Interpersonal violent behavior is a social problem that policymakers and researchers from various disciplines alike have tried to understand and resolve for a long time. Sociological and social psychological studies have put great effort into explaining interpersonal violent behavior. Many of these studies agree that individual values can potentially explain the support and prevalence of violent conduct (for earlier studies see, e.g., Cohen, 1955; Matza and Sykes, 1961; Lerman, 1968; Merton, 1968; Hirschi, 1969; for newer ones, see, e.g., Knafo, 2003; Knafo et al., 2008; Bilsky and Hermann, 2016). However, although the commonality of these studies is the use of values as explanatory variables of violent behavior, some (particularly earlier studies) mixed up values with other explanatory variables such as opinions, norms, or beliefs or used the value concept without any systematic consideration of a value theory (Cohen, 1955; Matza and Sykes, 1961; Lerman, 1968; Merton, 1968; Hirschi, 1969). This neglect is unfortunate because values can prove to be important individual explanations for violent behavior. Furthermore, some of the studies did not explain violent behavior per se, but classified it under a general category of delinquent behavior. However, the motivational basis for interpersonal violent behavior may be fundamentally different from that of other types of delinquent behavior (such as theft or tax evasion).

In the current study, we address this gap by building upon Schwartz (1992, 1994) theory of basic human values, which is one of the most systematically developed and empirically tested value theories to date, and by using it to explain interpersonal violent behavior. Based on the definitions and implications of the theory, we present the underlying mechanisms for the relations between specific values, attitudes toward violence, and violent behavior. We test the expected relationships with population data from a sample of young people in Germany using structural equation models. Before turning to the empirical parts in Sections “Data and Measures” and “Results”, we present our theoretical considerations in Sections “Basic Human Values, Attitudes, and Behavior” and “Values and Interpersonal Violence”, respectively.

Based on the early work of Rokeach (1973), Schwartz (1992, 1994) developed his theory of basic human values.¹ He defines values as desirable goals that are effective across situations, ordered relative to their importance, and direct an individual’s choice and evaluation of behavior, people, and events.² Values emerge from organic biological needs, requisites of coordinated social interaction, and survival and welfare needs of groups. Schwartz identified 10 distinct values, each expressing a motivational aspect that is more or less important to the individual. Universalism refers to tolerance and protection of the needs of others, care for nature, and a sense for justice. Benevolence is concerned with loyalty as well as the preservation and enhancement of the welfare of those whom one is close to. Tradition emphasizes commitment to the customs and ideas from one’s culture and religion as well as personal restraint and modesty. Conformity describes the adjustment of behavior to social norms and the expectations of others as well as self-restraint. Security expresses the need for harmony, social stability, and family and personal safety. Power implies the goals of prestige, social status, and control over other people and resources. Achievement is concerned with personal success and demonstration of competence. Hedonism refers to the motivational need for pleasure, enjoyment, and the sensual satisfaction of short-term desires. Stimulation expresses the desire for arousal and excitement as well as for change and novelty in life. Finally, self-direction involves creativity, freedom, autonomy, and independent thinking and action. The theory also assumes that the 10 values are arranged in a circular structure (see Figure 1). Adjacent values are assumed to share a common motivational basis and to be highly related to each other. Values on opposite sides of the circle are of a different motivational type, and their relationship is assumed to be small or even negative.

Individuals are generally motivated to maintain consistency between their values and actions (Rokeach, 1973; Bardi and Schwartz, 2003). If a behavior expresses the content of a particular value (i.e., a “value-expressive behavior,” see, e.g., Bardi and Schwartz, 2003; Schwartz and Butenko, 2014), the relationship between that value and the behavior is likely to be positive. The relationship is assumed to be strong when values are able to distinguish between behaviors that are relevant for the attainment of goals from other behaviors. If a value has a high priority for an individual, it is more likely that this person will perform a behavior which realizes the motivation behind that value (Cieciuch et al., 2015). An extensive body of empirical research shows support for the relationship between values and behavior (Roccas and Sagiv, 2017).

However, attitudes may mediate the relationship between values and behavior. Attitudes are summary evaluations of specific (social psychological) objects, such as behavior, persons, institution, or events (Ajzen, 2001) and they serve as major determinants of intention and behavior in the theory of planned behavior (Ajzen, 1991). Values may influence not only behavior, but also attitudes, especially when they express the content of the value (“value-expressive attitudes”; see Maio and Olson, 2000; Maio, 2010).³ The underlying mechanism between values and attitudes relates to the attainment of motivational goals that are either promoted or suppressed by the attitude object (e.g., behavior; Sagiv and Schwartz, 1995; see also Davidov et al., 2008a). If the object helps to realize the motivational goals behind a value, the value–attitude link is expected to be positive. If the object blocks the motivational goals, the value–attitude link may be negative. However, the more attitudes correspond to a specific behavior, the more their influence on behavior is expected to exceed the influence of values (Homer and Kahle, 1988; Feather, 1995; Schwartz et al., 2010; Boer and Fischer, 2013). Therefore, we expect the relation between values and interpersonal violent behavior to be fully mediated by attitudes toward interpersonal violent behavior, and we test this expectation empirically. Figure 2 illustrates the possible causal relation between values, attitudes, and behavior. It should be noted that both attitudes and behavior are not necessarily influenced by only one single value. Various values could exert an influence on attitudes and behavior simultaneously (Bardi and Schwartz, 2003).

Finally, a word of caution. We do not assess causality in this study in a strict sense. Indeed, consequences of behavioral decisions and attitudes may also have feedback effects on values (Vecchione et al., 2016). Thus, our findings may reflect evidence for associations rather than unidirectional causality. However, the proposed causal linkage between values, attitudes, and behavior lies on strong theoretical reasoning. We have solid reasons to believe that associations between values, attitudes, and behavior are at least indicative of some causal chain (Schwartz et al., 2017).⁴

Violent behavior can be considered a subcategory of aggressive behavior (Anderson and Bushman, 2002; Bushman and Huesmann, 2010). Aggression is a behavior that is performed with the intention to cause immediate harm to another person. The offender believes that the behavior will in fact cause harm and the potential victim is motivated to avoid the aggression. Although aggression is always overt and observable, it can appear in one of two forms (Archer and Coyne, 2005; Card et al., 2008): direct (e.g., slapping a person in the face, threating, mocking, name-calling) or indirect (e.g., gossiping, manipulation of the victim’s social status and relationships).⁵

Violent behavior is always direct and has extreme consequences (e.g., death or injury). Thus, not all forms of aggression are violent, but all forms of violence are aggressive. All forms of aggression and violence appear as a social behavior and involve an interaction between at least two persons. These criteria exclude a range of phenomena from interpersonal aggressive and violent behavior, such as violence directed toward inanimate objects without the intention to harm another person (e.g., hitting a wall), aggressive emotions (e.g., anger, hate), or cognitions (e.g., beliefs or attitudes toward violence). Hence, we refer to interpersonal violence as overt acts that are intentionally carried out to cause physical harm to another person, who is motivated to avoid physical harm.

Previous studies systematically demonstrate associations between various values and interpersonal violent behavior. Cole et al. (2007) found universalism, security, tradition, and conformity to be associated with lower rates of self-reported delinquent behaviors, and power and hedonism to be positively related with self-reported fighting among Bahamian adolescents. Knafo (2003) and Knafo et al. (2008) found universalism and conformity values to be negatively related and power to be positively related to self-report of violent or bullying behavior among Arab and Jewish adolescents in Israel. Menesini et al. (2013) found that self-reported physically harmful behavior was negatively related to the higher-order values conservation and self-transcendence and positively related to openness to change and self-enhancement among Italian pupils. Benish-Weisman and McDonald (2015) and Benish-Weisman (2015) used peer nominations in a sample of Israeli adolescents to assess aggressive behavior. Results showed that aggression is positively related with self-enhancement and openness to change values and negatively related with self-transcendence and conservation values. Finally, Bilsky and Hermann (2016) demonstrated that tradition and conformity are negatively and hedonism and stimulation positively related to self-reported delinquency of respondents in two German cities.

Based on these studies and our theoretical considerations, we expect to find an association between interpersonal violent behavior and most of the values. Universalism expresses concern for the welfare of others, and violence harms the welfare of others. Therefore, universalism is expected to have a negative impact on interpersonal behavior. Benevolence expresses concern for the welfare of those with whom one is close. Since its motivational goal is associated with ingroup harmony, friendship, honesty, helpfulness, and care for close others, we also expect benevolence to hinder the use of violence. Therefore, benevolence is expected to have a negative impact.

Tradition and conformity values are expected to have a negative impact as well. In most social contexts, violent behavior stands in contrast to traditional customs and social norms. Thus, the acceptance of traditional customs, modesty, and (even more) self-restraint and obedience to social norms are assumed to inhibit the evaluation of violence as an appropriate means of action. Similarly, security is also expected to have a negative impact on interpersonal behavior. Violent behavior poses a threat to the social order and to one’s personal and family security. Therefore, it is undesirable for individuals striving for order and security, and they are expected to refrain from violence or from supporting it.⁶

While universalism, benevolence, tradition/conformity, and security are expected to have a negative effect on interpersonal behavior, stimulation and power are expected to have a positive effect. Stimulation expresses a pursuit of excitement and risk-taking. Carrying out violent behavior can be both exciting and risky. Furthermore, violence may be regarded as a means to exercise control and power over people and resources and to gain respect. Thus, interpersonal violent behavior can prove itself to be beneficial for selfish purposes and as a useful means to gain power.

Achievement, which is located next to the power value on the value circle, is not expected to have any effect on interpersonal violence. On the one hand, ambition and success, which are important elements of achievement, may be enhanced by using violence. On the other hand, achievement is guided by social norms and conventional social standards, and most individuals motivated by this value are likely to conform to such standards. Therefore, overall we do not expect any association between achievement and interpersonal violence. Hedonism is also not expected to have any effect on interpersonal violent behavior. On the one hand, hedonism is concerned with seeking fun and pleasure. There are studies which link hedonism to psychopathy (e.g., Abraham and Rahardjo, 2015; Kajonius et al., 2015), showing that some individuals can find pleasure in violent behavior. On the other hand, violence is not comfortable or gratifying and neither fun nor pleasing for many other people. Therefore, overall we do not expect any relationship between hedonism and interpersonal violence. Finally, self-direction is not expected to have any effect. We see no mechanism to associate the goals of independence and creativity, which stand behind self-direction, with violence.

In sum, both previous studies and our theoretical considerations suggest single values that stand out with either positive or negative relations to aggression and interpersonal violence. Single values that stand out with negative relationships to interpersonal violence are universalism, tradition, conformity, and security. The strongest positive relationships to aggression and interpersonal violence are expected for power and stimulation values.

With the exception of Bilsky and Hermann (2016), all the previous studies discussed above assessed samples of adolescents. Adolescence is a dynamic period of social, cognitive, and neurobiological development that deserves special attention in the study of aggression and violence (see Benish-Weisman et al., 2017). The transition into adulthood (the so-called emerging adulthood, see Arnett, 2000, 2007) is a period of changes. Although violent behavior generally decreases with age (Farrington, 1986), it will not completely disappear in emerging adulthood. Therefore, our study contributes to the understanding of the behavior- and attitude-guiding functions of values in the period between adolescence and adulthood. Next, we will describe our dataset and test our theoretical expectations.

The relevance of human values for the study of the motivational sources of interpersonal violent behavior was investigated in different fields of the social sciences. However, various past studies failed to distinguish values from other dimensions such as attitudes, norms, or beliefs, and only very few systematically assessed the effect of values on violent behavior relying on a value theory. Furthermore, in various studies interpersonal violence was often analyzed in composite indexes of general delinquent behavior rather than in its own right. In the current study, we addressed these gaps in the literature by building upon Schwartz’ theory of basic human values, one of the most systematically developed and empirically tested value theories to date. We used it to explain attitudes toward interpersonal violence and interpersonal violent behavior. We analyzed data for young people in the period of emerging adulthood (drawn from a German study in Duisburg, Germany), which assessed violent behavior and different sources of violent behavior. We used structural equation models to estimate the relationships between values, attitudes, and violent behavior simultaneously.

Our results demonstrated that attitudes toward interpersonal violence displayed a significant relationship with interpersonal violent behavior regardless of the specific value that was analyzed. Thus, attitudes appeared to be an important prerequisite for performing such a behavior. Furthermore, we found that several specific values were associated with attitudes toward interpersonal violence. Some were also directly linked with interpersonal violent behavior (and not only indirectly via the attitudes). Universalism, benevolence, tradition/conformity, and security values were negatively related with attitudes toward interpersonal violence and, in addition, directly related with behavior. Thus, attitudes operated as a partial mediator between these values and violent behavior, and despite the dominant association between attitudes and behavior, values themselves had a significant additional contribution to the explanation of behavior, thus inhibiting violent behavior. Schwartz et al. (2017) argued that values have an effect on behavior especially when the action preserves and promotes the goals that these values express. This effect is particularly strong and potentially direct when the individual gives a high priority to these values. Since violence against others strongly contradicts the motivational goals of universalism, benevolence, tradition/conformity, and security values (e.g., concern for the welfare of others, compliance with social expectations and rules), these values are directly and inversely related with the behavior. These findings provided empirical support for our expectations that values from the dimensions self-transcendence and conservation are particularly relevant for and predictive of interpersonal violence.

By way of contrast and in line with our theoretical considerations, power and stimulation were positively related with attitudes toward interpersonal violence but had no direct relation to behavior. Thus, their relation with interpersonal violent behavior was fully mediated by attitudes toward the behavior. Indeed, pursuing dominance and thrill did not provide motivations, which were strong enough to be directly linked with violent behavior, and the relation of the values with behavior was mediated by attitudes toward violence. Thus, not all values from the dimensions self-enhancement and openness to change propel interpersonal violence. Only insofar as a motivation for dominance and thrill seeking is concerned, as is the case for the values power and stimulation, a positive effect of values on interpersonal violence could be empirically supported by the data. This finding underlines the need to test how the distinctive effect of specific values operates in predicting attitudes toward violence and violent behavior (Schwartz et al., 2017). Achievement and self-direction displayed a smaller negative effect on attitudes toward interpersonal behavior; thus, these values also displayed an inhibiting effect on attitudes toward interpersonal violent behavior. Finally, hedonism showed neither direct nor indirect associations with either attitudes or behavior.

This study is not without limitations. Since values were measured only once, it was not possible to assess the causal pattern between values, attitudes, and behavior. It might very well be the case that the more people engage in violence, the more they become inured to its consequences. Hence, violence may also lead to an assessment of violent acts as less serious and to an adaptation of the value priorities and attitudes. One would need panel data with repeated measures for values, attitudes toward interpersonal violence, and violent behavior to assess the direction of causality more closely or an experimental design using, for instance, a factorial survey (see, e.g., Atzmüller and Steiner, 2010; Dülmer, 2015). The possibility to test equivalent models with contemporaneous effects in the opposite direction is tempting, but deemed inadequate to determine causal dominance (Bollen and Pearl, 2013). Instead, our argumentation for the specified direction of effects relies on sound theoretical reasoning, and we have solid reasons to believe that associations between values, attitudes, and behavior are at least indicative of one possible causal chain (Schwartz et al., 2017). Since values are considered much more abstract and stable than attitudes or behavior, we assumed that the causal direction follows from values to attitudes to behavior (or directly from values to behavior), but we do not exclude the possibility that it might also operate in the other direction (e.g., Vecchione et al., 2016). Hence, the assessment of reversed causality and causal dominance is an important aspect for future research on the topic.

As values were measured only at one point in time, we also did not address the developmental process of values, attitudes, and behavior and how their developmental pattern may be linked in this study. This, too, could be an exciting avenue for future research. In addition, interpersonal violent behavior is a rare event and as such limits the analytical possibilities, which would be otherwise much more flexible for normal data (King and Zeng, 2001). Particularly interpersonal violent behavior (but to a certain extent also attitudes toward such behavior) may have suffered from underreporting, since respondents may have hidden their positive attitude toward violence or violent behavior due to social desirability bias (e.g., Meyerl, 2013). After all, violence is clearly a socially undesirable behavior. This in turn may result in an underestimation of the true effects for some values which are desirable like universalism or benevolence, or with an overestimation for other values that are considered rather undesirable, particularly tradition, conformity, and security (Schwartz et al., 2012). However, the designers of the study took precautions to reduce this problem to a minimum as described in the “Data and Measures” section, for example, by guaranteeing anonymity and avoiding the collection of any personal data that allows identifying specific respondents. Future research may try to address this problem by carefully considering the sensitivity of the questions asked and trying to control for it.

Finally, the 21 items to assess values are limited in their capacity to discriminate between 10 distinct values (Davidov, 2008, 2010; Davidov et al., 2008b). Although the scale worked well in this study, the tradition and conformity values had to be treated as a combined value. Longer versions of the questionnaire (e.g., PVQ-40, PVQ-RR) are needed to fully capture the continuum of human values Cieciuch and Schwartz, 2012).

In spite of these limitations, the current study demonstrated that values, attitudes toward violence, and violent behavior are linked, and that values can predict violent behavior, either indirectly via attitudes toward violence or directly. Indeed, although interpersonal violence is (fortunately) a rare event, abstract values – in combination with attitudes toward interpersonal violent behavior – are able to explain interpersonal violent behavior to a large extent and deepen our understanding of its motivational sources.

The study was carried out doing secondary data analysis of a project at the University of Bielefeld. For further details, one may contact the principal investigator Jost Reinecke (jost.reinecke@uni-bielefeld.de). No data were collected by the authors themselves for the preparation of the submitted study.

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer GM and handling Editor declared their shared affiliation and their involvement as co-editors in the Research Topic, and confirm the absence of any other collaboration.