The Enterprising Self and Prejudices toward Unemployed Persons

1 Introduction: objectives and theoretical frame

Neoliberalism has been a dominant way of governing societies and economies in Western societies since the late 90s (Dörre et al. 2014; Krasmann 2003; O’Malley 1996; Rose 1992, 1999). Many sociologists describe neoliberalism as a transnational political project (Krasmann & Volkmer 2007; Wacquant 2009) rather than as an economic theory or political ideology. In Western societies, such a project aims at a reconstruction of democratic-capitalistic welfare states with the figure of the enterprising self as the reference point for a new regulation of social matters and issues (Bröckling 2007; Bröckling et al. 2011a; Han 2014; Lessenich 2008).

In many political, social, and economic discourses, neoliberal societies are framed as economically highly efficient where individuals have more freedom than ever (e.g. Strassman 1976; Kapur et al. 1997; Stigler 1971). This paper sets out to shed light on the other side of the coin and ask how the enterprising self, a general orientation that translates into individual attitudes, may further exclusive and repressive tendencies in society by leading to strong associations with prejudices toward economically inefficient groups.

The focus will be on prejudices^[1] because they are highly functional legitimizing devices for establishing and reinforcing social inequality (Zick, Küpper & Heitmeyer 2010; Pratto et al. 2006) and serve to legitimize the collective and structural discrimination of groups (Sidanius & Pratto 1999). Prejudices toward unemployed persons are focused on since they form an exemplary group that does not conform with or, indeed, even contradicts the demands of neoliberal, activating, responsibilizing, and individualizing „workfare-societies” (Wacquant 2009).

In much of the sociological discourse on neoliberalism, such disciplining effects are rather hesitantly discussed since it is often assumed that the specific neoliberal form of exerting power via freedom and entrepreneurship translates frictionlessly into subjectivities across neoliberal systems (see, e.g., Burchell, Gordon & Miller 1991). Nonetheless, some sociological discourses on neoliberalism subtly implicate devaluations of unemployed persons (Bauman 2011; Günther 2002; Honneth 2002; Neckel & Dröge 2002; Offe 2001); yet, this aspect is – if explicitly mentioned – handled in a highly abstract, social-philosophical manner.^[2] This paper tries to get a hold of more concrete social mechanisms at the intergroup-level that are conducive to prejudices toward unemployed persons and thereby contribute to discrimination and social inequality in neoliberal societies. Such mechanisms have been studied too little in the sociological discourses. Getting an empirical hold of the general orientation of the enterprising self in terms of measurable attitudes is a new approach in this field and a precondition for such empirical examinations of inter-group mechanisms.

In order to grasp the respective mechanisms of devaluation, the notions of social dominance and legitimizing myths from the field of prejudice research will be invoked (Sidanius & Pratto 1999). Within this rationale, the enterprising self will be theoretically contrasted with classic elements of the Protestant Work Ethic (PWE), conceptualized as a contemporary transformation of these work values and beliefs, and its effects will be examined within the mechanism of prejudices toward unemployed persons.

Unemployed persons have, as a group, not yet received much attention with respect to prejudices and social distance, unlike, for example, the group of immigrants or other ethnically constructed groups (e.g. Pratto et al. 2006; Carvacho 2010; Zick et al. 2008). Against a background in which unemployed persons have become more and more salient as a socially constructed group and have experienced massive devaluations in contemporary public discourse that is in part intermixed with racism (e.g. Bohrer 2009; Sarrazin 2010; Berberich 2009), the research at hand helps to fill a gap in this research area.

First, Social Dominance Theory (SDT) (Sidanius & Pratto 1999) and the connection to more general research on social inequality will be introduced. The PWE and the enterprising self will then be discussed as two potential legitimizing myths in contemporary Germany,^[3] before their legitimizing function and thus their effects on prejudices toward unemployed persons are tested empirically.

2 Social dominance, legitimizing myths, and social inequality

Social Dominance Theory (SDT) aims at explaining functions and stabilities of group hierarchies, social inequalities, and discrimination within societies. In order to do so, the theory integrates political science, psychological, and sociological theories. SDT is thus neither a purely psychological nor a purely sociological theory, but rather represents an interdisciplinary theoretical frame that aims at understanding connections between individual attitudes, social structures, and institutional acting (Sidanius & Pratto 1999: 31).

The starting point of SDT is the insight that modern societies are ordered along categorical status systems that determine which groups are dominant and which groups are subdominant. The status position that an individual possesses via his or her belonging to a specific social group is decisive for his or her influence, possibilities, and resources within a given society. Members of dominant groups possess disproportionately positive values like power, authority, or possessions, whereas members of subdominant groups possess disproportionately negative social values like precarious positions/employment, bad health care, poorer housing, etc. (Sidanius & Pratto 1999: 32).

The authors differentiate between three basic social hierarchies: a hierarchy according to age, one ordered by gender, and an arbitrarily, culturally varying system, e.g. according to ethnic, religious, socio-economic, or cultural characteristics. While the first two hierarchical systems structure every society via the universal status features of gender and age, the third hierarchical system only appears in societies that function by division of labor and which create sustainable surplus (Sidanius & Pratto 1999: 38). This culturally varying and arbitrary status system is generally associated with the strongest forms of repression and violence toward low status groups (ibid.: 34).

The crucial assumption is that hierarchies and social inequality are (re-)produced and established through discrimination based on prejudices toward subordinate groups (Küpper & Zick 2005).

The subtheory of legitimizing myths

Prejudices are based on and connected with so-called legitimizing myths (LMs) that consist of culturally and time-specific attitudes, role models, general orientations, and ideologies. According to SDT, dominant groups use such ideologies or LMs to justify their dominant status and keep the subordinate groups in lower positions (see also Jost et al. 1994; Jost & Hunyady 2005). The theory argues that current LMs^[4] provide justification for social practices of exclusion that reproduce and maintain group-based social hierarchies (Sidanius & Pratto 1999).

The authors use the term LM interchangeable with „socio-political attitudes” (Pratto et al. 2006: 283), „social attitude” (Levin et al. 1998: 374), „cultural ideologies” (Pratto et al. 2006: 275), and ideology. They refer to Marx’ idea of false consciousness and ideology (Marx & Engels 1846), to Pareto (1901), and to Mosca (1896) within their subtheory of LMs (Sidanius & Pratto 1999: 21; see also Levin et al. 1998: 374). These classic political sociologists suggest that attitudes serve useful functions not only for individuals, but for entire social systems. Similarly SDT states that decisions and behaviors of individuals as well as the formation of new social practices and the operations of institutions are shaped by LMs. Certain types of social attitudes according to the classic political sociologists – referred to as „ideologies,” „legitimizations,” „ruling ideas,” or „political formulae” – are thought to serve justifications of hierarchical and inherently unequal relations between the ruling class and the working class.

SDT with a focus on LMs differs from the classic political sociologist approach in suggesting that „…traditional class conflict, rather than being the central form of conflict in human societies (as classical Marxists have long argued), is actually derivative of a much more general tendency for humans to establish relatively arbitrary ingroup-outgroup boundaries and to subsequently engage in various acts of ingroup favoritism and outgroup denigration on the basis of these distinctions.” (Levin et al. 1998: 376)

According to SDT, conflicts between economic classes are intertwined with and driven by similar mechanisms at the intergroup-level such as conflicts between different „races”, genders, ethnic groups, religions, age cohorts, nationalities, and any other salient, socially constructed group divisions where socio-political attitudes or ideologies, i.e. LMs play a decisive role (Levin et al. 1998).

Since the theoretical tool for the deduction of an empirically testable conceptual model for the research at hand is SDT, the socio-political attitudes described above will be referred to as LM (legitimizing myth) in the remainder of this paper.

SDT and social dominance orientation

According to SDT, depending on their „social dominance orientation” (SDO) (Sidanius & Pratto 1999: 48), individuals agree more or less to such LMs. SDO is a well-tested measure regarding internal and cross-time reliability, construct validity, and discriminant validity (Pratto et al. 1994). It represents „… the extent of individuals’ desires for group-based dominance and inequality.” (Pratto et al. 2006: 281) The basic idea is that „SDO orients people to find the most socially acceptable way of rationalizing inequality.” (Sidanius & Pratto 1999: 88) As several studies have shown, high-SDO individuals show more support for hierarchy-enhancing LMs than their low-SDO counterparts, and by means of these LMs they show more support for hierarchy-enhancing social policies (e.g. restrictive immigration policies or restrictive social policies) (Pratto et al. 2006).

One point of confusion with regard to SDT concerns the theoretical importance of the origins of and differences in SDO. Shouldn’t SDT be centrally concerned with the origins of this orientation? If so, what do sociologists have to do with this theory if it centrally concentrates on a psychological construct? However, as Pratto et al. have, in fact, stated, „… the theory is a theory of social dominance, not a theory of social dominance orientation.” (2006: 287) Measurements of SDO are viewed as a theoretical tool, rather than as a root cause of social hierarchy. More specifically, SDT argues that with the help of the concept of SDO one can empirically test whether a particular socio-political attitude or ideology serves as a LM that reinforces social inequalities or not. Thus, the utility of SDO lies in its being a research tool for understanding some of the processes that are conducive to group-based social hierarchy (Pratto et al. 2006). Thus, rather than being a strict personality theory of prejudice and discrimination, SDT operates at several levels of analysis, encompassing the level of individual differences, the level of social groups in context, and the level of competing ideologies within the social system as a whole (Pratto et al. 2006: 288).^[5]

SDT and social inequality

Just which groups are generally favored and which groups are disfavored and thus make up the high and the low status groups within a prevailing social structure is something reinforced by prejudices and discrimination as described above. „… When thousands of such acts of discrimination are aggregated over time, they stabilize groups-based social inequality.” (Pratto et al. 2006: 278) This is where SDT can be brought together with research on social inequality more generally – it sheds light on mechanisms at the intergroup-level that are conducive to what sociological researchers take stock of at a more abstract analytical level represented by objective criteria like access to income, education, or labor: social inequality.

Hradil (1999) defines social inequality as being present „… when persons regularly have access to more valuable goods in a society than others due to their status position in the systems of social relations.” (ibid.: 26, translation into English by the author) Following SDT, such structural social inequality and the mechanisms of discrimination at the intergroup level are in reciprocal dependency since „… [t]he structure of society itself, then, facilitates discrimination by individuals.” (Pratto et al. 2006: 278)^[6]

Sociologists themselves claim that sociological discourses on inequality and those on prejudices and (anti-)discrimination should refer to each other in terms of a „perspektivischen Dualismus [perspective dualism]” (Scherr 2010: 36). Socio-economical inequalities and prejudicial distinctions are seen as two distinct but interdependent types of hierarchy emergence that have – in their interplay – far-reaching consequences for living conditions and opportunities. Thus, discrimination and prejudices based on categorizations of groups, i.e. unemployed persons versus employed persons, and the respective ascriptions of character traits such as the unemployed being lazy, as transported by prejudices, are interrelated with socio-economic, political, and judicial disadvantages, which are objects of sociological research on social inequality (Scherr 2010: 55).

This does not mean that the social structure is determined by mechanisms of devaluation and discrimination at the intergroup-level. Rather, by additionally emphasizing these mechanisms at the intergroup-level, SDT highlights the reciprocal role that the value orientations, prejudices, ideologies, and socio-political attitudes of individuals play in both affecting and being affected by social hierarchy. Thus, SDT should be understood more as a complement to sociological research on social inequality than as a rival to it.

SDT has explicitly opened itself to sociological research questions by claiming that it is crucial to consider the material, cultural, institutional, gendered, and political contexts in which people live in order to develop truly social-psychological theories (Pratto et al. 2006). Sociological research could also open itself to complementary social psychological research such as SDT and in this fashion concur to more fully understand the emergence of social inequality.

SDT and principles of the just distribution of benefits and burdens

One obvious connection point is sociological research on principles of the just distribution of benefits and burdens (PJD) that serve to legitimize social inequality (see, e.g., Becker & Hauser 2009; Liebig 2007; Liebig & May 2009) and the subtheory of LM. Both ask the same fundamental question: what is it that collectively justifies unequal access to goods and thus helps to pacify social distributional conflicts? While sociological research on PJD generally holds on to the four basic, sociologically established principles (Becker & Hauser 2009), the subtheory of LM transports social-psychological knowledge that offers a fruitful frame for additional creative thinking on basic social justice orientations (BSJO) as well as alternative ways of empirical testing.

The notion within justice research that a new PJD of success is eroding the classic merit principle (Neckel 2008; Pongratz 2002) in neoliberal societies and the idea that hand in hand with such social change, new norms and values that transport new justice principles (i.e. new BSJO) need to develop in order to justify this social change (Liebig 2007: 133) are steps in this direction which could be enriched with ideas from SDT (see Klein & Groß 2011 for empirical research along this line).

In line with this notion from sociology, Pratto et al. (2006) see, as one of the major topics for future research, the creation and change of LMs (ibid.: 310), as well as the identification of why some ideologies become less powerful and others more so (ibid: 311). Sociology can help to fill this gap with its diagnosis of contemporary developments in neoliberal society. In the next section, the enterprising self and the PWE will be discussed and conceptualized as two distinct, potential LMs.

3 The enterprising self and the protestant work ethic

In using the notion of the enterprising self, Rose (1999) as well as Bröckling (2007) refer to Foucault’s (2007, 2008) analytical framework of governmentality (see also Burchell et al. 1991).^[7] The critical contribution of this concept to the study of neoliberal government lies in bridging practices bearing on the self, resulting in, for example, knowledge and truths to forms of power. In line with this notion of power and knowledge, a political value for the self as an enterprising self emerges (Han 2014; Lemke 1997, 2000, 2007; Lessenich 2008; Opitz 2004; Rose 1999; Wacquant 2009). The legitimizing underpinning of political activity in neoliberal governments consists of the presupposition of human nature as entrepreneurial, i.e. as active, innovative, creative,^[8] flexible, competitive, free, and highly self-reliant. This ideal self seeks continuously to maximize its individual human capital in order to be independent of state supervision. This general orientation and social role model, termed the enterprising self, inspires political mentalities and individual subjectivities within neoliberal societies (Rose 1999).^[9] In political discourse, national economic survival and competition in the world economy in the late 1980s were promised rescue by establishing these „entrepreneurial qualities” (Beckert 2009) in public institutions, especially education (see Lessenich 2008; Opitz 2004; Peters 2001; Wacquant 2009) as well as in individual conduct (see Kommission für Zukunftsfragen der Freistaaten Bayern und Sachsen 1997; PRWORA^[10]). The culture of dependency of the welfare states was in this manner to be overcome.

This culture of dependency associated with the Keynesian welfare regimes of the preceding decades was generally characterized by role models described in terms of the „disciplinary subject” by Opitz (2004: 297). The disciplinary subject is understood as passive: he or she is told how to move and act in order to ensure Fordist productivity (Prinz & Wuggenig 2007). It is less self-responsible for its own social security and thus less dependent on the forces of the free market than the enterprising self. Dominant forms of subjectivities and normed practices carried rather clear characteristics of the classic Protestant Work Ethic (PWE), stressing the disciplined performance of one’s normalized and routinized work duties, especially delayed gratification, hard work, and conformity (see Bauman 2011: 166; Hartmann 2002: 227; Sennett 2006).

Some authors describe the emerging neoliberal general orientation in terms of the enterprising self as a consequent extension of economic forms into all spheres of human behavior, including the private self-constructions and subjectivities of individuals (Lemke, Krasmann & Bröckling 2000: 16). Similarly, yet coming from a different theoretical tradition,^[11] Messner and Rosenfeld (2013) describe neoliberalism in the US in terms of a marketized ethic of anomie in contrast to the declining significance of the PWE.^[12] The enterprising self, similar to the anomic ethic, so the argument here, thus represents a new dimension of marketization of neoliberal societies in contrast to the PWE of Fordism that more than likely has an impact on the treatment and judgement of those groups that are currently labeled as economically unprofitable and burdensome, i.e. unemployed persons.

In accordance with the sociologist Kratzer (2006), who argues that radical marketization in post-Fordist societies produces and reproduces an escalation of social inequalities, the enterprising self is considered here to be a contemporary transformation of former work values and beliefs of Fordist-Taylorist regimes that represents a new dimension in the marketization of society. Moreover, transferred to the logic of SDT yet still in line with Kratzer (2006: 543), the enterprising self is considered a more powerful LM in the stabilization of present inequality than classic elements of the PWE, which were more potent in Fordist-Taylorist regimes. Following this argumentation along the rationale of SDT, hard work has today lost its power as an LM to discipline malfunctioning groups in comparison to the virtues and values of the enterprising self.

Since the PWE is a close companion or even an ancestor of the enterprising self, as is argued here, the PWE needs to be controlled for when the effects of the enterprising self on the group of unemployed persons are investigated. Even though some authors mention the notion of self-responsibility in the context of the PWE, which in this way then overlaps with the enterprising self, the central elements of the PWE are asceticism, discipline, and delayed gratification combined with hard work, conformity, and thrift (Furnham 1984, 1990; Maes & Schmitt 2001; Mirels & Garrett 1971), which are channeled into one’s „calling” in Weber’s classic formulation (Weber 1976 [1904–1905]). These central elements form the decisive point where the enterprising self can be distinguished from the classic PWE and conceptualized as a transformation of those earlier work values and beliefs. The enterprising self strives to be flexible and to break out of routine, to be creative, innovative, and unconventional in selling him- or herself and in maximizing human capital (Bröckling 2007). Thus, it is expected that the measures for the PWE as opposed to the enterprising self can be empirically distinguished as two distinct, latent constructs, which are likely to be positively correlated (Hypothesis 1).

The historical argument made in this paragraph cannot be tested in this study.^[13] Instead, the effects of both the PWE and the enterprising self as contemporary LMs within the mechanism of devaluation at the intergroup-level will be tested against each other in order to see which one is the more powerful contemporary LM.

In the next paragraph, the mechanism that leads to prejudices toward unemployed persons on the basis of the PWE and the enterprising self will be deduced in accord with the notion of social dominance.

4 The mechanism of devaluation at the intergroup-level – deduction of an emprical model

SDT suggests that an attitude only qualifies as a potent LM if it empirically mediates the relationship between SDO and discrimination against low status groups (Sidanius & Pratto 1999: 48). Since it is impossible to measure the actual act of discrimination in a survey on attitudes as used in this study, prejudices toward unemployed persons are referred to in the study instead. Following the mediation assumption within the rationale of SDT, it is expected that the enterprising self mediates the relationship between SDO and prejudices (Hypothesis 2), whereas the PWE does not (Hypothesis 3) after controlling for both potential myths simultaneously in one model (see figure 1).

Fig. 1:

Simplified conceptual model of the main variables

LMs are expected to have a strong hierarchy-maintaining and legitimizing function for the maintenance of the structural integrity of hierarchical relationships among social groups. This is an essential mechanism that furthers and reinforces social inequality according to the rationale of SDT (Levin et al.: 1998). Additionally, following assumptions about ideological asymmetry (Levin et al. 1998; Mitchell & Sidanius 1993; Sidanius et al. 1994), because of their hierarchy-maintaining function LMs are generally expected to be more strongly driven by SDO within higher status groups than within the lower status groups.^[14] Within the rationale of SDT, this asymmetry is one of several mechanisms „… helping to establish and maintain systems of group-based social hierarchy.” (Levin et al. 1998: 377)^[15] The rationale of ideological asymmetry as here described thus implies that the enterprising self is expected to be morestrongly driven by SDO in the higher status groups than in the lower status groups. Possibly the higher status groups have more resources for developing attitudes in accord with the enterprising self. This may be due to typical status specific professions in contrast to low status groups who rather work in areas where hard work actually still counts. Higher status groups may possibly be found in professions that force them more strongly into subjectivities in line with the enterprising self, i.e. where creativity and self-responsibility count more than working hard. This then results in high-SDOs in higher status groups being more likely to seek for the enterprising self as a legitimizing strategy than is the case in lower status groups. This eventually results in the phenomenon of ideological asymmetry with regard to the enterprising self here described. No such ideological asymmetry is expected with regard to the PWE since it is expected to be a less potent ideology with regard to reproducing the contemporary status hierarchy. Transferred to the expectation that the enterprising self mediates between SDO and prejudices, it may logically be expected to display stronger indirect effects between SDO and devaluation in higher status groups than in lower status groups (Hypothesis 4). Taken together, a moderated mediation (see Hayes 2013), where the strength of the mediation is contingent upon social status, is only expected for the enterprising self. All hypotheses can be expressed in the form of a conceptual path model as depicted in figure 1. This model will guide the empirical analyses.

5 An empirical test of the model

Participants and Procedure

The empirical study reported on here is based on a representative German sample from the „Group-focused Enmity” Project (N = 1780/N: representative split for items of the enterprising self = 833; e.g. Heitmeyer 2002; Zick et al. 2008). The data collection was conducted in 2010 by a professional survey institute using computer-assisted telephone interviewing (CATI). The standardized questionnaire was pretested beforehand. Only German participants were considered for analysis. The age of participants ranged from 16 to 95 years (mean = 51.46, SD = 16.3). The sample contained more females (55.8 %) than males (44.2 %). The highest school degree was reported by 17.6 % of the respondents, which is only slightly below the rate in the official statistics for 2010 (Statistisches Bundesamt, Germany). Missing values were treated with the full information maximum likelihood (FIML) procedure in the confirmatory factor analyses with latent constructs and with listwise deletion in the conditional process analyses.

Measurements^[16]

All items were used in German and translated into English for this paper.

The enterprising self. The items were developed by the author on the basis of Bröckling’s (2007) detailed qualitative analyses and reconstructions of this neoliberal general orientation, which was mainly based on self-help literature.^[17] Essential elements of the enterprising self are, on the one hand, responsibilization in terms of individualizing social security (Bröckling et al. 2011b) and, on the other hand, the general invocation of individuals to develop entrepreneurial virtues like self-responsibility, flexibility, creativity, and activity. Two dimensions were thus operationalized: Responsibilization and enterprising universalism. In this study, for the purposes of parsimony, only enterprising universalism was used. Nonetheless, the element of responsibilization is not ignored, since it is implicitly also transported in the dimension enterprising universalism. This dimension will be referred to as the enterprising self in the following. Participants responded to the four items of this measure on Likert-type scales ranging from 1 (completely disagree) to 4 (completely agree): „Nowadays, everyone can make something of oneself”; „If you can’t motivate yourself, you can only blame yourself if you fail”; „If you are not ready to risk something new, it’s your own fault if you fail”; and „If you have no ideas about how to sell yourself well, failure is your own fault.” The reliability was good (α = 0.75).

Status measurement. The survey included a subjective status measurement. Participants were asked to place themselves within the larger society: „In our society, there are social groups that stand more at the top and others that stand more at the bottom. Where would you place yourself on a scale from 1 (bottom) to 10 (top)?” (cf. „Unten-Oben-Skala” a top-bottom scale, Statistisches Bundesamt 2008, Cundiff et al. 2013). According to some authors (e.g. Neckel 2008; Ossowski 1957, Singh-Manoux et al. 2005), the subjective judgement of people with regard to their status in society as an indicator of their actual position within the arbitrary status hierarchy^[18] is at least as informative and close to reality as objective factors like occupational prestige, education, or income. In accordance with this assumption, the measurement of subjective status correlated positively with income and with education (see table 2 below), confirming that subjectively perceived social status corresponds in part to objectively measurable arbitrary status (see also Datenreport 2008: 176). At the same time, the correlations between subjective status and gender or age turned out to be rather weak, indicating that the subjective status measure used here rather expresses the arbitrary-set system of social hierarchy that – other than age and gender – is grounded in more obviously socially constructed and highly salient groups based on characteristics such as position, income, and education (Sidanius & Pratto 1999: 33 ff.). Since this arbitrary-set system of social hierarchy is characterized by a „… high degree of arbitrariness, plasticity, flexibility, and situational and contextual sensitivity” (ibid.: 33) in determining which group distinctions and characteristics are socially salient, they can never be fully captured in a single quantitative measurement. The use of only income or education as status indicators would hardly capture the full range of characteristics that make the neoliberal arbitrary-set system of social hierarchy. The subjective status measure offers a helpful alternative. When controlled for age and gender in the analyses, the subjective status measure captures the arbitrary-set status system pretty well since it even captures characteristics that were not explicitly asked for, as would be the case if only education or income were used. Moreover, as the correlation analysis shows, it does transport education and income, thus those two sociologically relevant dimensions are implicitly considered in the analyses when subjective status is used.

Prejudices toward unemployed persons. The dependent variable was prejudices against unemployed persons and expressed devaluations of this group. A measure taken from the „Group-focused Enmity” Survey (Zick et al. 2008) was used. Participants responded to the two items on Likert-type scales ranging from 1 (completely disagree) to 4 (completely agree): „Most unemployed are not really interested in finding a job” and „I find it scandalous when unemployed enjoy an easy life at the expense of the society.” The reliability value was good (α = 0.76).

SDO. Two firmly established items from the “Group-focused Enmity” Survey were used for the second dimension of SDO (opposition to equality, Sidanius & Pratto 1999: 62 ff).^[19] Participants responded on scales ranging from 1 (completely agree) to 4 (completely disagree), where high values indicate high SDO: „We should try to treat all groups as equal as much as possible”; „All people should be given an equal chance in life” (α = 0.72).

The Protestant Work Ethic (PWE). The operationalization of this dimension was based on a German Protestant Ethic scale (Maes & Schmitt 2001). For this survey, six items that represented two dimensions were pretested: frugality and appreciation of hard work. Only two items representing the hard-work dimension could, in the end, be used, ones which represent that dimension of the Protestant Work Ethic that is most distinct to the hypothesized new neoliberal ethic transported by the enterprising self. Certainly, it depends on how the PWE is defined whether elements of the enterprising self are conceptualized as parts of this ethic or not. The intention of this study is, nonetheless, to examine possible differences between a culture of work that was hypothesized to be dominant in Fordist-Taylorist regimes before the neoliberal turn and contemporary work values that are associated with the role model of the enterprising self. Thus is makes sense to specifically analyze those dominant elements of the two work regimes that are hypothesized as making the most difference theoretically, no matter what name they are actually given. The concrete items for the hard-work dimension of the PWE, which will be termed „PWE (hw)” in the rest of the paper are: „Hard work cultivates character” and “Working hard eventually makes you a better person.” Again, a Likert-type scale was used ranging from 1 (completely disagree) to 4 (completely agree). The reliability was acceptable (α = 0.69).

Controls. Age and gender (0=male, 1=female) were controlled for in all analyses since they are theoretically conceptualized as social stratification systems that appear to be universal (Sidanius & Pratto 1999: 36). In order to attain a more precise picture of the contemporary, arbitrary neoliberal social hierarchy and its reproduction, the effects of age and gender need to be factored out in the analyses where the above mentioned subjective status measure represents the main status indicator.

Data analyses

Hypothesis 1 suggests a confirmatory factor analysis (CFA) (Reinecke 2005) in order to test construct- and discriminant-validity of the two mediators the enterprising self and PWE (hw) and to see how they correspond to each other. Hypotheses 2, 3 and 4 collectively suggest a conditional process model, i.e. a moderated mediation model (Edwards & Lambert 2007; Muller et al. 2005; Preacher et al. 2007) with two mediators (Hayes 2013). Both distinct mediations, on the one hand the effects of SDO on devaluation through PWE (hw), on the other hand via the enterprising self, can be tested against each other in their respective strength that way. Moreover, ideological asymmetry can be assessed in the same model by conditioning the strength of the indirect effects on the value of a moderator, here subjective status. Regression based path analysis is employed for the assessment of this moderated mediation model with the aid of existing computational tools for estimating and probing interactions and conditional indirect effects in moderated mediation models (Hayes 2013).^[20] Confidence intervals for the population value of the (conditional) indirect effects were derived using bias corrected bootstrapping methods implemented in the computational tool. By applying bootstrapped confidence intervals, power problems that can be caused by asymmetric or non-normal sampling distributions of an indirect effect can be handled (MacKinnon, Lockwood & Williams 2004).

Results

To test Hypothesis 1, which postulates that „PWE (hw)” and „enterprising self” are two distinct, latent constructs, yet correlate positively with each other, confirmatory factor analyses were conducted. The two-factor model turned out to display considerably better model fit values than the one-factor model (see table 1). This means that the measures for the PWE (hw) and the enterprising self represent two distinct, latent constructs rather than different items for one and the same latent dimension, i.e. a more general contemporary work ethic or the like. Moreover, as expected in Hypothesis 1, the two distinct LMs are positively associated with each other (r=0.45***).

Moreover, descriptive analyses revealed that only 30 % of the respondent agree with the PWE (hw) whereas more than half of the respondents (51,4 %) agree with the enterprising self, giving a first hint that the enterprising self might be the dominant LM.

In the next step, bivariate correlations between all constructs were considered (see table 2). For the overall sample, the enterprising self and the PWE (hw), each separately, correlate positively with prejudice toward unemployed persons as well as with SDO. At the same time, SDO correlates significantly with prejudices. Taken together this implies that the preconditions for mediation of the effect of SDO on prejudice via the enterprising self as well as via PWE (hw) were fulfilled.

For the test of hypotheses 2 and 3, where it is expected that the enterprising self mediates the relationship between SDO (Hypothesis 2) and prejudice but the PWE (hw) does not (Hypothesis 3), mediation analysis with two parallel mediators, the enterprising self and the PWE (hw), was conducted using ordinary least square path analyses. As can be seen in table 3, contrary to expectation, both, the enterprising self and the PWE (hw) mediate the relationship between SDO and prejudice toward unemployed persons. Bias-corrected bootstrap confidence intervals for both indirect effects (uncond. indirect (ES) = .07, uncond. indirect (PWE (hw)) = .03) based on 10,000 bootstrap samples were each entirely above zero (.05 to .11 for ES and .02 to .06 for PWE (hw)).

Thus, within the rationale of SDT, both social attitudes serve as LMs. Nonetheless, the test whether there are any significant differences in the strength of the mediating effects between the two reveals that the enterprising self serves significantly more strongly as a transmitter of the effect of SDO on prejudice toward unemployed persons than the PWE (hw) does in the sample at hand (See table 3: contrast between the two indirect effects = .04 with the bootstrapped confidence interval remaining entirely above zero). Following the rationale of SDT, this indicates that the enterprising self displays a significantly stronger legitimizing effect on the prevailing arbitrary status hierarchy than the PWE (hw) does. This is in line with the expectations formulated in Hypotheses 2 and 3 though in a less intense way than expected, i.e. that the PWE (hw) does not display any indirect effect at all. Thus, highly dominance-oriented persons are more likely to strive to legitimize devaluation of the unemployed by adapting to the neoliberal ideology of the enterprising self than to the hard-work component of the PWE (hw). This indicates that the ideology of self-responsibilization is generally the more potent contemporary LM than commitment to hard work. At the same time, hard work does have some potential in legitimizing prejudices toward unemployed persons.

In the next step, conditional process modeling in terms of a test of moderated parallel mediation was conducted in order to assess ideological asymmetry as formulated in Hypothesis 4. More concretely, the hypothesized conceptual model posited that status would moderate the indirect relationship of SDO on prejudices toward unemployed person via the enterprising self and that this is due to an asymmetric relationship between SDO and the enterprising self in the mediational mechanism. This relationship is – following the thesis of ideological asymmetry – expected to be stronger when status is high than when people perceive their status as low. For the PWE (hw) no such ideological asymmetry was expected.

Fig. 2:

Plots of interaction between subjective status and SDO on the enterprising self and the PWE (hw); –/+1.54 is –/+1 SD away from mean (.00)

As shown in table 3, the cross-product between subjective status and SDO is significantly related to the enterprising self (.05, p=.04, CI does not include zero), indicating the existence of the phenomenon of ideological asymmetry with regard to the enterprising self. As expected, no such relationship was found for the PWE (hw) (.03, p=.20, CI includes zero).

The significant interaction for the enterprising self and the non-significant one for the PWE (hw) are visualized in figure 2. As displayed in this figure, where the regression lines are plotted for the mean value of the mean-centered subjective status (.00) and +/– 1 SD (+/– 1.54), a similar asymmetric tendency is found for the PWE (hw) as for the enterprising self since the regression line for the relationship between SDO and PWE (hw) also becomes steeper for people that judge their status to be higher. Nonetheless, this interaction does not reach significance for the PWE (hw).

Although these results are supportive of a significant, moderated mediation process with regard to the enterprising self (Hypothesis 4), bootstrapping results as a means to further verify the results were examined (table 4). The conditional indirect effect – the value of the indirect effect conditioned on values of the moderator (subjective status) – of SDO on devaluation of unemployed persons via the enterprising self is again examined at the three values of subjective status: the mean (.00), one standard deviation below the mean (–1.54), and one standard deviation above the mean (1.54).

As shown in table 4, bootstrap confidence intervals (bias corrected) indicate positive indirect effects for both mediators and for all, for low, for moderate, and for high status groups. Nonetheless, a closer examination of the coefficients suggests that the indirect effect grows with higher status for the enterprising self as mediator but remains relative stable when the PWE (hw) functions as mediator.

The bootstrap confidence intervals for the index of moderated mediation^[21] in table 4, confirm this impression. They indicate a significant conditional process at work for the enterprising self when modeled as a legitimizing myth (.02, CI does not include zero), but not for the PWE (hw) (.007, CI straddles zero). This means that the strength of the indirect effect of SDO on prejudices toward the unemployed on the part of the enterprising self differs significantly between the three status positions low, moderate and high in the expected direction:^[22] the higher the status, the stronger the enterprising self mediates the relationship, thus the more potent it is as a legitimizing ideology for the contemporary arbitrary status hierarchy. For the PWE (hw) as mediator, the indirect effects do not differ significantly between the different status positions. Thus, the data at hand indicate the phenomenon of ideological asymmetry for the enterprising self, but not for the PWE (hw).

In table 5 the conditional indirect effect via the enterprising self is additionally reported at multiple values of the moderator (subjective status) to complement the interaction displayed in figure 2 and in table 4 for the indirect effect. It also helps to identify regions of significance, viz. the values of subjective social status for which the conditional indirect effect is just statistically significant.

Persons that judge their social status as less than –1.45 on the mean-centered 10-point status scale (see the non-gray area in table 5, which is the region of non-significance, making up 23.3 % of the sample) do not use the ideology of the enterprising self to justify prejudices toward unemployed persons. The indirect effect reaches significance only for those people of higher status that place themselves higher than –1.45 on the mean-centered status scale (see gray area in table 5, which is the region of significance, making up 76.7 %).

This is another indication of the existence of ideological asymmetry regarding the enterprising self as legitimizing myth – an ideology that does not function as such among the 23.3 % who form the lowest status group.^[23]

Summary, limitations, and general discussion

The overarching objective of this research was to identify and empirically test mechanisms at the intergroup-level that are meant to complement more abstract social science research on social inequality in neoliberal society, i.e. in terms of examinations of structurally unequal access to education, income, or labor markets. In doing so, it follows the epistemological premise that any macro-level process or phenomenon needs to be grounded in some identifiable process involving individual actors and their activities. It should be possible, in other words, to „…illustrate the micro-instantiations” of the postulated macro-level phenomena (Jepperson & Mayer 2011: 66). Such mechanisms at the intergroup-level have not yet received much attention within sociological research on contemporary social inequality.

Referring to sociological rationales (e.g. Scherr 2010) as well as social-psychological thinking (e.g. Sidanius & Pratto 1999; Pratto et al. 2006), mechanisms that further prejudices toward low status groups at the intergroup-level were theoretically conceptualized in reciprocal relationships and interrelated with social structures of inequality. For the theoretical integration of sociological and social-psychological thought regarding structures of inequality and their reproduction, social dominance theory (SDT) was invoked in combination with the subtheory of legitimizing myths (LMs). This theoretical tool was combined with sociological insights on major turnovers in predominant normative patterns in neoliberal society, resulting in a „neoliberal paradigm” (Dörre et al. 2014) in contrast to Fordist-Taylorist forms of organisations in Keynesian welfare regimes. According to the authors, this neoliberal paradigm reveals the general neoliberal role model of the enterprising self that governs neoliberal subjectivities in contrast to the hard-work dimension of the Protestant Work Ethic (PWE (hw)), which formed parts of the disciplinary subject before the neoliberal turn (Opitz 2004). These two forms of subjectivity were measured in terms of individual attitudes aligned with the respective invocations in a representative survey in Germany and contrasted with each other in their potential to explain contemporary prejudices toward unemployed persons.

Unemployed persons make up a group that has not yet received much attention within social-psychological prejudice research even though this group has become more and more salient with respect to prejudices that carry accusations of inefficiency and cost-intensiveness under the neoliberal governmentality of the enterprising self (Groß et al. 2010; Hövermann et al. 2015). A conceptual model was deduced from the theoretical deliberations that guided the empirical analyses in order to help fill these gaps, i.e. the black box with regard to intergroup mechanisms conducive to prejudices within sociological research on inequality as well as the research gap regarding the group of unemployed persons within prejudice research. Central to these analyses was the interplay between dominance orientations, social status, the enterprising self, the PWE (hw), and prejudices toward unemployed persons.

Using representative data from a large survey in Germany, four theoretically derived hypotheses were assessed. The findings from conditional process modelling (Hayes 2013) were largely in accord with expectations. In contrast to expectations in line with SDT, where a significant mediation between social dominance orientation and prejudices toward unemployed persons was expected only via the enterprising self, both the enterprising self and the PWE displayed significant indirect effects. Within the rationale of SDT this means that the enterprising self has not fully displaced the PWE (hw) as a LM in neoliberal Germany; instead, the PWE (hw) functions as a LM parallel to the enterprising self in helping to justify the prevailing status hierarchy. Nonetheless, significant differences in the strength of the respective mediations were found, indicating that the enterprising self serves as a more potent LM than the PWE (hw) does for contemporary social hierarchies in Germany. Moreover, the assessment of ideological asymmetry for both potential LMs revealed that this asymmetry only seems to exist for the enterprising self and not for the PWE (hw). The results for the enterprising self as a LM could be replicated in an additional study with discriminatory intentions toward unemployed persons as the dependent variable. The results could not be presented in this paper due to the limited amount of space.

Within the rationale of SDT, the phenomenon of ideological asymmetry is considered to be a central mechanism in the societal reproduction of social inequality at the intergroup-level. The data at hand thus indicate that the ideology associated with Fordist-Taylorist society, the PWE (hw), is less potent in justifying the contemporary status hierarchy via devaluations of the unemployed in present-day Germany than is the neoliberal ideology of the enterprising self. These results corroborate the sociological diagnosis of a major change in the dominant normative patterns that have governed and structured society since the neoliberal turn as discussed, for example, by Dörre et al. (2014); Opitz (2004); and Kratzer (2006).

Certain limitations of the analyses need to be acknowledged. First, the operationalization of attitudes aligned with the general role model of the enterprising self must be regarded in a realistic way as a first attempt to conduct quantitative research in an area that is dominated by qualitative data. It was developed by the author on the basis of Bröckling’s (2007) qualitative analyses and reconstructions of this neoliberal social figure, mainly based on contemporary self-help literature. In order to find better quantitative indicators for the effects of the enterprising self as a general role model in terms of individual attitudes, more items need to be tested and validated. The items used in the present study contain explicit derogations of socially weak persons. This was intended because one essential element of the enterprising self is responsibilization in terms of individualizing social security (Bröckling et al. 2011b). Nonetheless, the individual accusations implicit in the present measurements may themselves transmit forms of devaluation, thus possibly leading to a tautology when testing the effect on prejudices. At this point, what is needed is to develop a measure that contains an individual alignment with entrepreneurial values without the facet of responsibilization. Possibly, it is only the responsibilization part of the measurement of the enterprising self, which stems from the business-related consulting literature, that has these effects in the high status groups. Recent discourses on governmentality that build on theories of subjectivization and are based on ethnographic studies focus more on consultant self-help and thus cybernetic aspects of contemporary governmentalities. These are comprised of techniques of the self, self-regulation, and the pressure to grasp approaching opportunities (Optionalisierung, see Traue 2010), the balance between self-recovery and entrepreneurship (Bührmann & Pongratz 2010) as well as the willingness to subordinate oneself after failure (e.g. Lessenich 2012). These aspects could serve as additional facets to be taken into account for a new measure, which may also help to better distinguish the measure for the neoliberal (the enterprising) self from the measure for the Fordist PWE.

Second, there are difficulties involved in assessing mediation (Bullock, Green & Ha 2010); thus, the results presented need to be interpreted carefully. Cross-sectional mediation models cannot prove or disprove causality; rather, only the fit of the model that implies causal hypotheses to empirical data can be assessed. To better test assumptions that imply causality, longitudinal or experimental data is needed.

Third, the measurement of the enterprising self as well as the models depicting the mediating mechanism of the SDO-prejudice relation via the PWE and the enterprising self that were presented here require additional testing in a cross-cultural setting. In the present research, it was understood as a given that a general neoliberal role model, as invoked by neoliberal politics, has an effect on personal attitudes. This idea is implicit in the way subjectivity is conceptualized within the rationale of governmentality (e.g. Butler 1997). As Guimond and colleagues (2013) have demonstrated, there is a distinction and a rather complex relationship between mental representations (personal attitudes, in this instance attitudes aligned with the general role model of the enterprising self) and cultural representations (cultural norms, comparable with the enterprising self as a general role model). At the same time, they found strong evidence that cultural representations have an effect on mental representations, not vice versa, which is in line with the implicit assumptions in the present research. Cross-cultural research is needed to find out more about the relationship of neoliberal politics and individual attitudes.

A central point of discussion raised by the analyses at hand that could be addressed within sociological discourse on governmentality is the fact that the enterprising self does not work well as legitimization for devaluations of unemployed persons within the subjective underclass in contrast to higher status groups. As mentioned before, in this sociological discourse, it is generally assumed that the specific neoliberal form of exerting power via freedom and entrepreneurship translates homogeneously and frictionlessly into subjectivities across social strata of neoliberal systems (see, e.g., Burchell et al. 1991). The results presented here indicate that this is not the case, i.e. that there are frictions in exerting power in this manner, thus impairing the assumption of homogeneity and posing additional questions concerning the mechanisms of governing at a distance and exerting power through freedom.^[24]

Finally, by way of closing, I suggest that the analyses at hand bear upon more general epistemological issues in the disciplines that deal with social inequality, viz. issues pertaining to theorizing about different levels of analysis. The debates between advocates of methodological individualism and methodological collectivism (or holism) are longstanding, well known, and sometimes contentious (e.g. Coleman 1986; Jepperson & Meyer 2011; Liska 1990; Sawyer 2001). Debates between sociologists and social-psychologists regarding the question about which of the two is the more meaningful perspective on the phenomenon of social inequality, the individual perspective intertwined with intergroup-mechanisms or the macro-sociological perspective, reflect those more fundamental debates. As soon as either perspective – methodological individualism or methodological collectivism – becomes „exclusivist and doctrinal” (Jepperson & Meyer 2011: 57), the utility of the other perspective is simply dismissed or minimized.

The approach of the research at hand was driven by the assumption that to fully understand a social phenomenon like social inequality along with its relatively peaceful reproduction and persistence, the insight needs to be taken seriously that higher level processes and phenomena like structural social inequality are usefully conceptualized as „‘emergent’ configurations” of lower level ones (Jepperson & Meyer 2011: 60) as are individual attitudes and processes at the intergroup-level. Put the other way around and taking the assumption of a reciprocal relationship between mechanisms of devaluation at the intergroup-level and structural social inequality seriously, macro-social processes, presumably also „penetrate” lower levels – not only the level of the individual in a society but the intergroup-mechanisms that are formed and operate in accordance with the dictates of cultural/institutional processes.

The results presented here are based on a novel attitude measure which was developed on the basis of sociological insights and analyses of neoliberal societies along with a specific concept of governmentality (Bröckling 2009) and which is meant to represent an attitude aligned with the neoliberal guiding principles of the enterprising self. Even though actual micro-macro connections could not be assessed with the data at hand due to limitations in the data-structure, this measure represents a „conceptual lynchpin” for linking the macro-cultural processes with the individual-level of analysis. The overall results for the models reveal processes at the individual and intergroup-levels that are – according to SDT – conducive to the reproduction of a neoliberal group-status hierarchy in which unemployed persons are assigned the positions with the lowest status. These processes – according to the theory – are in close reciprocal connection with the reproduction of structural social inequality. Even though, as mentioned above, it was not possible to test this macro-micro interrelationship in terms of empirical cross-level assessments with the data at hand, it is hoped to animate sociological thinking on the potential utility of theorizing at multiple levels of analysis and of explicating the linkages among these in matters of contemporary social inequality.