman, E., Adler, N. E., Kawachi, I., Frazier, A. L., Huang, B., & Colditz, G. A. (2001). Adolescents’ perceptions of social status: Devel- opment and evaluation of a new indicator. Pediatrics, 108, e31. http:// dx.doi.org/10.1542/peds.108.2.e31

Gosling, S. D., Rentfrow, P. J., & Swann, W. B., Jr. (2003). A very brief measure of the Big-Five personality domains. Journal of Research in Personality, 37, 504 –528. http://dx.doi.org/10.1016/S0092-6566 (03)00046-1

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

192 GREITEMEYER AND SAGIOGLOU

Greitemeyer, T. (2015). Everyday sadism predicts violent video game preferences. Personality and Individual Differences, 75, 19–23. http:// dx.doi.org/10.1016/j.paid.2014.10.049

Greitemeyer, T., & McLatchie, N. (2011). Denying humanness to others: A newly discovered mechanism by which violent video games increase aggressive behavior. Psychological Science, 22, 659–665. http://dx.doi .org/10.1177/0956797611403320

Griskevicius, V., Delton, A. W., Robertson, T. E., & Tybur, J. M. (2011). Environmental contingency in life history strategies: The influence of mortality and socioeconomic status on reproductive timing. Journal of Personality and Social Psychology, 100, 241–254. http://dx.doi.org/10 .1037/a0021082

Hayes, A. F. (2013). Introduction to mediation, moderation, and condi- tional process analysis: A regression-based approach. New York, NY: Guilford Press.

Johnson, W., & Krueger, R. F. (2006). How money buys happiness: Genetic and environmental processes linking finances and life satisfac- tion. Journal of Personality and Social Psychology, 90, 680–691. http:// dx.doi.org/10.1037/0022-3514.90.4.680

Jonason, P. K., & Webster, G. D. (2010). The dirty dozen: A concise measure of the dark triad. Psychological Assessment, 22, 420–432. http://dx.doi.org/10.1037/a0019265

Kennedy, B. P., Kawachi, I., Prothrow-Stith, D., Lochner, K., & Gupta, V. (1998). Social capital, income inequality, and firearm violent crime. Social Science & Medicine, 47, 7–17. http://dx.doi.org/10.1016/S0277- 9536(98)00097-5

Kiatpongsan, S., & Norton, M. I. (2014). How much (more) should CEOs make? A universal desire for more equal pay. Perspectives on Psycho- logical Science, 9, 587–593. http://dx.doi.org/10.1177/17456916145 49773

Kleit, R. G. (2005). HOPE VI new communities: Neighborhood relation- ships in mixed-income housing. Environment & Planning A, 37, 1413– 1441. http://dx.doi.org/10.1068/a3796

Kraus, M. W., Côté, S., & Keltner, D. (2010). Social class, contextualism, and empathic accuracy. Psychological Science, 21, 1716–1723. http:// dx.doi.org/10.1177/0956797610387613

Kraus, M. W., Horberg, E. J., Goetz, J. L., & Keltner, D. (2011). Social class rank, threat vigilance, and hostile reactivity. Personality and Social Psychology Bulletin, 37, 1376–1388. http://dx.doi.org/10.1177/0146167 211410987

Kraus, M. W., Piff, P. K., & Keltner, D. (2009). Social class, sense of control, and social explanation. Journal of Personality and Social Psy- chology, 97, 992–1004. http://dx.doi.org/10.1037/a0016357

Lott, B. (2002). Cognitive and behavioral distancing from the poor. Amer- ican Psychologist, 57, 100–110. http://dx.doi.org/10.1037/0003-066X .57.2.100

Loughnan, S., Haslam, N., Sutton, R. M., & Spencer, B. (2014). Dehu- manization and social class. Social Psychology, 45, 54–61. http://dx.doi .org/10.1027/1864-9335/a000159

Marcus-Newhall, A., Pedersen, W. C., Carlson, M., & Miller, N. (2000). Displaced aggression is alive and well: A meta-analytic review. Journal of Personality and Social Psychology, 78, 670–689. http://dx.doi.org/ 10.1037/0022-3514.78.4.670

Martini, T. S., Root, C. A., & Jenkins, J. M. (2004). Low and middle income mothers’ regulation of negative emotion: Effects of children’s temperament and situational emotional responses. Social Development, 13, 515–530. http://dx.doi.org/10.1111/j.1467-9507.2004.00281.x

Norton, M. I., & Ariely, D. (2011). Building a better America—One wealth quintile at a time. Perspectives on Psychological Science, 6, 9–12. http://dx.doi.org/10.1177/1745691610393524

Odgers, C. L., Donley, S., Caspi, A., Bates, C. J., & Moffitt, T. E. (2015). Living alongside more affluent neighbors predicts greater involvement in antisocial behavior among low-income boys. Journal of Child Psy-

chology and Psychiatry, 56, 1055–1064. http://dx.doi.org/10.1111/jcpp .12380

Ohbuchi, K., Kameda, M., & Agarie, N. (1989). Apology as aggression control: Its role in mediating appraisal of and response to harm. Journal of Personality and Social Psychology, 56, 219–227. http://dx.doi.org/10 .1037/0022-3514.56.2.219

Oppenheimer, D. M., Meyvis, T., & Davidenko, N. (2009). Instructional manipulation checks: Detecting satisficing to increase statistical power. Journal of Experimental Social Psychology, 45, 867–872. http://dx.doi .org/10.1016/j.jesp.2009.03.009

Osborne, D., Smith, H. J., & Huo, Y. J. (2012). More than a feeling: Discrete emotions mediate the relationship between relative deprivation and reactions to workplace furloughs. Personality and Social Psychol- ogy Bulletin, 38, 628–641. http://dx.doi.org/10.1177/014616721143 2766

Paolacci, G., & Chandler, J. (2014). Inside the Turk: Understanding Me- chanical Turk as a participant pool. Current Directions in Psychological Science, 23, 184–188. http://dx.doi.org/10.1177/0963721414531598

Pettigrew, T. F. (2015). Samuel Stouffer and relative deprivation. Social Psychology Quarterly, 78, 7–24. http://dx.doi.org/10.1177/019027 2514566793

Piff, P. K. (2014). Wealth and the inflated self: Class, entitlement, and narcissism. Personality and Social Psychology Bulletin, 40, 34–43. http://dx.doi.org/10.1177/0146167213501699

Piff, P. K., Stancato, D. M., Côté, S., Mendoza-Denton, R., & Keltner, D. (2012). Higher social class predicts increased unethical behavior. PNAS Proceedings of the National Academy of Sciences of the United States of America, 109, 4086–4091. http://dx.doi.org/10.1073/pnas.1118373109

Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Re- search Methods, Instruments, & Computers, 36, 717–731. http://dx.doi .org/10.3758/BF03206553

Rozin, P., Millman, L., & Nemeroff, C. (1986). Operation of the laws of sympathetic magic in disgust and other domains. Journal of Personality and Social Psychology, 50, 703–712. http://dx.doi.org/10.1037/0022- 3514.50.4.703

Saez, E., & Zucman, E. (2014). Exploding wealth inequality in the United States. Downloaded June, 16, 2015 from http://d3b0lhre2rgreb.cloudfront .net/ms-content/uploads/sites/10/2014/10/102014-wealth-brief.pdf

Sagioglou, C., & Greitemeyer, T. (2014). Bitter taste causes hostility. Personality and Social Psychology Bulletin, 40, 1589–1597. http://dx .doi.org/10.1177/0146167214552792

Sagioglou, C., & Greitemeyer, T. (2016). Individual differences in bitter taste preferences are associated with antisocial personality traits. Appe- tite, 96, 299–308. http://dx.doi.org/10.1016/j.appet.2015.09.031

Smith, H. J., & Huo, Y. J. (2014). Relative deprivation: How subjective experiences of inequality influence social behavior and health. Policy Insights from the Behavioral and Brain Sciences, 1, 231–238. http://dx .doi.org/10.1177/2372732214550165

Smith, H. J., & Pettigrew, T. F. (2014). The subjective interpretation of inequality: A model of the relative deprivation experience. Social and Personality Psychology Compass, 8, 755–765. http://dx.doi.org/10 .1111/spc3.12151

Smith, H. J., Pettigrew, T. F., Pippin, G. M., & Bialosiewicz, S. (2012). Relative deprivation: A theoretical and meta-analytic review. Personal- ity and Social Psychology Review, 16, 203–232. http://dx.doi.org/10 .1177/1088868311430825

Stouffer, S. A., Suchman, E. A., DeVinney, L. C., Star, S. A., & Williams, R. M., Jr. (1949). The American soldier: Vol. 1. Adjustment during army life. Princeton, NJ: Princeton University Press.

Twenge, J. M., Baumeister, R. F., Tice, D. M., & Stucke, T. S. (2001). If you can’t join them, beat them: Effects of social exclusion on aggressive behavior. Journal of Personality and Social Psychology, 81, 1058–1069. http://dx.doi.org/10.1037/0022-3514.81.6.1058

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

193SUBJECTIVE SES AND AGGRESSION

Walker, I., & Mann, L. (1987). Unemployment, relative deprivation, and social protest. Personality and Social Psychology Bulletin, 13, 275–283. http://dx.doi.org/10.1177/0146167287132012

Walker, I., & Smith, H. J. (2002). Relative deprivation: Specification, development, and integration. New York, NY: Cambridge University Press.

Wills, T. A. (1981). Downward comparison principles in social psychol- ogy. Psychological Bulletin, 90, 245–271. http://dx.doi.org/10.1037/ 0033-2909.90.2.245

Wolff, E. N. (2012). The asset price meltdown and the wealth of the middle class (No. w18559). Cambridge, MA: National Bureau of Economic Research. http://dx.doi.org/10.3386/w18559

Received March 1, 2016 Revision received March 23, 2016

Accepted April 4, 2016 �

Members of Underrepresented Groups: Reviewers for Journal Manuscripts Wanted

If you are interested in reviewing manuscripts for APA journals, the APA Publications and Communications Board would like to invite your participation. Manuscript reviewers are vital to the publications process. As a reviewer, you would gain valuable experience in publishing. The P&C Board is particularly interested in encouraging members of underrepresented groups to participate more in this process.

If you are interested in reviewing manuscripts, please write APA Journals at [email protected]. Please note the following important points:

• To be selected as a reviewer, you must have published articles in peer-reviewed journals. The experience of publishing provides a reviewer with the basis for preparing a thorough, objective review.

• To be selected, it is critical to be a regular reader of the five to six empirical journals that are most central to the area or journal for which you would like to review. Current knowledge of recently published research provides a reviewer with the knowledge base to evaluate a new submission within the context of existing research.

• To select the appropriate reviewers for each manuscript, the editor needs detailed information. Please include with your letter your vita. In the letter, please identify which APA journal(s) you are interested in, and describe your area of expertise. Be as specific as possible. For example, “social psychology” is not sufficient—you would need to specify “social cognition” or “attitude change” as well.

• Reviewing a manuscript takes time (1–4 hours per manuscript reviewed). If you are selected to review a manuscript, be prepared to invest the necessary time to evaluate the manuscript thoroughly.

APA now has an online video course that provides guidance in reviewing manuscripts. To learn more about the course and to access the video, visit http://www.apa.org/pubs/authors/review- manuscript-ce-video.aspx.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an Ps

yc ho

lo gi

ca l

A ss

oc ia

tio n

or on

e of

its al

lie d

pu bl

is he

rs .

T hi

s ar

tic le

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

194 GREITEMEYER AND SAGIOGLOU