Moral disengagement at work

Using moral disengagement, people is able to deactivate their moral standards under certain situations (^{Khan et al., 2021}). Moral disengagement refers to "the way that people cognitively process decisions and behavior with ethical import that allows those inclined to morally disengage to behave unethically without feeling distress" (^{Moore et al., 2012}, p.2). It comprises eight mechanisms to solve the conflict between self-concept and performing inappropriate behaviors (^{Fida et al., 2015}; ^{Kapoor et al., 2021}; ^{Navas et al., 2023}). These mechanisms can be even more, attending to the way in which the mechanisms help to reconstructing morality ("behavior is not morally wrong"), reconstructing agency ("actor is not responsible") or both (^{Schaefer & Bouwmeester, 2021}).

Going deeper with moral disengagement, two different perspectives exists (^{Ogunfowora et al., 2022}a): moral disengagement as a state or as a trait. As a state, it describes the process of reconstruing moral judgments (^{Bandura, 2016}). As a trait, it refers to the predisposition of people to engage the moral disengagement process (^{Zhao et al., 2019}). This distinction is important because, although related, we cannot be ascertained that both constructs have the same determinants, mediators, and consequences. According with ^{Schaefer and Bouwmeester (2021}), researchers have not paid attention to this distinction until now. As a consequence, is difficult to distinguish whether empirical evidence refers to moral disengagement process (state) or propensity to moral disengagement (trait). Bearing this distinction in mind, the present research is focused on the latter, propensity to moral disengagement (PMD).

Recently, a systematic review (^{Newman et al., 2020}), a meta-analysis (^{Ogunfowora et al., 2022}b) and a theoretical review (^{Schaefer & Bouwmeester, 2021}) on moral disengagement has been published. Following these articles, we can outline the individual antecedents, correlates and outcomes of PMD. Among the individual antecedents, PMD receives support for gender (men tend to have higher moral disengagement than women), variables related with leadership performance (e.g., leadership self-efficacy, affective motivation to lead, non-calculative motivation to lead) and personality-based variables like conscientiousness, honesty-humility, moral identity, emotions, and creativity (Schaefer & Bouwmeester, 2021). Dark personality traits, which comprises antisocial behaviors with unique and distinct features (^{Paulhus & Williams, 2002}) remains with an unclear role, because some authors suggest that are antecedents of PMD (e.g., ^{Egan et al., 2015}; ^{Kapoor et al., 2021}) and others that really should be considered as correlates (Ogunfowora et al., 2022b). Regarding the outcomes, moral disengagement is related with different ethical-related constructs like counterproductive behaviors and organizational citizenship behaviors (Newman et al., 2020). The relationship with outcomes that are not moral in nature, however, remains practically unexplored (Ogunfowora et al., 2022b): two studies found that moral disengagement increases turnover intentions (^{Christian et al., 2017}; ^{Huang et al., 2017}), and one that PDM diminishes task performance (Ogunfowora et al., 2022a).

The present study

Although the study of moral disengagement at work has come a long way, there are still many unknowns to be answered that lead to the present study. First of all, we need research that verify which results from prior research refers to moral disengagement as state or as trait (^{Schaefer & Bouwmeester, 2021}). A good example is the relationship with Big Five personality traits, where previous research has shown mixed results, suggesting associations with agreeableness (positive, always the highest), conscientiousness (positive), and neuroticism (negative), meanwhile the role of openness and extraversion is still ambiguous (^{Rengifo & Laham, 2022}).

Secondly, meta-analytic prior research has demonstrated a positive relationship between Dark Triad and moral disengagement (^{Ogunfowora et al., 2022}b), but we do not know what happen if Big Five is also considered. Further research in that sense may help to un understand the relationship among constructs. Additionally, the research on dark personality has growth up including everyday sadism to the components of the triad, turning it into the Dark Tetrad (^{Fernández-del-Río et al., 2020}). Thus, seems necessary research that explore if the relationship between PMD and sadism is the same as the remaining dark personality traits.

Thirdly, more research is needed regarding the non-ethics outcomes of PMD. Job performance is among the most relevant individual variables at the workplace (^{Ramos-Villagrasa et al., 2019}). It is composed by four dimensions that differentiate how workers' behavior contributes to organizational goals in (^{Ramos-Villagrasa et al., 2022}): (1) task performance, the degree in which the worker is able to do their job duties successfully; (2) organizational citizenship behavior or contextual performance, the extra-role behaviors that contributes to the social environment and organization' functioning; (3) counterproductive work behaviors, including all unethical behaviors that reduces productivity or impairs the social environment; and (4) adaptive performance, the capacity of workers to deal with changes in their job and work environment. Previous research had provided strong support for the influence of PMD on contextual performance and counterproductive work behaviors (e.g., ^{Fida et al., 2015}; ^{Zhao et al., 2022}). This is due to both dimensions are ethic-based, whilst the remaining (task performance and adaptive performance) are not. In fact, until the best of our knowledge, they are only one study investigating the relationship of task performance (^{Ogunfowora et al., 2020}) and none with adaptive performance. Thus, our research is focused on the role of PMD as predictor of these two non-ethics dimensions of job performance. The scarce amount of studies focused on non-ethics outcomes suggests that PMD is positively associated with negative outcomes and negatively associated with positive ones (Ogunfowora et al., 2022b). Thus, we expected a negative relationship between PMD and both dimensions of job performance.

Continuing with non-ethics constructs, we want to investigate is work engagement. Work engagement refers to a positive and motivated state of being that is composed by (^{Bakker & Demerouti, 2017}; ^{Reig-Botella et al., 2024}): (1) vigor, which refers to enthusiasm and mental stamina; (2) dedication, the feeling of pride and passion at work; and (3) absorption, the degree in which the worker is completely concentrated in its work. Work engagement is a determinant of several relevant outcomes at the workplace like job satisfaction, task performance, and lesser turnover intentions (^{Mazzetti et al., 2021}). Since PMD indicates a predisposition to break the trust between the person and the organization displaying unethical behaviors, we believe that is difficult for an employee to feel the fulfillment and connection to work suggested by work engagement. Thus, we expect a negative relationship between PMD and work engagement.

Fourthly, as far as we known, Big Five, Dark Tetrad and PMD has been not considered at the same time until now. Previous research supported the consideration of Dark Tetrad in addition to Big Five in the prediction of task performance (^{Fernández-del-Río et al., 2020}) and adaptive performance (^{Ramos-Villagrasa et al., 2020}). PMD showed incremental validity over Dark Triad in the prediction of workplace misconduct and organizational citizenship behaviors (^{Ogunfowora et al., 2022}b), but without considering Big Five, which are antecedents of PMD (^{Rengifo & Laham, 2022}). By investigating the role of PMD over "Bright" and "Dark" personality in the prediction of two different dimensions of performance (task and adaptative) and work engagement, the nomological network of moral disengagement can be clarified and expanded.

Participants and procedure

The sample comprised 405 employees (54% female; M_age = 39.52, SD_age = 13.52) from different organizations. Their average job tenure was 16.63 years (SD = 12.65). Most of them have a permanent job (57.7%), followed by temporary contracts (23.0%) and other forms like internships (19.3%). Following CNO-11 classification, 16.8% of workers belong to catering, personal, and protection services and salespersons; 15.1% are technicians and support workers; 13.6% work in accountancy, administrative and other office jobs; 13.3% to scientific and intellectual technicians and professionals; 12.1% are craftspersons and skilled workers in manufacturing and construction; 10.1% belong to elementary occupations; 9.1% to managerial staff; 4.0% are agricultural, livestock, forestry and fishing sector workers; 3.2% are installation and machinery operators and assemblers; and only 2.7% had a military occupation Regarding education, most participants has vocational training (27.9%), followed by university degree (26.4%), high school (23.7%), primary studies (11.1%) and postgraduate studies (10.9%).

Data for the study were obtained following a non-probability sampling with the aid of trained university students. All workers voluntarily agreed to participate and were informed orally and by written about: (1) study purposes; (2) the type of information to be collected from them; (3) how data would be treated.; (4) that no personal information that allow identification (e.g., names, address, emails) was collected; (5) were ask for the research or its results. The informed consent was obtained orally, because we are not gathering any data that allow identification of participants and the data were analysed anonymously.

Measures

Sociodemographic and work behavior characteristics. We asked participants about their gender, age, and job tenure.

Personality. The Big Five was assessed with the 60-item Spanish version of the Big Five Inventory-2- Short (BFI-2-S, ^{Soto & John, 2017}), rated on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). It has 12 items per personality trait.

Dark Tetrad. We applied the Spanish version of the Short Dark Tetrad (SD4, ^{Paulhus et al., 2021}). This scale comprises 28 items rated on a 5-point Likert type scale, ranging from 1 (strongly disagree) to 5 (strongly agree), with 7 items per dimension.

Propensity to moral disengagement (PMD). We used the 8-item scale developed by ^{Moore et al. (2012}). It is rated on a 7-point Liker scale ranging from 1 (strongly disagree) to 7 (strongly agree). To date, it is considered the best scale to measure PMD (^{Newman et al., 2020}).

Task performance. The scale of task performance from Individual Work Performance Questionnaire (IWPQ), adapted into Spanish by ^{Ramos-Villagrasa et al. (2019}) has been used. It is rated on a 5-point Likert scale ranging from 0 (seldom) to 4 (always).

Adaptive performance. Adaptive performance was assessed with scale by ^{Ramos-Villagrasa et al. (2020}). It has 8 items rated on a 7-point Likert scale, ranging from 1 (totally ineffective) to 7 (totally effective).

Work engagement. Engagement at work has been measured with the Spanish version of the Utrecht Work Engagement Scale (UWES, ^{Schaufeli et al., 2002}). It is comprised by 17 items in a 7-point Likert scale ranging from 0 (never) to 6 (always).

Statistical analysis

Regarding descriptives, we computed means, standard deviations, and reliabilities (α). As this paper is focused on PMD, we also estimate differences between sociodemographical variables (T-test for gender and type of contract, and one-factor ANOVA for occupation and education). Associations between the variables were assessed with Spearman correlations. Regarding regression analyses, three different series of analyses were performed: (1) with PMD as criterion and control variables in step 1 (gender and job experience), Big Five in step 2, and Dark Tetrad in step 3; (2) with non-ethics outcomes as criteria and control variables in step 1 (gender and job experience), and PMD in step 2; (3) with non-ethics outcomes as criteria and control variables in step 1 (gender and job experience), Big Five in step 2, Dark Tetrad in step 3, and PMD in step 4. We did not include age among the control variable due to high association with job experience (r = .89, p < .001, see Table 1).

Table 1. Descriptive statistics, reliability, and correlations of the variables. 

^{Note. a}Coding: Male = 0, Female = 1; Skw = skewness; Kur = kurtosis; α = Cronbach’s alpha. PMD = propensity to moral disengagement.

^{* =}p ≤ .05;

^**p ≤ .01.

Descriptive statistics, reliabilities, and correlations

Descriptive statistics, reliabilities, and correlations among PMD and the remaining variables are shown in Table 1. Given that skewness of psychopathy and PMD is high (2.63 and 2.18) we computed nonparametric correlations. Internal consistency coefficients have values of .64 or higher excepting extraversion, which have a value below standards (α = .55). Therefore, the results related to this variable should be interpreted carefully.

Focusing on associations between the variables of interest (personality, PMD, and non-ethics outcomes), it is remarkable that PMD is associated with all the variables (M_|r| = .26, range (-.29, .40)) excepting extraversion (r = -.04, p = .385) and adaptive performance (r = -.09, p = .095). Focusing on criteria, all of them display similar results: all Big Five personality traits are related with criteria (M_|r| = .26, range (-.23, .42)) excepting between agreeableness and task performance (r = .09, p = .078). Dark Tetrad, however, depends on the personality trait and the criterion. Thus, Machiavellianism is not associated with criteria, Narcissism and sadism with all of them, the first positively (M_r = .15, range (.14, .16)) and the second one negatively (M_r = -.12, range (-.15, -.10)). Psychopathy is associated with task performance (r = -.13, p = .009) and work engagement (r = -.19, p < .001). Regarding PMD, as we stated before is not related with adaptive performance, but is associated with task performance (r = -.16, p = .001) and work engagement (r = -.21, p < .001).

Differences in PMD according with sociodemographical data

Two variables (gender and type of contract) are examined using T-test. They are differences regarding gender, with men scoring higher in PMD than women (Men = 17.22, Women = 15.03; T = 2.99, p = .003, Cohen's D = .299). They are not differences regarding type of contract (Permanent contract = 17.13; Temporary contract = 15.77; T = 1.50, p = .134).

The remaining two sociodemographical variables (occupation and education) are analyzed using one-factor ANOVA. They are differences regarding occupation (F_{9, 95.9}) = 2,85, p = .004). Post-hoc tests allow us to identify those differences are that workers of the agricultural, livestock, forestry and fishing sector scores substantially high than workers of three sectors: managerial staff (T = -3.76, p = .025), scientific, intellectual technicians and professional workers (T = -4.44, p = .007), and technicians and support professionals (T = -7.85, p = .016). They are no differences regarding education.

Predictive models of PMD

Table 2 displays the results regarding the prediction of PMD. All steps are statistically significant. In step 1 (control variables), being male plays a role as antecedent of PMD (β = -.154, p = .002), but is not involved in further models. The same happen with agreeableness (β = -.209, p < .001) and conscientiousness (β = -.127, p = .024) when Dark Tetrad is included (ΔR² = .156). Thus, the final model explains 30.2% of PMD with openness (β = -.151, p = .002), Machiavellianism (β = .274, p < .001), psychopathy (β = .175, p = .002), and sadism (β = .167, p = .004) as predictors.

Table 2. Hierarchical regression with PMD as criterion. 

^{Note. a}Coding: Male = 0, Female = 1; PMD = propensity to moral disengagement.

Predictive models of non-ethics outcomes

Firstly, we computed models using PMD as the only predictor. The third step, concerning the potential moderation between the variable gender and PDM, was not significant in any case. Results, displayed in Table 3, show that PMD contributes to explain 2.4% variance of task performance (β = -.131), 1.6% of adaptive performance (β = -.106), and 3.5% of work engagement (β = -.190).

Table 3. Hierarchical regression with PMD as predictor. 

^{Note. a}Coding: Male = 0, Female = 1; PMD = propensity to moral disengagement.

Secondly, we computed analyses with Big Five and Dark Tetrad, considering PMD in the last step, which are displayed in Table 4. As in the previous predictive models, the incorporation of the potential moderation between the variable gender and PDM (step 5), was not significant in any case. Thus, task performance and work engagement has Big Five, Dark Tetrad, and PMD as predictors. On the other hand, adaptive performance only has Big Five as predictors. Given that the focus of the present article is on PMD, we are going to discuss only models of task performance and work engagement.

Table 4. Hierarchical regression with personality and PMD as predictors. 

^{Note. a}Coding: Male = 0, Female = 1; PMD = propensity to moral disengagement.

Predictive model of task performance explains 26.4% of variance with conscientiousness (β = 0.428, p < .001), Machiavellianism (β = 0.151, p = .004), narcissism (β = 0.122, p = .037), psychopathy (β = -0.119, p = .048), and PMD (β = -0.107, p = .045) as predictors. The incorporation of PMD improves the predictive model by 0.8%.

Predictive model of work engagement explains 22.0% of variance with job tenure (β = 0.116, p = .017), openness (β = 0.17, p = .001), conscientiousness (β = 0.146, p = .010), narcissism (β = 0.191, p = .002), and PMD (β = -0.129, p = .019). Moral disengagement improves the predictive model by 1.1%.

Limitations and future research

The present study has some limitations that should be acknowledged. The cross-sectional design implies that we cannot be ascertain about the causal relationships between variables. However, as the research about moral disengagement in the work context and its influence on organizational results are still in its infancy, we believe that our study is useful to guide further research. Additionally, the use of self-reports in job performance should be considered a limitation also, because tend to be more favourable than supervisor' ratings or objective data (^{Ramos-Villagrasa et al., 2019}). Thus, further research may improve current research using methods different than self-reports. Continuing with limitations, as we mentioned in results, the observed reliability of extraversion is quite low. In consequence, results regarding this variable should be considered with caution. We recommend that extraversion continue to be included in future studies to verify its role in the prediction of criteria.

Besides the aforementioned, we believe that further research should investigate if PMD explains additional variance of the ethics-based dimensions of job performance (i.e., contextual performance and counterproductive behaviors at work) after controlling the effect of Big Five and Dark Tetrad at the same time. Another interesting issue is the role of organizational justice, which can mediate the relationship between PMD and job performance and work engagement. Finally, our study also has found that workers in the agricultural, livestock, forestry and fishing sector workers has higher scores in PMD than other sectors. Researchers may be interested in delving deeper into its relationship in the future to explain this finding and its implications for vocational behavior.