Moral Motivation, Moral Judgment, And Antisocial Behavior

For decades, moral judgment was considered central to moral functioning and moral action (e.g., Kohlberg, 1981; 1984). Despite extensive evidence for the link between moral judgment and behavior (e.g., Derryberry & Thoma, 2005; Thoma, 1994), it is a weak relationship. People often do not take the action they think is the moral choice (Blasi, 1980). The evidence supporting the weak link between sophisticated judgment and action is so prevalent that some researchers now consider Plato’s famous statement that “to know the good is to do the good,” to be empirically invalid (Frimer & Walker, 2008). The contemporary view is that moral judgment alone is insufficient to explain moral or immoral actions because it neglects other areas of moral functioning. For example, moral judgment does not indicate how sensitive someone is to moral issues, whether she has the skills to implement moral goals, or whether she is motivated to take moral action (Rest, 1983, 1986). In describing the nature of moral behavior, Rest (1983) suggested at least three other psychological processes beyond moral judgment. Here, we adopt Rest’s view of multiple component processes and apply it to a model for inhibiting antisocial behavior. We focus specifically on the components of judgment and motivation.

In a review of empirical studies, Rest (1983) posited that at least four psychological processes must occur in order for a moral behavior to take place: sensitivity, judgment, motivation and implementation. The four components are interactive and mutually influential. They are described in a logical (but not necessarily chronological) order.¹ (1) Moral sensitivity comprises perceptual and interpretive capacities. For example, does the person notice the moral situation (moral perception)? If so, what is the emotional response, and how elaborately does the individual understand the configuration of events (moral imagination)? Moral sensitivity includes perceptual attunement, emotional response (e.g., empathy), and discerning what types of actions are possible, as well as who or what might be affected by possible actions and the outcomes that might result from them. (2) Moral reasoning and judgment involve weighing the costs and benefits of each previously identified action and choosing the best one to take in the circumstances. Numerous considerations can be brought to bear on the process of coming to a moral judgment, including ethical responsibilities as a person or professional, the needs of other involved individuals and communities and anticipated outcomes. (3) Moral motivation concerns giving priority to the chosen moral action above all other concurrent priorities with the intention of fulfilling it. Moral motivation is reflected in the extent to which one gives priority to moral values, as opposed to other values of the moment (situationally) and over the long term (dispositionally). (4) Moral implementation involves having the strength and skills necessary to carry out the moral action. This includes knowing what steps to take and how to complete them, and managing any distractions along the way. In other words, to complete a chosen action, one must have competence and perseverance. Failure in any one of these components will result in failure to implement moral action. Thus the weak link between judgment and action can be explained by components beyond moral judgment, such as motivation or action skills.

Antisocial behavior entails various delinquent acts such as theft or assault, and is associated with dishonesty and a disregard for social norms (Rowe, 1985). Much research has attempted to explain antisocial actions relating them to moral reasoning or judgment. In essence, moral judgment measures are proxies for moral cognitive development which moves toward greater capacities for abstraction and inclusion of other viewpoints when resolving moral dilemmas. Greater abstract and inclusive thinking should allow for the moral course of action to take priority. Indeed, using the Defining Issues Test (Rest, 1979) and matching delinquents with nondelinquent peers, moral judgment scores are significantly lower among delinquents (Rest, 1979; Thoma, Rest, & Barnett, 1986). Several studies support this finding using different methodologies. For instance, using structural equation modeling, Raaijmakers, Engels, and van Hoof (2005), found that antisocial behavior scores of 24- to 26-year-old Dutch participants were negatively related to more sophisticated moral reasoning as assessed three years earlier by the Defining Issues Test (Rest et al., 1999). In a study using the Sociomoral Reflection Measure-Short Form (SRM-SF; Gibbs, Basinger, and Fuller, 1992) young Taiwanese offenders demonstrated less mature moral reasoning than a comparison group of Taiwanese nonoffenders (Chen & Howwit, 2007). A meta-analytic study, testing the significance of a pooled mean effect size from 15 studies that used different methodologies, confirmed the negative relation between moral reasoning and antisocial behaviors (Nelson, Smith, & Dodd, 1990).

Despite the usefulness of measuring the relation of moral judgment to antisocial behavior, it does not inform us about other psychological processes of moral functioning that may be related to behavior. For example, reasoning about moral dilemmas does not reveal whether or not a person centralizes moral concerns or follows through with moral ideals. Therefore, in order to account for this important facet of moral functioning, many have begun to look beyond moral judgment.

Factors other than moral judgment have been noted in studies of antisocial behavior. Tarry and Emler (2007) found that moral reasoning as assessed by the SRM-SF (Gibbs et al., 1992) was not significantly related to self-reported delinquency among young boys aged 12 to 15 years old in London, but instead was negatively related to attitudes toward formal authority and support for moral values. A comparison study of delinquent and nondelinquent high school students showed that delinquent children were more likely to cognitively distort information (e.g., blaming others for their antisocial actions), which in turn led to greater frequency of reported antisocial behaviors (Liau, Barriga, & Gibbs, 1998). In other studies we see cognitive distortions of various kinds. In a study of youth incarcerated for highly antisocial aggressive actions (Slaby & Guerra, 1988), it was found that these youth not only viewed their aggressive actions as more legitimate than other comparison aggressive or nonaggressive youth (moral judgment), they also thought the actions helped their self-esteem and that their victims did not really suffer (moral insensitivity). Another study of 150 young male delinquents from Scotland (Kiriakidis, 2008) showed that delinquents scored higher on moral disengagement (believing that moral standards do not pertain to them) than a community sample. These findings suggest that an examination of components beyond moral judgment is warranted. We examined moral motivation.

As the judgment action gap has received more attention, moral motivation research has increased (e.g., Lapsley & Narvaez, 2004; Narvaez & Lapsley, 2009). Prioritizing moral commitments and actions (over other actions or life concerns) is the crux of moral motivation. As described earlier, moral motivation can be situational and/or more dispositional. In this study we examine dispositional moral motivation. Dispositional moral motivation entails a habitual focus on the moral course of action to the extent that it reflects traitlike characteristics (e.g., being unselfish) and attitudes (e.g., it is important for people to work together to solve problems) that can be measured in a way that is akin to personality assessment. The facets of moral motivation that we address include specifically, ethical identity, unselfishness and citizenship.

Moral Identity. Blasi (1980) proposed that motivation and subsequent action emerge from identity. Moral identity reflects how central moral concerns are to one’s self concept (Blasi, 1980, 1983). People who have a moral or ethical identity prioritize moral commitments and feel obligated to live consistently with their respective moral concerns. When the self-concept is organized around moral virtues, it influences behavior (Aquino & Reed, 2002). Desires are kept in line with moral principles (Hardy & Carlo, 2005). Those who take ownership of moral standards as part of the self are less likely to morally disengage (Paciello, Fida, Tramontano, Lupinetti, & Caprara, 2008). Similarly, Aquino and colleagues (Aquino, Reed, Thau, & Freeman, 2007) have shown that people with strong moral identities are less likely to use moral disengagement maneuvers that endorse the moral acceptability of inflicting harm on others. The positive relationship between moral identity and moral behavior is a central theme found in studies of nominated moral exemplars, or people that have been recognized as having taken part in significant moral action (e.g., Walker & Frimer, 2007), who typically exhibit a merging of personal and moral goals (Colby & Damon, 1992). In contrast, people who commit antisocial behaviors may not prioritize moral goals as central to their self-concept, and as a result, may be more willing to inflict unkind and harmful acts upon others, especially strangers or “outgroup” members. Reed and Aquino’s (2003) study provides support for this line of thinking: participants with stronger moral identities were less willing to harm outgroup members. Therefore we predicted a negative relation between ethical identity and antisocial behavior.

Unselfishness. Many moral actions require that the moral actor give weight to the interests of others, possibly even having to make uncompensated sacrifices. In doing so, the moral actor’s motivation would likely not have self-interested goals as most highly prioritized.² Thus, moral motivation often requires a person to be unselfish. On the flip side, selfishness, which entails narcissism, egotism, and inflated expectations of personal entitlement, is likely to contribute to a wide range of immoral and antisocial behaviors. People who are self-centered and regard themselves as superior often feel entitled to help themselves to the resources of others without regard for their feelings or considerations (Baumeister, Smart, & Boden, 1996; Wink, 1991). This lack of regard for others has been shown to be positively related to violent behavior in particular (Baumeister et al., 1996; Costello & Dunaway, 2003), likely because with such an inflated self-concept, it is easy to feel that one has been insulted or treated with less than adequate respect (Costello & Dunaway, 2003). A person who feels entitled to special treatment is likely to encounter external obstacles in the way that do not permit special treatment to be attained, resulting in a sense of frustration and eventual aggression in order to overcome resistance (Raphling, 1998). For instance, a person may feel that he deserves greater attention in comparison to everyone else in the room, but the attention is divided evenly among everyone. Due to a grandiose self-concept, he is very prone to feel as if he has been treated with disrespect even though he has been treated equally among others. For egotists, even neutral or slightly positive feedback can sometimes be perceived as inadequate and even insulting (Costello & Dunaway, 2003). A review of several empirical studies on gang members, delinquents, and other violent youth consistently show that they are constantly preoccupied with self-deserved respect and self-assertion, and think that they are better than others (Baumeister et al., 1996). This is in strong contrast to what moral motivation entails, which is giving priority to moral values that support the greater good as opposed to selfish, egocentric values. We thus suspected that antisocial behavior will be positively related to selfishness.

Citizenship. The third facet to constitute moral motivation in this study is citizenship. Here, we conceptualized citizenship within a communitarian framework (Jones & Gaventa, 2002), defining it as a demonstration of particular civic virtues (Smith, 1998) such as honesty, trustworthiness, rule following, and conscientiousness. Citizenship involves characteristics that help bind citizens to one another as a community, including prioritizing the common good over individual interests (Jones & Gaventa, 2002). It includes encouraging others to do their share of work and supporting those who follow the rules. In our view citizenship motivates people toward moral action because they understand how their behaviors promote the interests of the community. Conforming to social norms and other aspects of citizenship may interfere with personal interests and impulses. As mentioned earlier, antisocial behavior is associated with greater dishonesty and disregard for social norms (Rowe, 1985). Given that moral motivation entails giving priority to moral values as opposed to other values in the moment (Rest, 1983), we posited that citizenship would negatively predict antisocial behavior.

The current study examined the relations between past antisocial actions and moral variables with a non-antisocial population, college undergraduates, in order to examine which variables are most negatively related to antisocial behaviors. First, because the traditional view is that reasoning drives moral action, we tested how strongly moral judgment was related to antisocial behaviors. We expected a small negative relation between antisocial behaviors and more sophisticated forms of moral judgment. Second, as moral motivation has become of greater interest in predicting moral behavior than moral judgment (Blasi, 1980), we tested how well a group of moral motivation variables (ethical identity, unselfishness and citizenship) predicted antisocial behavior. We expected scores on these variables to be more strongly predictive of antisocial behavior than moral judgment scores.

To measure antisocial behavior, we used the Rowe Delinquency Scale (Rowe, 1985), a measure commonly used with normal samples. For example, research into adolescent personality and behavior commonly measures past antisocial behavior (e.g., Charles & Egan, 2005). So we are in keeping with typical use of the measure. Although there is temporal discontinuity in the focus of the measurement tools (with the delinquency scale focused on the past and the other measures on present views), we do not mean to measure causality, only correlation. Again, this conforms with prior research practices (e.g., Lapsley & Hill, 2010).

We tested populations known for a low frequency of antisocial behaviors, 440 undergraduates from two Midwestern universities in the United States. Although our sample of participants is not ideal for studying frequent antisocial offenders, our focus for this study was on what was negatively related to antisocial behaviors. Additionally, although most of this population was on the low end of antisocial behaviors committed, there was still considerable variance among the participants.

The participants received course credit for participating. Of these students, 52% were female. Ages ranged from 17 to 43 (M = 19.3, SD = 1.79). To measure ethnicity, we asked respondents to mark all categories that applied to them. The breakdown of ethnicity was as follows in decreasing order: White (83.4%), Asian (5.4%), African American (3.8%), Hispanic (3.8%), Other (2.5%), and Indian/Native American (1.1%).

Participants completed a paper survey that included several measures. The participants took the survey individually or in small groups in psychology department laboratory rooms. The time it took for participants to complete the survey ranged from 30 to 50 minutes. Participants’ survey responses were entered into a statistical program (SPSS) for analysis.

The survey contained several scales measuring antisocial behavior, moral judgment and moral motivation.

AntiSocial Behavior. Antisocial behavior was measured with the shorter, 20-item version of the Antisocial Behavior scale (Rowe, 1985), which assesses the frequency of past antisocial behaviors. The revised scale (Lapsley & Duggan, 2001) is shorter and better suited to a non-antisocial population. The acts listed span a variety of activities, including theft (e.g., shoplifting), aggression (e.g., starting a fight), vandalism (e.g., breaking a window), and miscellaneous others (e.g., trespassing, lying about age). For each antisocial act, participants responded with a 4-point Likert scale (1 = no/never, 2 = one time, 3 = several times, 4 = very often), similar to other antisocial behavior measures for both adolescent and adult populations (e.g., Jacobson, Prescott, & Kendler, 2002; Rowe, Maughan, & Eley, 2006). Higher scores indicate greater frequency of antisocial behavior and range from 20-80. Cronbach’s alpha for this sample was .83.

The remaining scales represent seven independent variables that were organized into two larger, more theoretical components: moral judgment and moral motivation.

Moral Judgment. The Defining Issues Test (DIT; Rest et al., 1999) was used to measure moral judgment. The DIT is a paper-and-pencil measure of moral judgment derived from Kohlberg’s theory (Kohlberg, 1984) that measures primarily tacit moral understanding (Narvaez & Bock, 2002), specifically, how one makes judgments about dilemmas concerning justice. After each of six dilemmas, respondents choose the best action for the protagonist and then rate and rank twelve issues in terms of their importance in making the action choice. The rating and ranking of these items determines scores on a range of indices. The scores represent the degree to which a participant uses three schemas. (1) Personal Interests Schema focuses on how a decision affects the welfare of the individual or close associates; (2) Maintaining Norms Schema centers on upholding the status quo of law and order; and (3) Postconventional Schema focuses on the ideals of social justice and cooperation that may or may not be represented in current laws. With age and education, postconventional schema scores increase (Rest et al., 1999; 2000). The validity of the DIT is well established (Rest et al., 1999; Thoma, 2006).

We report scores and correlations for Personal Interest (PI), Maintaining Norms (MN), Postconventional (P score) thinking and N2 (a combination of rejection of Personal Interest items and preference for Postconventional items), but focused on two in our structural equation model: the Maintaining Norms (MN) schema score and the N2 score. These two scores comprise the three main indices of the DIT (N2 is a combination of Personal Interests and Postconventional schemas).

Moral Motivation. Three dispositionally oriented measures were used to determine the extent to which an individual gives priority to moral values. The three scales emerged from the Minnesota Community Voices and Character Education project (Anderson, Narvaez, Bock, Endicott, & Lies, 2003). Participants responded to each measure with a Likert-type scale ranging from 1 (Never Agree) to 5 (Always Agree). The 12-item Citizenship scale (Narvaez, Bock, Endicott, & Lies, 2004) is a self-report assessment of attitudes toward honesty, trustworthiness, rule following, and conscientiousness. A sample item is “It is important to support those who are following the rules.” We summed participants’ response to each question, which created a possible range of scores from 12 to 60, with higher scores indicating stronger citizenship. Cronbach’s alpha was .79. The Selfishness scale (formerly, Prosocial Responsibility Scale; Anderson et al., 2003) is a 5-item scale that measures attitudes toward relating to others in selfish vs. unselfish ways. A sample item is “It is important to take up other people's time for what you want.” Scores for each item were added together, with a total score ranging from 5 to 25, whereby higher scores represent unselfishness and lower scores indicate selfishness. Cronbach’s alpha reliability was .68. Finally, the Ethical Identity scale (Narvaez et al., 2004) was used to measure responsibility and commitment to goodness. It consists of 21 items. A sample item is “Being a good person at school is important to me.” Participants’ response to each question was summed, with total scores possibly ranging from 21 to 105. Higher scores represent a more salient ethical identity. Cronbach’s alpha was .81.

To examine the relations of different variable types to antisocial behavior, we examined the predictor variables by construct (moral judgment and moral motivation). First we conducted correlation analyses. Then we used structural equation modeling (SEM) to examine how both moral judgment and moral motivation predicted antisocial behavior controlling for all variables simultaneously. All significance tests were conducted at a .05 level of significance, and all tests were two-tailed. Table 1 shows the means and standard deviations of key variables including by sex (due to the fact that males on average have higher rates of antisocial behavior). We also examined normal probability plots (Chambers, Cleveland, Kleiner, & Tukey, 1983) and found that antisocial behavior scores for this sample represented a normal distribution.

We conducted correlations on all main variables (see Table 2). Several correlations were expected within construct groups. The moral judgment scores, which are ipsative, were negatively correlated: maintaining norms (MN) with N2, r(440) = -.33, p < .001. The moral motivation measures were positively correlated: Citizenship with Unselfishness, r(440) = .55, p < .001, and Ethical Identity, r(440) = .55, p < .001; Unselfishness with Ethical Identity, r(440) = .40, p < .001. In relation to the dependent variable, the moral judgment variables of primary interest were slightly but negatively correlated with antisocial behavior (maintaining norms, MN: r(440) = -.18, p < .001; N2: r(440) = -.10,p < .05. The measures of moral motivation were more strongly negatively related to antisocial behavior: Citizenship, r(440) = -.42, p < .001, Unselfishness, r(440) = -.37, p < .001, and Ethical Identity, r(440) = -.40, p < .001.

Path Model for Antisocial Behavior.Figure 1 presents our hypothesized SEM model for antisocial behavior, with circles representing latent constructs and squares representing manifest variables. Moral motivation was a latent construct indicated by three variables: Citizenship, Unselfishness, and Ethical Identity. The two moral judgment variables were treated independently due to the fact that they represent distinctive types of thinking. The model predicted that MN, N2, and Moral Motivation directly influenced antisocial behavior.³MN and N2 were allowed to correlate with each other, and with Moral Motivation. The moral motivation variables (Citizenship, Unselfishness, and Ethical Identity) were allowed to correlate with each other as well.

Analyses were conducted through EQS. Due to violations of multivariate normality (Mardia’s Normalized coefficient = 9.92) we estimated the model with maximum likelihood estimation using the Satorra-Bentler scaled chi square (Satorra & Bentler, 1988) to adjust for nonnormality. Using the observed information matrix to obtain more adequate standard errors (Bentler, 2006), the model provided good fit, Satorra-Bentler χ2(6, N = 440) = 4.59, p = .60, CFI = 1.00, RMSEA = .000. Moral motivation was correlated significantly with MN (unstandardized coefficient = 8.34, p < .05, standardized = .17), but not with N2 (unstandardized coefficient = 3.12, ns). MN and N2 were negatively correlated with each other (unstandardized coefficient = -50.84,p < .05, standardized = -.33). All three predictors were significantly related to less antisocial behavior, with moral motivation being the strongest predictor: MN (unstandardized coefficient = -.07, p < .05, standardized = -.13), N2 (unstandardized coefficient = -.06,p < .05, standardized = -.11), and Moral Motivation (unstandardized coefficient = -.84,p < .05, standardized = -.51). Together, moral judgment and moral motivation accounted for 31% of the variance associated with reported past antisocial behaviors.

In this study, we used SEM to examine the extent to which moral judgment and moral motivation predicted self-reported past antisocial actions among a college student population in order to determine which construct is most negatively related to immoral action. Although moral judgment had some predictive power, as hypothesized, moral motivation accounted for the majority of the variance associated with antisocial behavior scores. Moral judgment (as determined by the Maintaining Norms and N2 indices of the DIT) also negatively predicted antisocial behavior, but these relationships were weaker in comparison to moral motivation.

We tested the relation of three moral motivation variables to antisocial behavior: ethical identity, unselfishness, and citizenship. All three of these variables were negatively related to antisocial behavior, independently and when jointly loaded onto the latent construct of moral motivation. Compared to the two moral judgment variables, the moral motivation construct accounted for more variance in self-reported antisocial behavior scores. As explained earlier, people with an ethical (or moral) identity are those who make moral standards and commitment central to who they perceive themselves to be (Blasi, 1980; Paciello et al., 2008). Moral standards that concern others, for example, “respect the property of others,” may compete with personal interests, for example, “grab what you can.”

Unselfishness, entailing relating to others in a less egocentric and more considerate way, was also negatively related to antisocial behavior. As perspective-taking skills become increasingly complex, our understanding of relations between persons also become more complex, which allows us to monitor just how our decisions and actions have an impact on society as a whole (Lapsley, 2006). Perspective taking is a clear staple of moral motivation, as moral motivation entails giving priority to moral interests (including interrelational interests) as opposed to selfish, inconsiderate interests. This may suggest that those who perform more antisocial acts are less likely to consider the perspectives of others. Indeed, a preintervention assessment between a group of delinquent and nondelinquent subjects showed that delinquents were markedly worse in their ability to adopt the role or perspectives of others in comparison to nondelinquent subjects, and also more egocentric (Chandler, 1973). Results from the intervention, however, showed that the delinquent subjects who were placed into a program that provided remedial training in role-taking skills subsequently showed significant reductions in both egocentrism and further delinquent activities than a delinquent control condition.

Finally, citizenship was also negatively related to antisocial behavior. This is perhaps the least surprising of all the findings, given that citizenship is related to what antisocial behavior is not, which is positive, active involvement in one’s community. Indeed, among a large sample of 16-year-old Belgian adolescents asked to choose what characteristics were representative of a good citizen, one of the prominent factors in their choices was civic engagement (e.g., helping people that are worse off and being active in the community) (Dejaeghere & Hooghe, 2009). Dispositional moral motivation entails giving priority to prosocial (not antisocial) values and making them central to the self (Blasi, 1983). This is much different than simply subscribing to moral principles. It cannot be assumed that simply laying out a set of rules for children and adolescents will ensure that they internalize them (Rowe, 2006). In order for moral principles to have influence on anyone’s life, a personal, relational meaning must be attached to them. This may be why moral motivation variables had stronger, negative relations with antisocial behavior than moral judgment variables.

We tested whether moral judgment variables would predict antisocial behavior when moral motivation was included. To do this, we entered two indices of the DIT into our SEM model (Maintaining Norms and N2). The negative relationship between moral judgment and antisocial behavior was significant. This finding echoes previous research relating measures of moral judgment to frequency of delinquent and antisocial actions (e.g., Chien-An & Howitt, 2007; Nelson et al., 1990; Raaijmakers et al., 2005). With increasing moral cognitive development come greater capacities for abstraction and inclusion of other points of view that allow one to better take the moral course of action. Interestingly however, the indices of the DIT, Maintaining Norms and N2, though negatively correlated with each other, both showed a negative relationship to antisocial behavior. We suspect that they negatively influence antisocial behavior for different reasons. For the person who is oriented toward maintaining norms, being antisocial is against the law and could cause dysfunction and loss of order in society. High scores on N2 (Postconventional score plus rejection of Personal Interests) were also negatively correlated with antisocial behavior, but P score was not significantly correlated with antisocial behavior suggesting that the rejection of personal interests as a moral motivation drove the effect, emphasizing further a maintaining norms orientation. In short, individuals who refrain from antisocial or immoral actions may do so for different reasons. So our prediction that antisocial behavior would be negatively related to more sophisticated moral judgment is supported in terms of rejecting personal interests and embracing maintaining norms. In short, within the model, moral judgment explained additional variance beyond moral motivation variables suggesting its importance for preventing antisocial behavior.

One must also consider that moral judgment itself may be influenced by past antisocial behavior. For example, a longitudinal study by Raaijmakers et al. (2005) found that antisocial actions at age 21 influenced moral reasoning at age 23; however, moral reasoning at 21 did not influence antisocial actions at age 23. They concluded that moral reasoning not only affected concurrent antisocial behavior but was affected by previous antisocial behavior. Practicing antisocial behavior may decrease the use of more sophisticated reasoning as time goes on, corroborating Aristotle’s (1988) view that extensive practice influences what virtues or vices one develops.

Although treatments and interventions for delinquency and antisocial behavior exist (see Guerra & Slaby, 1990; Kazdin, 1987; Mulvey, Arthur, & Reppucci, 1993), the ideal would be to prevent them in the first place. The results from our study suggest that prevention education efforts for antisocial behavior should encompass both moral motivation and moral reasoning, perhaps with more emphasis on the moral motivation variables specific to citizenship, unselfishness, and ethical identity. This, of course, is assuming that young people’s moral motivation can be influenced by education. Several studies, briefly summarized below, have shown evidence supporting the notion that targeted programs and educational endeavors can positively influence youth’s citizenship, unselfishness, and ethical identity.

Citizenship, as we have defined it, focuses on civic virtues within a communitarian framework. In line with our definition, citizenship education that may help prevent antisocial behavior could include community-based dialogues and prosocial community-oriented activities for adolescents. For example, an intervention at a high school in Massachusetts consisted of students and teachers meeting for just one hour each week to discuss school-related issues, and perform community service and other related activities (McIntosh, Berman, & Youniss, 2010). Data collected subsequently over four years showed significant increases in community service, political knowledge, and community concern among students. The Just Community approach (Power & Higgins-D’Alessandro, 2008), which brings together students and teachers in a school to discuss the life and discipline of their community using participatory democratic practices, has shown similar effects. Adolescents who were involved in the just community approach were far more involved in civic affairs after they graduated compared to their peers in the same school who were not involved in the program (Grady, 1994, cited in Powers & Higgins-D’Alessandro, 2008).

Citizenship education efforts should not only be school based but also be embedded within the adolescent’s family and family context. For instance, individuals who come from warm and responsive families often take part in community service (Hart, Atkins, & Donnelly, 2006). Adolescents who live within contexts marked by proper modeling of civic behaviors by parents as well as increased social interaction with peers are more likely to become engaged in civic behaviors that have moral relevance later in young adulthood (Zaff, Malanchuk, & Eccles, 2008). Indeed, a study consisting of a large sample of adolescents from seven different countries found that the parents’ emphasis on social responsibility (e.g., being attentive to others and helping less fortunate members of society) strongly predicted civic commitment in every country (Flanagan, Bowes, Jonsson, Csapo, Sheblanova, 1998).

Education aimed at increasing unselfishness, another of our moral motivation variables, could focus on role-taking skills. Chandler (1973) employed drama and the making of video films as a method to bolster role-taking skills in delinquent youth. Compared to two different control groups, Chandler found that his role-taking intervention resulted in reducing both social egocentrism and delinquent behavior. Thus, role-taking appears to be a critically important skill related to selfishness and an important facet of antisocial prevention efforts.

Ethical identity, the last of our moral motivation variables, has received increasing attention from moral psychologists in the last two decades. However, the focus of their attention has been on theoretical and empirical investigations rather than educational interventions. Lapsley (2008), nonetheless, discusses educational implications of the recent work in ethical identity. He argues that the foundation of ethical identity is deeply relational and starts with a strong positive parent-child relationship in early childhood (Kochanska, 2002). Recent research from Hardy, Bhattacharjee, Reed, and Aquino (2010) suggests that parenting continues to play an important role in adolescent ethical identity, with authoritative parenting being positively related to it.

In addition, a young person’s ethical identity appears to be influenced by strong, positive relationships to individuals and groups in contexts other than the family (Lapsley, 2008). Friends, classroom, and school community are all important contexts for the development of ethical identity. Most relevant to education are the classroom and school aspects. Two school-based moral education approaches have shown positive influences on ethical identity. Research from the Child Development Project (Battistich, 2008), a comprehensive moral education approach for elementary schools intended to develop the entire moral person, showed positive changes in intrinsic prosocial motivation and a commitment to democratic values. Both of these effects would seem to be important precursors for prioritizing moral commitments within one’s self-concept. Though the Child Development Project is a wide-ranging moral education approach, its conceptualization of “schools as ‘caring communities of learners’” was a central organizing construct for their program (Battistich, 2008, p. 333). Thus, a primary focus of the project was to create a positive community within both the classroom and school. This educational strategy appears to be important. In fact, reanalysis of the Minnesota Community Voices and Character Education Project data (Narvaez, 2011a) corroborate the importance of a positive classroom climate for developing ethical identity. In pretest-posttest analyses, increases in ethical identity were shown in the participating schools, but only when classroom climate was included as a covariate. Thus, it appears that having a positive, caring classroom climate promotes a student’s prioritization of moral concerns within their own self-concept.

The study has several limitations worth noting and discussing. First, our data is correlational, limiting discussion of causal direction. Second, although we discuss antisocial behavior, our sample was from a group not prone to committing antisocial behaviors, undergraduates (although some of whom may be adolescence-limited delinquents but very few, if any, maintain this course of antisocial behavior; Moffitt, 2006). We used Rowe’s (1985) modified scale (Lapsley & Duggan, 2001) that removed some of the more serious antisocial acts, intentionally making it appropriate for non-antisocial samples. Indeed, given that our sample reported very few antisocial behaviors as it was, the constriction of the range of antisocial acts to only include the minor and moderately serious range had no effect on our scores. Whereas our focus for this study was on what was negatively related to antisocial behaviors, our sample of participants is not ideal for studying frequent antisocial offenders. Our results in fact suggest this, as demonstrated by the low mean antisocial behavior score.

Third, we used self-report measures. Analyses and discussions of self-report antisocial measures have received some attention in the literature, and in fact, these studies have generally shown that these self-report measures have produced acceptable evidence of reliability and validity (Piquero, Macintosh, & Hickman, 2002). One concern that many have about self-report antisocial behavior scales is whether participants are under- or overreporting their antisocial behaviors. However, Huizinga and Elliott (1986) found that self-report errors are less likely to occur with low frequency offenders, as in our sample, compared to high frequency offenders.

It is also important to note that the moral motivation measures were also self-report and therefore may be susceptible to social desirability bias. Participants may have responded in a way that presented themselves in an artificially high positive light when answering questions about their selfishness, ethical goodness, and citizenship. The other factor that might be at play is common method bias, a problem when constructs are all measured in the same way at the same time. Our positive correlations, particularly between antisocial behavior and moral motivation, could be artificially inflated because they were both self-report.

Last but not least, the antisocial behavior measure assessed past behaviors whereas the moral motivation and judgment measures assessed present attitudes, beliefs, and behaviors. The assumption inherent within our analyses was stability of these constructs over time; however, this may not be the case. Participants may have changed over time with respect to antisocial behaviors and/or their moral motivation. Due to lack of longitudinal data, we do not know whether a positive change in moral motivation lead to less antisocial behaviors or whether a decrease in antisocial acts facilitated an improvement in moral motivation. Thus, our model is underdefined and suggests that longitudinal changes in antisocial behavior need to be studied.

These findings should be replicated with a more diverse sample that includes more life-course-persistent offenders. Additional modifications for a replication study would be to include a measure of social desirability and to diversify the assessments such that not all are self-report.

Another important future direction is studying moral motivation and antisocial behavior over time. As noted earlier, moral character and motivation is developed with the help of several aspects of the community (Aristotle, 1988; Narvaez, 2006b, in press; Narvaez et al., 2004; Narvaez & Lapsley, in press). Moral motivation may be dependent upon relatively stable conditions that provide optimal support for its manifestation (Narvaez, 2011b). How and whether ethical motivations are consistent across situations needs further research. Longitudinal research examining changes in individual’s moral motivations from educational interventions, relocation to other places, and life-changing events (e.g., witnessing of terrorist attacks) may provide evidence for whether or not moral motivations remain consistent.

Another important facet of moral functioning that this study did not address is the relationship between specific emotional states and moral motivation. It has been demonstrated that positive emotional states are related to prosocial behaviors (Isen & Levin, 1972) and negative emotional states are related to antisocial behaviors (Hollist et al. 2009). More specific emotions may be important for antisocial behavior such as guilt. Guilt has been shown to be related to moral behavior in that the individual feels responsible to address a moral issue (Basil, Ridgway, & Basil, 2006; Miceli, 1992). Those who frequently commit antisocial acts may not have fully adopted moral standards and are therefore less prone to experience emotions such as guilt for committing them.

Both antisocial and prosocial behavior may be best understood by focusing on multiple components of morality. This entails not just understanding moral judgment but how and why people are motivated toward moral (or antisocial) actions.