The Protective Role of Empathy and Emotional Self-efficacy in Predicting Moral Disengagement in Adolescents Separated from Illegal Armed Groups

Introduction

In the context of the Colombian armed conflict more than two million children, girls, and adolescents have been victims of displacement and forced recruitment by illegal armed groups with the objective to turn them into instruments of war (Centro Nacional de Memoria Histórica [CNMH, 2017]; Springer, 2012; Fondo de las Naciones Unidas para la Infancia [UNICEF, 2016]). Their incorporation into paramilitary and guerrilla groups limited their opportunities to live out a normal childhood, enjoy their rights, and condemned them to live in illegality and alienation. On having been recovered and separated from the power of the illegal armed groups, the Colombian State recognizes them as passive victims of political violence and war (CNMH, 2017; Congreso de la República de Colombia, 2002). This view of former child combatants nullifies the possibility for them to participate in their dual condition as victims and perpetrators in the uncovering of the facts as a mechanism for reparation and reconciliation with society (Gómez, 2019b; Lugo, 2018).

Awareness of the damage caused and recognition of individual and collective moral responsibility is an important aspect in the process of reintegration of former child and adolescent combatants into society (Gómez, 2019b; Ospina et al., 2018). This requires a comprehensive educational, psychosocial, and right to restorative justice accompaniment for ex-combatants, oriented towards the identification of the emotional vulnerabilities that facilitated recruitment and to identify in them potential psychological resources to socialize and communicate assertively with people (Galván & Durán, 2019). Both aspects are important for the achievement of transitional justice in Colombia in the pursuit of non-repetition, victim reparation, and genuine social reconciliation.

In the psychosocial accompaniment of minors with transgressive behavior, individual or collective, empathy, along with moral emotions, guilt, and shame, are factors that modulate moral authority and inhibit antisocial behavior (Gómez, 2019b; Gómez & Narváez, 2019; Kokkinos & Kipritsi, 2017; Muratori et al., 2017; Tangney et al., 2007; Williford et al., 2015). Moral emotions allow people to anticipate the harmful results of moral transgressions which helps change their behaviors to have a more positive result (Malti et al., 2009). Empathy and emotional self-efficacy have been shown to be psychological factors that can inhibit violent and anti-social behavior in childhood and adolescence, while promoting positive social behaviors (Cohen & Strayer, 1996; Correa, 2017; Hoffman, 2000; Kokkinos & Kipritsi, 2012; Miller & Eisenberg, 1988; Richaud de Minzi & Mesurado, 2016; Valois et al., 2017).

Empathy has been understood as the ability to understand and share someone else’s emotional state (Cohen & Strayer, 1996). Several authors have suggested that empathy is a multidimensional psychological construct in which both cognitive and affective aspects are integrated (Davis, 1983; Mestre et al., 2004). The cognitive mastery of empathy implies the ability to understand or adopt another person’s point of view, while the affective domain is characterized by the ability to share another person’s emotional state or experience feelings of concern or sympathy for others (Arango et al., 2018; Davis, 1994; Hoffman, 2001; Richaud de Minzi, 2014).

On the other hand, emotional self-efficacy is another psychological factor of high relevance to the understanding of both prosocial and antisocial behavior, while playing an important role in the regulation of one’s emotions and moral agency (Bandura, 1990c; Bandura et al., 2003). Emotional self-efficacy refers to the subjective self-assessment of a person’s emotional competence, both to express positive emotions, including satisfaction, joy or liking, and to modulate negative emotions, for example, anger, anxiety, or sadness (Alessandri et al., 2014; Valois et al., 2017; Wang et al., 2018).

Several studies have suggested that empathy is a mediating variable between aggressive and prosocial behavior in childhood and adolescence (Gómez, 2019c; Richaud de Minzi & Mesurado, 2016). Other studies have shown that a deficit in the affective domain of empathy is a predictor of violent and antisocial behavior, while inhibiting the ability to perceive victims’ suffering and consequences caused (Jolliffe & Farrington, 2006; Kokkinos & Kipritsi, 2012, 2017; Muratori et al., 2017; Williford et al., 2015). In addition, high emotional self-efficacy has been found to be a factor regarding the protection of aggression, that reduces the likelihood of antisocial behaviors in the future and promotes prosocial behaviors and psychological adjustment in children and adolescents (Alessandri et al., 2014; Bandura et al., 2003; Caprara et al., 2010; Gerbino et al., 2016; Gómez, 2019a; Mesurado et al., 2018; Valois et al., 2017).

Traditionally, both empathy and emotional self-efficacy have been considered to be directly associated with different types of social behavior, leaving aside the role of moral cognition in understanding transgressive behavior. In this sense, Bandura’s (1990a, 1990b, 1991a, 1991b, 1996, 1999, 2002) proposed theory of moral disengagement turns out to be very useful in studying cognitive-moral processes underlying the relationships between emotional factors and transgressive behaviors.

According to Bandura et al. (1996), individuals socialize, internalize, and build an understanding of the moral standards of their community, which subsequently serves to guide their behavior. Once these moral standards are internalized, most people regulate their actions according to these standards. In doing so, this is satisfactory to them, improves the sense of their self-esteem, and allows them to exercise their moral agency in accordance with the ethical principles established in the social structure (Bandura et al., 1996).

Using affective processes of self-regulation, the emergence of self-sanctions motivate and allow cognitive regulation of behavior that conforms to socially established moral norms. However, these self-regulatory functions only have an impact on actual behavior when activated and, according to the theory of moral disengagement, there is a variety of psychosocial processes that effectively disconnect self-sanctions from inhumane behavior, freeing individuals from self-censorship and possible guilt (Bandura et al., 1996). Consequently, people can selectively activate and deactivate internal control to allow different types of behavior with the same moral standards (Bandura, 2002; Bandura et al., 1996). In other words, it is not necessary to renounce one’s moral standards to transgress them, because, through the process of moral disengagement, it is possible for a person to maintain his/her moral standards while justifying actions that violate those morality patterns, convincing him/herself that his/her standard does not apply to a particular situation or person.

The theory of moral disengagement postulates that moral division can occur through eight different mechanisms, which allow subjects to disconnect moral censorship and thus commit inhuman acts, isolating negative feelings, associated with acts that violate the moral canons of a society. Bandura (1999, 2002, 2016) described these mechanisms, grouping them into domains: (1) cognitive restructuring of immoral behavior, implemented through the use of moral justification mechanisms, palliative comparison, and euphemistic labelling; (2) minimization of the role of the agent, facilitated by the mechanisms of displacement of responsibility and diffusion of responsibility; (3) misrepresentation of harmful consequences, through the use of mechanisms of minimizing, ignoring or misconstruing the consequences of the action itself; and (4) blaming the victim, using the mechanisms of dehumanization and attribution of the blame for the harm caused to the abused person (see Figure 1).

Figure 1

Domains and Mechanisms of Moral Disengagement.

Taken from Bandura (1999, p. 194).

The research undertaken by Bandura et al. (1996), Bandura et al. (2001), Barchia & Bussey (2011), Paciello et al., (2008), Pelton et al., (2004), and Pornari & Wood (2010) have shown that there is a positive relationship between the use of moral disengagement and aggressive and antisocial behavior, and a negative association between prosocial behaviors, empathy, and moral disengagement (Gómez, 2019a; Gómez & Narváez, 2019).

Additional studies have shown that emotions are an important source of motivation for cognition and moral conduct (Eisenberg, 2000; Haidt, 2003; Hoffman, 2000) and it is the emotions themselves that make us react to rape and the perpetration of inhumane acts (Haidt, 2003). A review of the literature on the study of emotions highlights that empathy, sympathy, guilt, and shame are predictors of prosocial moral reasoning in addition to the existence of a positive association between empathy and prosocial behavior (Eisenberg et al., 2010; Eisenberg et al., 2006; Gómez & Narváez, 2019; Hoffman, 2000; Olthof, 2012) and a negative relationship between empathy and aggression (Carlo et al., 2010; Gómez & Narváez, 2019; MacEvoy & Leff, 2012).

In this sense, emotional factors, including shame, guilt, affective empathy, self-efficacy to express positive emotions, and modulating negative emotions, have an important influence on moral behavior. Hence, selective deactivation of these factors of emotional self-regulation plays an important role in the use of moral disengagement from morally unacceptable behaviors (Doyle & Bussey, 2017; Giulio et al., 2018; Moore, 2015). However, moral disengagement has been studied mainly on the basis of the mediating or predicting role of aggressive and antisocial behavior, and there is little information regarding the role of empathy and emotions in understanding cognitive mechanisms of moral disengagement, especially in a population as particular as teenagers separated from illegal armed groups.

Regarding the aforementioned, emotions and the ability to govern them are fundamental in how subjects feel about themselves and how they relate to others (Eisenberg et al., 1995; Eisenberg et al., 1997), including self-regulation of emotions that provoke anger, anxiety, frustration, and sadness (Caprara & Gerbino, 2001; Gómez, 2019b; Lewis et al., 2006; Valois et al., 2017).

According to Cole et al. (2004), emotions are inherently regulating behaviors, but the process of regulating emotions varies by context. Poor regulation of emotions, especially the most negative ones, predicts problems regarding control of aggression, behavioral problems, substance abuse, and antisocial behaviors (Eisenberg et al., 1995, 1996; Frick et al., 2003; Frick & Morris, 2004). Likewise, the absence of moral emotions such as guilt and empathy are related to cruelty, manipulation, deception, and behavioral disorders in childhood and adolescence (Frick et al., 2003).

Guilt and shame are relevant emotions when experiencing subjective unrest when an immoral, socially unacceptable action has been taken (Tangney et al., 1996). These emotions involve cognition of self-awareness, self-understanding, and understanding of moral standards (Eisenberg, 2000; Lewis, 1992). Therefore, these emotions are necessary to accept responsibility for the acts performed that caused others harm and the emergence of repentance and the desire to amend the harm caused. These are essential conditions for the achievement of the objectives of all restorative justice in the context of psychosocial intervention with adolescents who are lawbreakers.

The conditions of forced recruitment of these young people, coming from rural contexts – where the armed conflict was lived with extreme harshness, in conditions of poverty and combined with a lack of opportunities – involves the importance of exploring factors determining their participation in inhuman acts and the psychological mechanisms involved in the moral disconnection with their victims, in the context of war.

It is important to mention that many previous studies on moral disconnection have focused on explaining violent behavior, especially in youth populations with criminal, antisocial, and school bullying behaviors, within school, urban, and deprivation of liberty contexts. However, while there are investigative precedents on the role of moral disengagement, and particularly the mechanism of dehumanization, in the understanding of war, terrorism and war behaviors (Bandura, 2004, 2016; Blanco et al., 2020; Villegas de Posada et al., 2018), studies are scarce regarding rural adolescent populations that break the law as a result of forced recruitment by illegal armed groups, and who are forced to commit inhumane acts contrary to all social morality and that constitute in front of jurisprudence punishable crimes and even human rights violations.

After their disengagement from these illegal groups, and in specialized psychosocial care processes, ex-combatant Colombian rural adolescents show emotional discomfort, positive affections, and empathetic concerns in describing their participation in war actions that under other conditions of existence they probably would not have committed as a personal decision (Gómez, 2019b). In this sense, this work is a contribution to the psychosocial processes and programs of re-education and reintegration of young ex-combatant rural youths with favorable prognosis for the restoration of their rights and positive participation in social life.

The objective of this paper was to analyze the effects of empathy and emotional self-efficacy in predicting moral disengagement in a group of adolescents disengaged from illegal armed groups who are under the protection of the Colombian Institute of Family Welfare (ICBF) for the restoration of their rights. Based on the results obtained in previous research and theoretical conceptualization, two hypotheses were put forward:

H1: Empathy and emotional self-efficacy present significant correlations, of a negative sign, with total moral disengagement and its socio-cognitive mechanisms.

H2: The dimensions of empathy and emotional self-efficacy present negative effects in the prediction of total moral disengagement and its socio-cognitive mechanisms.

Method

This is a quantitative approach study, with non-experimental cross-sectional design, and explanatory scope.

Participants

The population was composed by 35 adolescents belonging to the specialized attention program for the reestablishment of rights of boys, girls, and adolescent victims of illicit recruitment who dissociated themselves from organized armed groups outside the law, located in the modality of Tutorial Home of the Colombian Institute of Family Welfare (ICBF) in the city of Manizales, Colombia. Given the small number of adolescents located in this scheme and being the only one in the department of Caldas, the total population was taken, consisting of 19 men (54.3%) and 16 women (47.5%).

The ages covered ranged from 12 to 18 years old, of which 60% (n = 21) were in the range of 12 to 14 years old and the remaining 40% (n = 14) between 15 and 18 years old. The average age was 16.34 years (SD = 2.02).

The origin of the population that was part of the study was quite heterogeneous; however, the departments where the most teenagers came from were Chocó (25.7%) and Antioquia (17.1%). Similarly, it is Chocó department from which most of the people in the research were recruited by illegal armed groups (28.6%).

As for the armed group to which the adolescents belonged, 45.7% were recruited by the ELN, 40% by the FARC, and the remaining 24.4% was distributed among BACRIM, ERG, and EPL. The time spent in these groups was quite variable (M = 23.7) due to some atypical figures for youths who stayed for more than 48 months (4 youths), but for the most part, about 69%, they stayed for up to 26 months.

Following disengagement of the armed group, teenagers entered ICBF rights restoration program in the Tutor Substitute Home modality. On average, teenagers stayed in the program for about 18 months and, like the time spent in the armed group, great variability is observed (SD = 22.21).

Instruments

Mechanisms of Moral Disengagement (MMDS; Bandura et al., 1996). It is a Likert-type questionnaire containing 32 items distributed in eight subscales, with five response options (from 1 = strongly disagreed up to 5 = totally agree), to measure inclination to use the mechanisms of moral disengagement and the effect on aggressive behavior. The Spanish-validated version of Rubio-Garay et al. (2017) was used, which retains the factorial structure of eight mechanisms, as proposed in the original version. Subscales evaluate the eight mechanisms of moral disengagement: moral justification (“It is right to use force with those who offend your family”), euphemistic labelling (“Giving nudges is just a way to joke”), palliative comparison (“It’s not serious to insult a partner, it would be much worse to hit them”), displacement of responsibility (“Young people who do not receive a proper education cannot be blamed for bad behavior”), diffusion of responsibility (“You can’t blame a gang member for the damage caused by the gang”), distortion (minimizing, ignoring, or misconstruing) of consequences (“Insults among friends are harmless”), attribution of blame (“People who neglect their things are to blame if they are stolen”), and dehumanization of the victim (“Some people deserve to be treated as animals”).

Different studies have reported internal consistency rates ranging from .82 to .93 for the total scale and the subscales (Bandura et al., 1996; Gini et al., 2014; Gómez et al., 2019; Hardy et al., 2015; Paciello et al., 2008).

Interpersonal Reactivity Index (IRI; Davis, 1983). It is a multidimensional Likert scale with five response options, from 1 (does not describe me at all) to 5 (describes me very well) designed to measure the cognitive and affective factors of empathy in four dimensions: perspective taking, fantasy, empathetic concern, and emotional discomfort. The scale consists of 28 items. Three of the four dimensions were used for this study: empathetic concern evaluates empathetic responses mediated by emotional factors, in terms of feelings of warmth, compassion, and concern for others (“I often have tender, concerned feelings toward people less fortunate than me”); capture of perspective refers to the tendency to spontaneously assume others’ point of view, that is, an empathy mediated by cognitive factors (“I often try to better understand my friends by imagining how they see things (putting myself in their place”); and personal discomfort measures personal feelings of anxiety and discomfort that arise from observing someone else’s negative experience (“When I’m in an emotionally tense situation, I get scared”). The Spanish version of the IRI was validated in Spanish with a sample of 1,285 adolescents, obtaining acceptable reliability coefficients for the full scale and subscales (Mestre et al., 2004).

Regulatory Emotional Self-Efficacy (RESE; Caprara & Gerbino, 2001; Caprara et al., 2008). This is a 32-item questionnaire designed to evaluate perceived self-efficacy in the management of negative and positive emotions. Each item is evaluated on a Likert scale of five response options, i.e., 1 = incapable, 2 = slightly capable, 3 = average capability, 4 = very capable, and 5 = completely capable. The dimensions evaluated are negative emotion management (NEG), which measures beliefs about a person’s ability to regulate negative emotions appropriately (“Are you able to overcome frustration if others don’t appreciate you the way you want?”), and positive emotion expression (POS), measuring beliefs about the ability to express positive emotions (“Are you able to express happiness when something good happens to you?”).

This instrument has been validated in adolescent population in Italy, the United States, and Bolivia, showing a similar factorial structure in all three countries (Alessandri et al., 2014; Caprara et al., 2008). In general, partial invariance was found at both metric and scalar levels in all genders and countries, and the reported internal consistency rates ranged from .70 to .97.

Ethical Aspects

In accordance with Law 1090, 2006, Resolution 008430, 1993, and Law of Childhood and Adolescence 1098, 2006, this study obeys ethical principles regarding respect, privacy, and dignity, ensuring confidentiality and anonymity of participants, as established in Article 26 and 50 of Law 1090. It was endorsed by the Protection Directorate of the Colombian Institute of Family Welfare (ICBF) at national level and the regional center in Manizales. Additionally it had the informed consent of adolescent participants and the endorsement of the ethics committee of the Luis Amigó Catholic University (Colombia).

Procedure and Data Analysis

Application of questionnaires and interviews was carried out individually in the facilities of the service institution, Asociación Mundos Hermanos, and in some cases in guardian substitute households. Each session lasted one hour, with 10-minute breaks between tests.

SPSS version 25 statistical package was used for the analysis (IBM Corporation, 2017b). An internal consistency analysis of the scales and subscales was performed, using omega coefficient (ω). Omega coefficient, unlike alpha coefficient works with factorial loads (Gerbing & Anderson, 1988). This allows greater stability in calculations for multidimensional scales, since it does not depend on the number of items (McDonald, 1999). Subsequently, a univariate descriptive analysis of means and standard deviations of instruments applied was carried out. The Kolmogorov-Smirnov test was applied, showing that variables have a normal distribution. Subsequently, a correlation analysis was carried out using Pearson’s r coefficient. Additionally, a multiple linear regression analysis was performed using the forward successive step entry method. Finally, a structural equations model was used to determine total, direct, and indirect effect between variables considered in this study. For modeling structural equations Amos version 24.0 software (IBM Corporation, 2017a) was used.

Table 1

Descriptive Statistics on the Scores Obtained by the Sample of Adolescents Disassociated from Armed Groups

Note. Own elaboration; ω = Mcdonald’s omega coefficient.

Results

Table 1 presents descriptive analyses of study variables. In moral disengagement, it was found that the mechanisms of attribution of blame (M = 2.22), followed by moral justification (M = 2.21) and the distortion of consequences (M = 2.20), present the highest mean values. In empathy, the emotional factor of empathetic concern had the highest mean value. In emotional self-efficacy, the variables of negative affection management presented the lowest scores and expression of positive affection the highest values.

Table 2 shows the results of correlation analysis of the total scores of the study variables, using Pearson’s r. Statistically significant negative correlations were found between moral disengagement and emotional self-efficacy (p = .014). Although the correlation between moral disengagement and empathy was negative, it was not statistically significant. Emotional self-efficacy correlates positively with empathy (p = .002).

Table 2

Correlations between the Total Scores of Variables of Moral Disengagement, Empathy, and Emotional Self-efficacy

Note. Own elaboration.

*p < .05 (bilateral), **p < .01 (bilateral).

Table 3 presents correlations between the mechanisms of moral disconnection and the dimensions of empathy and emotional self-efficacy. In terms of empathy, the use of perspective correlates negatively with moral justification, euphemistic language, displacement of responsibility, and distortion of consequences. The emotional factor of empathic concern presents a unique and significant negative correlation with de-humanization.

Table 3

Correlations between the Mechanisms of Moral Disengagement and the Dimensions of Empathy and Emotional Self-efficacy

Note. Own elaboration. TP = perspective-taking; PE = empathic concern; MP = personal discomfort; AEN = self-efficacy for handling negative emotions; EMV: self-efficacy for moral emotions: shame; AEP = self-efficacy for the expression of positive emotions; AE = empathic self-efficacy; Hum = humor; JM = moral justification; LE = euphemistic labelling; CV = palliative comparison; DR = displacement of responsibility; DIFR = diffusion of responsibility; DC = distortion of consequences; AC = attribution of blame; Des = dehumanization.

*p < .05 (bilateral),**p <.01 (bilateral).

Self-efficacy for the expression of positive emotions correlates negatively with advantageous comparison, diffusion of responsibility, distortion of consequences, and dehumanization. Empathic self-efficacy presents a single significant negative correlation with the diffusion of responsibility. Variables of personal discomfort and mood did not report significant correlations with mechanisms of moral disconnection.

Table 4 shows the results of the multiple linear regression analysis, in order to detect the factors with the greatest predictor effect on moral disengagement mechanisms, with empathy and emotional self-efficacy being predictor variables. The fact that the beta coefficient (β) of predictor variables is negative indicates that they are variables that reduce the likelihood of moral disengagement in adolescents detached from armed groups in this study.

Table 4

Ineffectiveness of Empathy and Emotional Self-efficacy about Moral Disengagement

Note. Own elaboration.

Self-efficacy for the management of negative emotions was found to decrease the probability by 21% of total moral disengagement. Likewise, total emotional self-efficacy and self-efficacy to express positive emotions explain 28% of the variance of the moral justification mechanism. Perspective, which is a cognitive dimension of empathy, decreases the likelihood of euphemistic labelling by 11%, and displacement of responsibility and distorting consequences by 20%. Self-efficacy for moral emotions, specifically shame, reduces the probability of palliative comparison by 15% and dehumanization by 24%. Finally, empathetic self-efficacy reduces probability by 24% of diffusion of responsibility. Independent variables had no effect on the attribution of blame (see Table 4).

To establish total, direct, and indirect effect of empathy and emotional self-efficacy on total moral disengagement, a model of structural equations was estimated through the weighted least squares method, as it provides consistent and less biased estimates with relatively small sample sizes (Byrne, 2016).

The fitting criterion test was made taking into consideration the values of chi square (c²), incremental fit index (IFI and CFI), fitting criterion index (GFI), normed fit index (NFI), root of the average quadratic residue of approach (RMSEA), and the root of the average quadratic residue (RMR). The obtained model presented a good absolute adjustment, c²(1) = 0.841, p = .40, and presented a good comparative adjustment, which is deduced from the fitting criterion index which is deduced from the comparative fitting criterion index (CFI), the normalized adjustment index (NFI) and the Tucker-Lewis index (TLI) (CFI = 1.0 > .90, NFI = .997 > .90, IFI = 1.079 > .90, TLI = 1.805 > .90). It also shows a good fitting index (GFI) and its corrected correspondent (AGFI), as well as an optimal root mean square residue (RMSEA) value (GFI = .999 > 0.90, AGFI = .994 > .90, RMSEA = .000 < .08). The results show that the proposed model presents an adequate adjustment to the data (McArdle & Nesselroade, 2014).

Figure 2

Empirical Model of Structural Equations: Self-efficacy for the Management of Negative Emotions as a Mediating Variable of Moral Disengagement.

Note. Own elaboration. *p < .05, **p < .01.

For the structural equation model, self-efficacy for the management of negative emotions was taken as a mediating variable between empathy and moral disconnection (see Figure 2). The results show that empathy and self-efficacy for the management of negative emotions explain 22% of the variation in moral disconnection (R² =.22, CI 95% [0.016, 0.570], p = .01), whose total, direct, and indirect standardized effects are negative. Likewise, it was found that self-efficacy for the expression of positive emotions is a predictor of self-efficacy for the management of negative emotions (R² = .46, CI 95% [0.233, 0.730, p = .01).

Table 5 shows the total, direct, and indirect standardized effects of study variables. It was found that the effects of empathy, self-efficacy for the expression of positive emotions, and management of negatives over moral disengagement are all negative.

Table 5

Standardized Total, Direct, and Indirect Effects of Predictor Variables on Response Variables

Note. Own elaboration.

*p < .05, **p < .01.

Discussion

The objective of this study was to analyze the role of empathy and emotional self-efficacy in the prediction of moral disengagement in a group of adolescents disengaged from illegal armed groups who are under the protection of the Colombian Institute of Family Welfare-ICBF for the restoration of their rights.

In general terms, the proposed hypotheses were partially supported. In relation to the first hypothesis (H1), only total emotional self-efficacy presented a negative association with total moral disconnection and only some dimensions of empathy and emotional self-efficacy were significantly and negatively correlated with the socio-cognitive mechanisms of moral disconnection. The second hypothesis (H2) was also partially supported, given that only the cognitive factor of empathy, self-efficacy in handling negative emotions, and shame showed significant and negative effects on total moral disconnection and socio-cognitive mechanisms.

Correlation and predictive analyses revealed that emotional self-efficacy is negatively associated with moral disconnect, indicating that the capacity for control over one’s own emotions, both positive and negative, plays a protective role against the use of moral disconnect. These findings are consistent with several studies which have shown that a poor ability to regulate negative emotions is a predictor of psychological maladjustment, behavioral problems, aggression, and antisocial behavior in adolescence (Bandura et al., 2003; Caprara et al., 2010; Carlo et al., 2010; Eisenberg et al., 1995; Eisenberg et al., 1996; Frick et al., 2003; Frick & Morris, 2004; Richaud de Minzi & Mesurado, 2016).

Adolescent ex-combatants present a greater self-informed capacity for the expression of positive emotions, an aspect that protects them against the use of moral disconnection mechanisms, especially those that spread responsibility, distort consequences, and dehumanize. However, it is worth noting that these young people have less capacity for self-efficacy in the handling of negative emotions, which can be explained by the conditions of psychosocial vulnerability to which they were exposed, including physical and psychological subjection by armed groups, legitimization of social violence, exposure to war and armed confrontation, and poverty in rural contexts, among other factors, which have been reported in various studies and reports on the recruitment of children and adolescents in the Colombian armed conflict (CNMH, 2013, 2017; Gómez, 2019a, 2019b; Human Rights Watch, 2004; ICBF, OIM, & UNICEF, 2014); Lugo, 2018; Springer, 2012; Valencia-Suescún et al., 2015).

Consistent with the above, Cole et al. (2004) suggested that emotions are inherently regulating social competence, from childhood to adulthood, but the development of skills for the agency and regulation of those emotions vary according to the sociocultural context and experiences of social modeling. In this sense, the experience of adolescents in illegal armed groups has an ideological background in which the exercise of violence imposes ways to feel and reason according to the dynamics of war (Blanco et al., 2020; Gómez,2019b; Villegas de Posada et al., 2018), while the process of social reintegration, through experience in state protection environments – substitute families – seeks, justly, to break war ideologies that involve moral disengagement and enable a connection with the social bond, based on the recognition of the other (Gómez, 2019b; Gómez & Narváez, 2018).

In particular, it is evident that self-efficacy in the management of negative emotions plays a protective role in predicting moral disengagement in general and the mechanism of moral justification, while presenting negative correlations with euphemistic labelling, spreading responsibility, distorting consequences, and dehumanization. The model of structural equations showed, consistent with the findings already reported and the second hypothesis (H2), self-efficacy in the management of medium negative emotions between empathy and moral disengagement. Likewise, self-efficacy in the expression of positive emotions had a significant indirect effect on moral disengagement. These findings are consistent with various studies with offending adolescents, highlighting the importance of emotional self-regulation, empathy, recognition of the experience of others in mitigating violent, and antisocial behavior (Giulio et al., 2018; Muratori et al., 2017; Petruccelli et al., 2017). Other studies have found that a deficit in the capacity for empathy and sensitivity for the well-being of others is a predictor of moral disengagement and increases the likelihood of antisocial behaviors in the future (Arango et al., 2018; Kokkinos & Kipritsi, 2017; Miller & Eisenberg, 1988; Muratori et al., 2017).

On the other hand, self-efficacy in regulating moral emotions, specifically guilt and shame, is a factor negatively associated with the displacement and diffusion of responsibility, while playing a protective role in predicting advantageous comparison and dehumanization. This is particularly because the capacity for moral emotions involves the recognition of the behaviors exercised towards others, allowing the person to take blame and shame, without having to resort to using moral disengagement as a strategy to avoid self-censorship. In this respect, the study by Tangney et al., (1996), which explored the relationship between the propensity to shame and guilt and constructive and destructive anger responses in children, adolescents and adults, found that the tendency to experience shame is related to dysfunctional responses associated with anger, direct and indirect aggression, and self-directed hostility in all the groups studied.

In line with the above, the study by Bandura et al. (2001), which analyzed the socio-cognitive self-regulation mechanisms that regulate transgressive behavior, proposed an empirical model of structural equations in which social effectiveness and regulatory self-efficacy or self-regulatory capacity in children and adolescents present a negative effect on moral disconnection and transgressive behavior, while self-efficacy was positively associated with prosocial behavior. These findings confirm the importance of self-regulation of emotions in moral agency and its protective effect in preventing transgressive behavior in the processes of social reintegration in adolescents disengaged from armed groups who were part of this study.

Another important finding is the role played by the cognitive domain of empathy, specifically perspective taking, in understanding the moral disconnect in the adolescents studied. The findings show that perspective-taking has a negative predictive effect on the mechanisms of euphemistic language, displacement of responsibility, and distortion of consequences. Likewise, empathic concern presented a significant negative correlation with dehumanization. These findings indicate that the ability to adopt the point of view of others and recognize their personal experience is a protective factor for the adolescents in this study, in the face of the use of moral disconnection mechanisms.

The results obtained find legitimacy in Bandura’s (1999, 2002, 2016) explanations in that moral disconnection implies changing the perception of inhuman or antisocial behaviors and damages they may cause to others in order to avoid emotional discomfort and moral censure, which implies, consequently, an exercise of de-legitimization and disengagement from the suffering of others. These results have important implications for psychosocial theory and practice with this population in that they recognize the importance of emotions and the recognition of others’ experience in the regulatory cognitive processes of moral agency. Also, it corroborates the importance of empathy in the processes of psychosocial accompaniment for the social reintegration of these adolescents and the prevention of moral disconnection, which affects prevention of recidivism in criminal activities.

Similar studies with adolescents detached from armed groups in Colombia have shown that empathy, both in their cognitive and affective domain, is a predictor of prosocial behavior and management of negative emotions has been associated with altruistic prosocial motivation (Gómez 2019a, 2019c; Gómez et al., 2019). Other studies with adolescent survivors of armed confrontation have shown that companionship, social support, and tendency to prosocial behaviors is related to the reduction of symptoms of stress and emotional discomfort generated by war (Gómez, 2019b; Haroz et al., 2013; Taylor et al., 2018), which has a positive impact on the processes of reintegration into social and family life. Finally, the positive correlation between emotional self-efficacy and empathy found in this research is consistent with what is stated in other studies in which results highlight that empathy and positive emotions are predictors of prosocial behavior (Eisenberg et al., 2010; Eisenberg et al., 2006; Gómez & Durán 2020; Hoffman, 2000; Olthof, 2012).

Conclusions

The empirical evidence provided by this research allows us to conclude that emotional self-efficacy and empathy decrease the probability of the appearance of transgressive behaviors and harm towards people. However, linking adolescents to experiences of direct violence negatively impacts their ability to regularly and clearly recognize their own emotions and have control over them. Emotional suppression and dehumanization of the enemy are the greatest training strategy for war, suppressing in the minors recruited the ability to empathically understand others’ emotions.

Moral performance is not only based on the knowledge of the concepts of good and evil or on the conventional moral duties. Moral judgments and acts require empathic involvement, which is essential for social coexistence and personal well-being. Therefore, the psychosocial accompaniment of adolescents disengaged from illegal armed groups must be projected towards the promotion of emotional competencies, which links the awareness of one’s own and other people’s emotions with principles and the display of empathic attitudes, oriented towards prosociality.

Limitations

The study has several limitations, the first of which is the small sample size. This is due to the difficulty in accessing a population with these characteristics. Access to young people detached from armed groups for research is restricted by government institutions for security reasons, so this study was supported by the Colombian Institute of Family Welfare. A second limitation is the cross-cutting design of the study, which restricts the explanatory possibilities regarding the impact of experience in armed groups and social reintegration measures on study variables.