1. Introduction
This study examined correlates of the concept of Employee Disenchantment (ED) (Furnham & Treglown, 2018; Treglown & Furnham, 2022). Furnham (2015) suggested employee disenchantment, the opposite of job satisfaction or engagement, had five different facets or factors: organisational lying and hypocrisy, perceived inequity, bullying and mistreatment, distrust and broken promises. These factors formed a coherent concept and a validated self-report measure of them was developed (Treglown & Furnham, 2022). It was argued that disengagement explains why people at work go from being engaged to disengage; productive to subversive; and a friend to an enemy of the organisation.
For our purposes ED has five distinct by related factors: Organisational Hypocrisy where companies do PR about their mission, vision, methods, as well as their integrity and transparency, customer and employee care, welfare, etc. Yet employees are aware that this is untrue. Second, perceived inequity which occurs when some people in the organisation are treated very differently from others. The feeling people are being unfairly held back while a few succeed can stimulate a great deal of resentment. Third, bullying and mistreatment which is the belief that some senior people are callous, uncaring, and manipulative, and that many are victim. Further, some organisations have a management style that is essentially aggressive and Machiavellian. It is then not only the oversensitive type who buckles under the acute and chronic bullying that leads from disenchantment to the need for revenge. Fourth, distrust which is the feeling that the organisation does not even trust its own employees. It may have put in place a number of devious and not-admitted (often electronic monitoring) systems to spy on its own people. Whilst top management may talk about, and demand, loyalty from their staff, it is clear that they do not trust their own employees. Fifth, broken promises which is all about expectations not being met, and assurances about such things as training and promotion being ignored.
This study looks at correlates of this measure, particularly its relationship to Counterwork Behaviours (CWBs) which have often been described as a function of both environmental and individual difference factors (Penney, Hunter, & Perry, 2011). CWBs have been divided into Interpersonal Deviance occurs when misconduct target(s) specific stakeholders such as co-workers and Organizational Deviance which encompasses production and property deviance.
In this study, we focus on individual difference correlates of different but related measures of CWB focusing on both bright-and dark-side personality variables as well as demographic factors and moral disengagement. The primary interest however is in ED and whether this factor has incremental validity over the above factors in explaining CWBs. Further, to look at the reliability of our findings we use different measures of CWBs.
Many studies have looked at the role of the “bright” personality in CWB. Meta-analyses have noted that Neuroticism, Agreeableness, and Conscientiousness are strongly related to both interpersonal and organisational deviance (Berry, Ones, & Sackett, 2007). Other researchers have investigated “dark-side” individual differences in CWB by focusing on the propensity to “misbehave” at work; the “Bad Apples” hypothesis (Furnham & Taylor, 2011). This approach suggests that there are specific “dark” traits that increase the likelihood of CWB because of the inherent malevolence associated with them. For instance, an important meta-analysis found that the Dark Triad of Psychopathy, Narcissism, and Machiavellianism (Paulhus & Williamson, 2002) was significantly related to increased CWBs (O’Boyle Jr. et al., 2012). However, DeShong et al. (2015) argued that extremely low levels of Big Five personality traits are enough to account for the variance in interpersonal and organisational CWBs. Despite disagreement over whether “bright” or “dark” traits are more important, there is clear evidence that individual differences help account for why employees engage in CWB.
Beyond bright and dark personality, many studies were dedicated to the role of moral disengagement in explaining CWB. Moral disengagement depicts the socio-cognitive dissociation of deviant acts from the negative aspects (i.e. guilt) that should deter the actor from engaging in said acts (Fida et al., 2015). Moral disengagement has been considered a form of cognitive distortion that works in three broad ways, helping individuals view their deviant acts and the negative consequences as morally and socially appropriate. The first set of socio-cognitive mechanisms facilitate the restructuring of behaviours to appear less unethical (Bandura, 1991), i.e. labelling the action with euphemisms or advantageously comparing the action with other more adverse action. The second group of mechanisms work by minimising the role of the act, i.e. displacing or diffusing responsibility on colleagues or supervisors. The third type of mechanisms aims to distort the effects of the action, i.e. dehumanising the victim or minimising the perceived consequences. Previous research has demonstrated that moral disengagement mediates the relationship between negative affect experienced in response to job stressors and organisational and individual-directed CWB (Fida et al., 2015). Therefore, it is important to understand how disenchantment relates to moral disengagement in relation to various forms of CWB.
2. Disenchantment, Bright and Dark Side Personality
“Bright”, or normal, personality encompasses the behaviours we engage in when we are being purposeful and positive (Kaiser, LeBrenton, & Hogan 2015). The Five-Factor Model (FFM) is the most comprehensive taxonomy of bright personality. This study expects two of the FFM traits to be related with disenchantment, namely Agreeableness and Neuroticism (H1; H2). Agreeable individuals are predisposed to be more trusting, forgiving, optimistic, and tolerant (Thompson et al., 2005; McCrae & Costa, 1987). Disagreeable individuals, however, are prone to cynicism, anger, and mistrust (Ozer & Benet-Martínez, 2006). Recent conceptualisations of Agreeableness suggests that it aids the effect management of frustration and anger (Graziano, Jensen-Campbell, & Hair, 1996). Due to their forgiving nature and ability to manage anger, it is expected that agreeable individuals will be negatively associated with disenchantment. Individuals high on Neuroticism experience more frequent negative emotion, greater instability of affect, and higher sensitivity to adverse stimuli (McCullough et al., 2001; Watson, Clark, & Harkness, 1994). Neurotic individuals are predisposed to have a lower tolerance for stressors, making it more likely for them to experience disenchantment.
Dark-side traits represent interpersonally maladaptive and malevolent traits (Treglown et al., 2016). One of the most prominent taxonomies of dark personality is the dark triad of Machiavellianism, Psychopathy and Narcissism (Paulhus & Williams, 2002). It is expected that each of the dark triad traits will lead to increased propensity of disenchantment (H2a, H2b, H2c). Due to the high level of aggression associated with Psychopathy (Furnham, Richards, & Paulhus, 2013), it is expected that those who score high on this trait will react stronger to unfair treatment at work. Machiavellians experience greater paranoia and hyper-vigilance to manipulation (Monaghan, Bizumic, & Sellbom, 2016; Christoffersen & Stamp, 1995); they are constantly survey their environment for threats. As they are constantly looking for evidence that they are being unfairly treated, is likely that Machiavellians will be predisposed to experience higher levels of disenchantment. Finally, Narcissists react aggressively in response to threats to their ego (Jones & Paulhus, 2010). If Narcissists appraise job stressors as threats to their ego, it is likely that they will experience higher levels of disenchantment.
Honesty-humility, the sixth factor of the HEXACO model of personality (Lee & Ashton, 2004) depicts an individual’s fairness, modesty, and sincerity versus greed, deceit, and entitlement. Honesty-humility has been described as a broad dark side dimension due to its ability to account for common elements across the Dark Triad (Lee & Ashton, 2005; Furnham et al., 2013; Treglown et al., 2016). Previous research has demonstrated that honesty-humility is negatively correlated with negative reciprocity norms and a readiness to retaliate (Perugini et al., 2003). The lack of forgiveness associated with honesty-humility is thought to predispose individuals to react to transgressions, fostering feelings that others should “pay the price” for mistreatment (Lee & Ashton, 2012). It is hypothesised that the sharp reactions associated with low honesty-humility will manifest disenchantment with greater magnitude (H2d).
We test two hypotheses with respect to moral disengagement: first that it will be correlated with Obsesssive-Compulsive Disorder (OCD) (H3a), but also for reasons discussed above it will mediate the relationship between disenchantment and CWBs (H3b).We will also investigate the extent to which disenchantment is linked to specific as well as overall CWB. We hypothesise that all correlations will be positive and significant (H4a, H4b, H4c). Most importantly we are interested in the incremental validity issue assuming that disenchantment will add variance given that all the above individual trait variables have been controlled for (H5).
H1a: Agreeableness will be negatively associated with disenchantment.
H1b: Neuroticism will be positively associated with disenchantment.
H 2a: Psychopathy will be positively associated with disenchantment.
H2b: Machiavellianism will be positively associated with disenchantment.
H2c: Narcissism will be positively associated with disenchantment.
H2d: Honesty-humility will be positively associated with disenchantment.
H3a: Moral disengagement will positively predict the occurrence of CWB.
H3b: Moral disengagement will mediate the relationship between disenchantment and CWB.
H4a: Disenchantment will positively relate to CWB through abuse, product deviance, sabotage, theft, and withdrawal.
H4b: Disenchantment will positively relate to interpersonal deviance.
H4c: Disenchantment will positively relate to organisational deviance.
H5: Disenchantment will explain additional variance in CWB after controlling for individual difference variables.
3. Method
3.1. Participants
246 participants took part in this study; 142 of which were female, 104 were male, and 1 participant did not wish to specify their gender. Ages ranged from 18 to 65, with the majority of participants were between 18 and 24 years old (n = 100; 40.1%). We gathered information about which industry they worked in as well as education, total number of years worked, number of previous organisations, hours worked per week, or income level.
3.2. Measures
Employee Disenchantment: Employee disenchantment was assessed through the 45-item inventory. Totals for the five factors of disenchantment and overall disenchantment were calculated by summing the scores of all relevant items. The internal consistency of Employee Disenchantment was also high in α = .98. It has been used in a number of studies (Furnham & Horne, 2022).
Counterproductive Work Behaviour: In order to gain a broad understanding of the relationship between disenchantment and differing levels of deviant behaviour, three different measures were used to assess CWB.
1) Robinson and O’Leary (1998)’s Individual Anti-Social Behaviour Inventory, assessing deviant and destructive behaviours at work. Participants responded on a binary scale, indicating whether they had engaged in any of the nine behaviours in the past six months. A high score indicated a high level of individual anti-social behaviour. The average score for this scale was 11.16 (SD = 1.96). This scale was found to have good internal consistency (α = .71).
2) Bennett and Robinson (2000)’s Workplace Deviance Checklist. This inventory assessed two types of CWB: those directed at the organisation, i.e. theft or sabotage, and those directed at individuals within the organisation, i.e. bullying or rumour-spreading. Participants were asked 19 questions and indicated the frequency to which they engaged in each CWB (1 = Never to 5 = Always). The internal validity was high for both interpersonal (α = .84) and organisational deviance (α = .82).
3) CWB-Checklist (Spector et al., 2006). The CWB-Checklist assesses five different types of CWB: abuse, product deviance, sabotage, theft, and withdrawal. This scale was scored in a binary fashion; for 32 statements, participants indicated whether or not they had engaged in that behaviour at work within the past 6 months. An overall CWB-Checklist score and scores for the five CWB types were generated by summing the number of relevant CWBs a participant had engaged in. Internal reliability for this scale was high (α = .88).
Bright Personality: “Bright” personality was measured using Donnellan et al. (2006)’s Mini-IPIP measure. This inventory assesses the Big Five personality traits: Neuroticism, Extraversion, Conscientiousness, Agreeableness, and Intellect/Imagination. The Mini-IPIP consists of 20-items, with four items measuring each trait. Each item described behaviour, where respondents were asked to rate how applicable this statement was to them on a 5-point Likert scale. Totals for the five traits were calculated by summing the respondent’s scores for each relevant item. Previous studies have demonstrated that the Mini-IPIP has high internal consistency (Cooper, Smillie, & Corr, 2010), as well as comparable convergent and discriminant validity with other five-factor model measures (Donnellan et al., 2006).
Dark Personality: This study used two approaches to measuring “dark” personality. The first was to measure Honesty-Humility, one of the six traits found in Lee & Ashton (2004)’s HEXACO model of personality. Participants indicated on a five-point Likert scale the extent to which they agreed with the 16 Honesty-Humility items. Honesty-Humility was found to have a high internal consistency (α = .85).
The second approach assessed dark personality with more distinguishability, looking at the Dark Triad. This study used Jones and Paulhus (2014)’s Short Dark Triad (SD3) measure of sub-clinical Psychopathy (α = .76), Machiavellianism (α = 78), and Narcissism (α = .76). This scale consists of 27 items, with nine dedicated to each of the three Dark Triad traits. This scale has been found to be the most robust short measure of the Dark Triad (Maples et al., 2014).
Moral Disengagement: Individual propensity to morally disengage was assessed through Moore et al. (2012)’s short, eight-item measure. Each item represented one of eight inter-related cognitive mechanisms that Bandura (1986) proposed would expedite unethical behaviours. Participants rated the extent to which they agreed with statements on a 7-point Likert scale (1 = Strongly Disagree; 7 = Strongly Agree). Moral disengagement was calculated by summing the eight items, where a higher score indicated a greater propensity to morally disengage. This scale had strong internal consistency (α = .79).
3.3. Procedure
130 participants were gathered through opportunistic sampling on social media sites. Participants recruited through this method were asked their current type of employment. Due to the nature of the questions, only participants who were currently working (full-time, part-time, or self-employed) were used for the study. An additional 116 participants were recruited through Prolific by the first author an online market that recruits participants to take part in surveys. The researchers specified that the sample should be in full-or part-time employment due to the nature of the survey. The questionnaires were hosted on Prolific, for which participants were exchanged £1.50 for fully completing all questions. Written and informed consent was provided by all participants before engaging in the study.
4. Results
4.1. Demographics
Table 1 details how disenchantment correlates with demographic information. Disenchantment was significantly positively correlated with the number of years an employee has worked with their current employer (r = .16; p = .014). Regression analysis revealed, however, that this was not a significant predictor after controlling for age and gender (β = .13; p = .10).
Table 1. Correlations between disenchantment factors and demographics.
|
Disenchantment |
Organisational Lying |
Perceived Inequity |
Respect |
Distrust |
Broken Promises |
Age |
.09 |
.10 |
.05 |
.03 |
.14* |
.06 |
Gender |
.07 |
.02 |
.12 |
.02 |
.12 |
.04 |
Education |
−.03 |
.02 |
−.02 |
−.11 |
.01 |
−.05 |
Years w/Current |
.16* |
.16* |
.12 |
.12 |
.11 |
.17** |
Work Location |
−.09 |
−.12 |
−.08 |
−.04 |
−.04 |
−.06 |
Years Worked |
.06 |
.05 |
.03 |
.02 |
.09 |
.07 |
No. Orgs. |
.04 |
.00 |
.02 |
.03 |
.10 |
.03 |
Hours per Week |
−.10 |
−.13* |
−.06 |
.01 |
−.13* |
−.07 |
Income |
−.07 |
−.05 |
−.07 |
−.03 |
−.08 |
−.07 |
*p < .05; n = 246; **p < .01; Years w/Current = Years worked with current employer; No. Orgs = Number of organisations worked at prior to current.
A one-way ANOVA revealed that Disenchantment significantly between differed across industries (F(15, 228) = 2.27; p = .004). A Tukey post-hoc analysis revealed that participants who worked in Property were significantly less disenchanted than participants who worked in Construction (p = .037), Information and Communication (p = .037). It was assumed that Property elicited uncharacteristically positive scores, so a second ANOVA was run without scores from this industry. The significance of the ANOVA remained, reaffirming the result. Significant differences across industry were also noted for Organisational Lying (F(15, 227) = 1.83; p = .024), Perceive Inequity (F(15, 227) = 2.71; p < .001), Distrust (F(15, 227) = 2.26; p = .004), and Broken Promises (F(15, 227) = 1.74; p = .035). Tukey post-hoc analysis revealed that those who worked in Property had significantly lower disenchantment scores than Construction for Perceived Inequity (p = .005) and Distrust (p = .024). Total disenchantment was not significantly correlated with any other demographic variable: age, gender, education, total number of years worked, number of previous organisations, hours worked per week, or income level.
Distrust was significantly positively correlated with age (r = .14; p = .030). The number of years an employee worked with their current employer was positively correlated with Organisational Lying (r = .16; p = .015) and Broken Promises (r = .17; p = .009). The number of hours an employee works per week was negatively correlated with perceptions of Organisational Lying (r = −.13; p = .042) and Distrust (r = −.13; p = .045).
4.2. Disenchantment and Individual Differences
Firstly, this study analysed whether disenchantment was related to particular “bright” personality, namely the Big Five traits. Correlations between disenchantment and bright personality can be seen in Table 2. Neuroticism was positively correlated with total Disenchantment (r = .16; p = .013) and perceptions of Respect/Bullying (r = .22; p < .001), Distrust (r = .17; p = .008), and Broken Promises (r = .14; p = .028). Agreeableness was negatively correlated with total Disenchantment (r = −.16; p = .011) and perceptions of Organisational Lying (r = −.16; p = .010), Perceived Inequity (r = −.15; p = .021), and Broken Promises (r = −.15; p = .022). Intellect/Imagination was negatively correlated with total Disenchantment (r = −.22; p = .001), and with perceptions of Organisational Lying (r = −.21; p = .001), Perceived Inequity (r = −.19; p = .003), Respect/Bullying (r = −.14; p = .029), Distrust (r = −.20; p = .002), and Broken Promises (r = −.18; p = .004). This confirms H1a and H1b.
Table 2. Correlations between disenchantment and IPIP “Bright” personality traits.
|
Disenchantment |
Organisational Lying |
Perceived Inequity |
Respect |
Distrust |
Broken Promises |
Extraversion |
−.03 |
−.03 |
−.07 |
.02 |
−.04 |
−.01 |
Agreeableness |
−.16* |
−.16* |
−.15* |
−.11 |
−.12 |
−.15* |
Conscientiousness |
−.06 |
−.10 |
−.05 |
.01 |
−.08 |
−.03 |
Neuroticism |
.16* |
.06 |
.12 |
.22** |
.17** |
.14* |
Imaginative/Intellect |
−.22** |
−.21** |
−.19** |
−.14* |
−.20** |
−.18** |
*p < .05; n = 246; **p < .01.
This study looked at the Honesty-Humility dimension of the HEXACO, and the Dark Triad of Psychopathy, Machiavellianism, and Narcissism. These correlations can be seen in Table 3. This confirms H2a and H2b but not H2c.
Honesty-Humility was negatively correlated with total Disenchantment (r = −.21; p = .001), and specifically with perceptions of Organisational Lying (r = −.19; p = .002), Perceived Inequity (r = −.15; p = .019), Distrust (r = −.17; p = .009), and Broken Promises (r = −.24; p < .001). This confirms H2d. Honesty-Humility was not significantly correlated with perceptions of Bullying. Total Disenchantment was positively correlated with Psychopathy (r = .24; p < .001) and Machiavellianism (r = .29; p < .001); all five sub-factors of disenchantment were also positively correlated to Psychopathy and Machiavellianism. Narcissism was not significantly correlated with disenchantment.
Table 3. Correlations between disenchantment facets and “Dark” personality traits.
|
Disenchantment |
Organisational Lying |
Perceived Inequity |
Respect |
Distrust |
Broken Promises |
Psychopathy |
.24*** |
.25*** |
.14* |
.26*** |
.13* |
.25*** |
Machiavellianism |
.29*** |
.27*** |
.26*** |
.16* |
.24*** |
.29*** |
Narcissism |
.05 |
.05 |
.02 |
.06 |
−.00 |
.09 |
Moral Disengage |
.19** |
.19** |
.10 |
.20** |
.13* |
.19** |
Honesty-Humility |
−.21** |
−.19** |
−.15* |
−.12 |
−.17** |
−.24*** |
*p < .05; n = 246; **p < .01; ***p < .001.
A three-step hierarchical regression was conducted to analyse which personality variables predicted disenchantment. The first step contained age and gender, the second step included the Big Five personality traits, and the third step included Honesty-Humility and the Dark Triad. The regression accounted for 13.6% of the variance. The regression revealed that individuals who have lower Intellect/Imagination and higher levels of Psychopathy and Machiavellianism were more likely to be disenchanted. Interestingly there was some variance across the five factors: for instance both Neuroticism and Imagination was significant on four of the five factors. Also the introduction of Psychopathy, Machiavellianism, and Narcissism in Step 3 consistently improves the model fit (significant F Change), but the impact of narcissism remains inconsistent. Results can be seen in Table 4.
Table 4. Regression of demographics, bright personality, and dark personality on disenchantment.
|
|
Disenchantment |
Organisational Lying |
Perceived Inequity |
Respect |
Distrust |
Broken Promises |
|
|
β |
t |
β |
t |
β |
t |
β |
t |
β |
t |
β |
t |
Step 1 |
Age |
.084 |
1.31 |
.093 |
1.45 |
.041 |
.636 |
.022 |
.336 |
.132 |
2.08* |
.058 |
.901 |
|
Gender |
.071 |
1.10 |
.013 |
.195 |
.126 |
1.96 |
.013 |
.195 |
.112 |
1.77 |
.047 |
.736 |
F Change |
|
|
1.54 |
|
1.08 |
|
2.19 |
|
.078 |
|
3.88* |
|
.708 |
R2 |
|
|
.013 |
|
.009 |
|
.018 |
|
.001 |
|
.031 |
|
.006 |
AdjR2 |
|
|
.004 |
|
.001 |
|
.010 |
|
−.008 |
|
.023 |
|
−.002 |
Step 2 |
Extraversion |
.049 |
0.71 |
.035 |
.525 |
−.005 |
−.076 |
.092 |
1.40 |
.030 |
.463 |
.065 |
.983 |
|
Agreeableness |
−.177 |
−2.63** |
−.157 |
−2.29** |
−.155 |
−2.27* |
−.144 |
−2.12* |
−.141 |
−2.09* |
−.164 |
−2.40* |
|
Conscientiousness |
−.014 |
−0.21 |
−.068 |
−1.03 |
−.028 |
−.431 |
.090 |
1.39 |
−.040 |
−.611 |
.012 |
.181 |
|
Neuroticism |
.148 |
2.19* |
.041 |
.601 |
.073 |
1.07 |
.268 |
3.97*** |
.144 |
2.14* |
.145 |
2.12* |
|
Imagination |
−.169 |
−2.64** |
−.172 |
−2.65** |
−.149 |
−2.31 |
−.112 |
−1.75 |
−.141 |
−2.20* |
−.141 |
−2.17* |
F Change |
|
|
4.42*** |
|
3.42** |
|
3.04* |
|
4.85*** |
|
3.54** |
|
3.37** |
R2 |
|
|
0.09 |
|
.076 |
|
.078 |
|
.094 |
|
.099 |
|
.072 |
AdjR2 |
|
|
.071 |
|
.049 |
|
.050 |
|
.067 |
|
.072 |
|
.045 |
Step 3 |
Psychopathy |
.230 |
2.70** |
.260 |
3.01** |
.120 |
1.373 |
.308 |
3.56*** |
.104 |
1.19 |
.199 |
2.30* |
|
Machiavellianism |
.187 |
2.30* |
.159 |
1.93 |
.218 |
2.61** |
.075 |
.917 |
.170 |
2.05* |
.171 |
2.07* |
|
Narcissism |
.018 |
.217 |
−.036 |
−.420 |
.043 |
.500 |
.041 |
.480 |
−.006 |
−.067 |
.053 |
.623 |
|
Honesty-Humility |
.086 |
.989 |
.072 |
.808 |
.092 |
1.03 |
.170 |
1.93 |
.026 |
.291 |
.032 |
.366 |
F Change |
|
|
5.44** |
|
5.16** |
|
3.62** |
|
4.35** |
|
2.65* |
|
5.47*** |
R2 |
|
|
.175 |
|
.152 |
|
.132 |
|
.158 |
|
.139 |
|
.153 |
AdjR2 |
|
|
.136 |
|
.112 |
|
.091 |
|
.117 |
|
.098 |
|
.112 |
*p < .05; **p < .01; ***p < .001.
4.3. Disenchantment and Moral Disengagement
Moral disengagement was correlated with Disenchantment (r = .19; p = .003). This confirmed H3a. Specifically, moral disengagement was correlated with Organisational Lying (r = .19; p = .003), Respect/Bullying (r = .20; p = .002), Distrust (r = .13; p = .038), and Broken Promises (r = .19; p = .003). Moral disengagement was not significantly correlated perceptions of Perceived Inequity.
A three-step hierarchical regression was conducted to assess the extent to which disenchantment can explain variance in the propensity to morally disengage. The first step contained age and gender, the second disenchantment, and the third step included the bright and dark personality variables. Age was a significant predictor of moral disengagement (β = −.142; p = .027). Disenchantment was a significant predictor of moral disengagement in second step (β = .204; p = .001), explaining 5.2% of the variance. This initially indicated that individuals with higher disenchantment are more likely to morally disengage. However, disenchantment became non-significant in the third step. Intellect/ Intelligence and Honesty-Humility were negative predictors, whilst Psychopathy, Machiavellianism, and Narcissism were positive predictors. This final step accounted for 45.8% of the variance. Results can be seen in Table 5.
Table 5. Regression of demographics, disenchantment, and personality on propensity to morally disengagement.
|
|
Moral Disengagement |
|
|
β |
t |
Step 1 |
Age |
.062 |
0.96 |
|
Gender |
−.142 |
−2.23* |
F Change |
|
2.86 |
|
R2 |
|
.023 |
|
AdjR2 |
|
.015 |
|
Step 2 |
Disenchantment |
.204 |
3.24** |
F Change |
|
10.46** |
|
R2 |
|
.064 |
|
AdjR2 |
|
.052 |
|
Step 3 |
Extraversion |
−.101 |
−1.63 |
|
Agreeableness |
−.084 |
−1.51 |
|
Conscientiousness |
−.017 |
−0.34 |
|
Neuroticism |
−.027 |
−0.51 |
|
Imagination |
−.223 |
−4.37*** |
|
Psychopathy |
.233 |
3.40** |
|
Machiavellianism |
.213 |
3.27** |
|
Narcissism |
.177 |
2.64** |
|
Honesty-Humility |
−.201 |
−2.90** |
F Change |
|
20.84*** |
|
R2 |
|
.485 |
|
AdjR2 |
|
.458 |
|
*p < .05; **p < .01; ***p < .001.
4.4. Disenchantment and CWBs
Table 6 shows the correlations between Disenchantment, personality, moral disengagement, and the three measures of CWB. Disenchantment was significantly positively correlated with total scores on the CWB-Checklist (r = .29; p < .001), abuse (r = .23; p < .001), product deviance (r = .28; p < .001), sabotage (r = .23; p < .001), withdrawal (r = .26; p < .001), organisational deviance (r = .30; p < .001), interpersonal deviance (r = .26; p < .001), and individual anti-social behaviour (r = .39; p < .001). The relationship between disenchantment and theft bordered on significant (r = .12; p = .059). However, increased Broken Promises was significantly correlated with theft (r = .13; p = .042).
Table 6. Correlations between disenchantment facets, personality, dark triad, honesty-humility, and counterproductive work behaviour measures.
|
Disench. |
Org. Lying |
Per. In. |
Respect |
Distrust |
Br. Prom. |
E. |
A. |
C. |
N. |
I. |
SDT-P. |
SDT-M |
SDT-N |
HH |
MD |
CWB Checklist—Total |
.29*** |
.22*** |
.25*** |
.31*** |
.21*** |
.28*** |
.05 |
−.22** |
−.08 |
.06 |
−.02 |
.50*** |
.33*** |
.22** |
−.34*** |
.33*** |
CWB Checklist—Abuse |
.23*** |
.14* |
.20** |
.30*** |
.16* |
.22** |
.06 |
−.24*** |
−.08 |
.07 |
.01 |
.46*** |
.27*** |
.24*** |
−.27*** |
.28*** |
CWB Checklist
—Product Deviance |
.28*** |
.27*** |
.25*** |
.25*** |
.21*** |
.23*** |
−.12 |
−.28*** |
−.06 |
.12 |
−.05 |
.32*** |
.25*** |
.09 |
−.23*** |
.26*** |
CWB Checklist—Sabotage |
.23*** |
.14* |
.19** |
.29*** |
.17*** |
.20** |
.07 |
−.12 |
−.06 |
.08 |
−.06 |
.35*** |
.13* |
.11 |
−.19** |
.19** |
CWB Checklist—Theft |
.121 |
.087 |
.09 |
.120 |
.098 |
.13* |
.034 |
−.17** |
−.04 |
−.01 |
−.02 |
.27*** |
.26*** |
.10 |
−.28** |
.27*** |
CWB Checklist
—Withdrawal |
.26*** |
.23*** |
.22** |
.19** |
.19** |
.27*** |
.07 |
.01 |
−.04 |
−.01 |
−.04 |
.35*** |
.24*** |
.13 |
−.25*** |
.19** |
Interpersonal Deviance |
.26*** |
.18** |
.24*** |
.35*** |
.20** |
.21** |
−.03 |
−.22** |
−.21** |
.05 |
−.01 |
.45*** |
.40*** |
.22** |
−.36*** |
.36*** |
Organisational Deviance |
.30*** |
.28*** |
.28*** |
.22** |
.24*** |
.25*** |
.08 |
−.22*** |
−.11 |
.11 |
−.01 |
.48*** |
.31*** |
.23*** |
−.34*** |
.29*** |
Anti-Social Behaviours |
.39*** |
.31*** |
.37*** |
.36*** |
.33*** |
.33*** |
−.06 |
−.22** |
−.18** |
.16* |
−.03 |
.41*** |
.41*** |
.15* |
−.30*** |
.33*** |
Disench.—Disenchantment; Org. Lying—Organisational Lying; Per. In.—Perceived Inequity; Br. Prom.—Broken Promises; E—Extraversion; A—Agreeableness; C—Conscientiousness; N—Neuroticism; I—Intellect/Imagination; SDT-P—Psychopathy; SDT-M—Machiavellianism; SDT-N—Narcissm; HH—Honesty-Humility; MD—Moral Disengagement. *p < .05; n = 246; **p < .01; ***p <.001.
Two-step regressions (controlling for age and gender in the first step) were conducted to assess the extent to which disenchantment predicted four main types of CWB. Gender was found to be a significant negative for CWB Checklist, Organisational deviance, and Interpersonal Deviance. Age was also a negative predictor for interpersonal deviance. Disenchantment was found to be a significant positive predictor for all four main types of CWB (see Table 7 for the results).
Table 7. Regression of disenchantment on four measures of CWB.
|
|
CWB Checklist |
Interpersonal Deviance |
Organisational Deviance |
Anti-Social Behaviour |
|
|
β |
t |
β |
t |
β |
t |
β |
t |
Step 1 |
Age |
.019 |
.299 |
−.070 |
−1.11 |
−.137 |
−2.20** |
.006 |
.090 |
|
Gender |
−.181 |
−2.86** |
−.132 |
−2.08* |
−.20 |
−2.86** |
−.081 |
−1.26 |
F Change |
|
|
4.10** |
|
2.88 |
|
6.79** |
|
.793 |
R2 |
|
|
.033 |
|
.023 |
|
.053 |
|
.007 |
Adj.R2 |
|
|
.025 |
|
.015 |
|
.045 |
|
−.002 |
Step 2 |
Disenchantment |
.308 |
5.10*** |
.284 |
4.64*** |
.326 |
5.49*** |
.403 |
6.82*** |
F Change |
|
|
11.69*** |
|
21.49*** |
|
30.09*** |
|
46.49*** |
R2 |
|
|
.127 |
|
.103 |
|
.158 |
|
.167 |
Adj.R2 |
|
|
.116 |
|
.092 |
|
.147 |
|
.157 |
*p < .05; **p < .01; ***p < .001.
A series of five-step hierarchical regression were conducted to analyse the additional variance employee disenchantment explained in different types of CWB, after accounting for age, bright and dark personality, and propensity to morally disengage.
The first set of regressions analysed individual anti-social behaviour. Age and gender were not significant predictors of anti-social behaviour. In the second step, which accounted for 7.2% of the variance, Agreeableness negatively (β = −.21; p = .002), whilst Neuroticism positively (β = .18; p = .008), predicted individual anti-social behaviour. In the third step accounted for 23.2% of the variance, with Psychopathy (β = 3.63; p < .001) and Machiavellianism (β = 3.78; p < .001) being positive predictors. Agreeableness was no longer a significant predictor in this step. Moral disengagement was not a significant predictor of anti-social behaviour in the fourth step. Disenchantment was a significant positive predictor in the fifth step (β = 4.48; p < .001). The five-step model accounted for 29.1% of the variance, with Psychopathy, Machiavellianism, and Disenchantment being the only significant predictors. Results can be seen in Table 8.
Table 8. Regression of demographics, bright personality, dark personality, moral disengagement, and disenchantment on antisocial behaviour, and interpersonal and organisational deviance.
|
|
Anti-Social Behaviour |
Interpersonal Deviance |
Organisational Deviance |
|
|
β |
t |
β |
t |
β |
t |
Step 1 |
Age |
.007 |
.105 |
−.070 |
−1.10 |
−.130 |
−2.07* |
|
Gender |
−.085 |
−1.32 |
−.143 |
−2.24* |
−.182 |
−2.89** |
F Change |
|
|
.872 |
|
3.21* |
|
6.58** |
R2 |
|
|
.007 |
|
.026 |
|
.052 |
AdjR2 |
|
|
−.001 |
|
.018 |
|
.044 |
Step 1 |
Extraversion |
.027 |
.417 |
.165 |
2.54* |
.023 |
.361 |
|
Agreeableness |
−.211 |
−3.12** |
−.256 |
−3.823*** |
−.180 |
−2.69** |
|
Conscientiousness |
−.100 |
−1.55 |
−.019 |
−.290 |
−.145 |
−2.26* |
|
Neuroticism |
.180 |
2.67** |
.169 |
2.53** |
.058 |
.865 |
|
Imagination |
.000 |
−.002 |
−.005 |
−.076 |
−.025 |
−.392 |
F Change |
|
|
4.79*** |
|
4.70*** |
|
3.47** |
R2 |
|
|
.099 |
|
.115 |
|
.117 |
AdjR2 |
|
|
.072 |
|
.088 |
|
.091 |
Step 2 |
Psychopathy |
.291 |
3.63*** |
.376 |
4.66*** |
.276 |
3.48** |
|
Machiavellianism |
.290 |
3.78*** |
.062 |
.799 |
.234 |
3.09** |
|
Narcissism |
.048 |
.606 |
.037 |
.456 |
.026 |
.327 |
|
Honesty-Humility |
.093 |
1.13 |
−.019 |
−.224 |
−.020 |
−.241 |
F Change |
|
|
13.24*** |
|
11.28*** |
|
13.75*** |
R2 |
|
|
.267 |
|
.259 |
|
.287 |
AdjR2 |
|
|
.232 |
|
.224 |
|
.253 |
Step 3 |
Moral Disengagement |
.075 |
.961 |
−.011 |
−.135 |
.079 |
1.023 |
F Change |
|
|
.924 |
|
.018 |
|
1.05 |
R2 |
|
|
.270 |
|
.259 |
|
.290 |
AdjR2 |
|
|
.232 |
|
.221 |
|
.253 |
Step 4 |
Disenchantment |
.267 |
4.479*** |
.145 |
2.34* |
.202 |
3.37** |
F Change |
|
|
20.07*** |
|
5.46* |
|
11.36** |
R2 |
|
|
0.329 |
|
.277 |
|
.324 |
AdjR2 |
|
|
0.291 |
|
.236 |
|
.285 |
*p < .05; **p < .01; ***p < .001.
The second set of regressions look at Interpersonal and Organisational deviance. In the first steps, age predicted organisational deviance (β = −.14; p = .026), whilst males were significantly more likely to engage in both interpersonal (β = ; p = ) and organisational (β = −.18; p = .004) deviance. In the second step, Extraversion (β = .17; p = .012) and Neuroticism (β = .17; p = .012) positively predicted interpersonal deviance, Conscientiousness negatively predicted organisational deviance (β = −.15; p = .025), and Agreeableness negatively predicted both interpersonal (β = −.27; p < .001) and organisation (β = −.18; p = .008) deviance. In the third step, Machiavellianism positively predicted organisational deviance (β = .23; p − .002), whilst Psychopathy positively predicted both interpersonal (β = .38; p < .001) and organisational deviance (β = .28; p = .001). Moral disengagement was not a significant predictor of either types of deviance. Disenchantment was a positive predictor in the fifth step for both interpersonal (β = .15; p = .020) and organisational (β = .20; p = .001) deviance. The five-step model of interpersonal deviance accounted for 23.6%, with Psychopathy and Disenchantment being the only predictors. The five-step model of organisational deviance accounted for 28.5% of the variance, with Conscientiousness being a negative predictor whilst Psychopathy, Machiavellianism, and Disenchantment acting as positive predictors. The results can be seen in Table 9.
Table 9. Regression of demographics, bright personality, dark personality, moral disengagement, and disenchantment on counterproductive work behaviour.
|
|
CWB Checklist |
Abuse |
Product Deviance |
Sabotage |
Theft |
Withdrawal |
|
|
β |
t |
β |
t |
β |
t |
β |
t |
β |
t |
β |
t |
Step 1 |
Age |
.022 |
.350 |
.041 |
0.652 |
−.039 |
−.619 |
−.044 |
−.679 |
.035 |
.538 |
.007 |
.103 |
|
Gender |
−.189 |
−2.98** |
−.170 |
−2.68** |
−.158 |
−2.49* |
−.091 |
−1.41 |
−.033 |
−.512 |
−.183 |
−2.88** |
F Change |
|
|
4.45* |
|
3.73* |
|
3.36* |
|
1.27 |
|
.264 |
|
4.14* |
R2 |
|
|
.036 |
|
0.030 |
|
.027 |
|
.011 |
|
.002 |
|
.033 |
AdjR2 |
|
|
.028 |
|
0.022 |
|
.019 |
|
.002 |
|
−.006 |
|
.025 |
Step 1 |
Extraversion |
.138 |
2.17* |
.156 |
2.41** |
−.024 |
−.368 |
.119 |
1.772 |
.085 |
1.25 |
.083 |
1.239 |
|
Agreeableness |
−.238 |
−3.53*** |
−.278 |
−4.17*** |
−.248 |
−3.72*** |
−.140 |
−2.02* |
−.211 |
−3.04** |
.031 |
.453 |
|
Conscientiousness |
.006 |
.094 |
.011 |
0.172 |
.022 |
.351 |
.002 |
.029 |
.012 |
.184 |
−.020 |
−.294 |
|
Neuroticism |
.146 |
2.17** |
.161 |
2.41** |
.173 |
2.60** |
.121 |
1.742 |
.019 |
.280 |
.051 |
.731 |
|
Imagination |
−.004 |
−.055 |
.041 |
0.654 |
−.027 |
−.424 |
−.061 |
−.924 |
−.009 |
−.142 |
−.049 |
−.750 |
F Change |
|
|
3.60** |
|
4.71*** |
|
4.85*** |
|
1.94 |
|
1.94 |
|
.661 |
R2 |
|
|
.104 |
|
0.118 |
|
.118 |
|
.050 |
|
.042 |
|
.047 |
AdjR2 |
|
|
.078 |
|
0.092 |
|
.092 |
|
.021 |
|
.013 |
|
.018 |
Step 2 |
Psychopathy |
.421 |
5.26*** |
.405 |
5.00*** |
.224 |
2.63** |
.389 |
4.47*** |
.166 |
1.89 |
.298 |
3.48** |
|
Machiavellianism |
.081 |
1.06 |
.031 |
0.394 |
.068 |
.840 |
−.052 |
−.631 |
.092 |
1.09 |
.136 |
1.67 |
|
Narcissism |
.009 |
.115 |
.102 |
1.27 |
.066 |
.786 |
−.052 |
−.603 |
−.085 |
−.968 |
−.117 |
−1.374 |
|
Honesty-Humility |
.000 |
−.004 |
.070 |
0.836 |
.019 |
.219 |
.022 |
.246 |
−.134 |
−1.48 |
−.057 |
−.651 |
F Change |
|
|
13.37*** |
|
10.38*** |
|
4.19** |
|
6.05*** |
|
4.95** |
|
8.41*** |
R2 |
|
|
.273 |
|
0.253 |
|
.178 |
|
.140 |
|
.117 |
|
.168 |
AdjR2 |
|
|
.238 |
|
0.217 |
|
.139 |
|
.099 |
|
.075 |
|
.128 |
Step 3 |
Moral Disengagement |
.046 |
.583 |
.021 |
.268 |
.085 |
1.025 |
.015 |
.180 |
.106 |
1.23 |
−.004 |
−.042 |
F Change |
|
|
.340 |
|
.072 |
|
1.05 |
|
.033 |
|
1.51 |
|
.002 |
R2 |
|
|
.274 |
|
.253 |
|
.182 |
|
.140 |
|
.123 |
|
.168 |
AdjR2 |
|
|
.236 |
|
.214 |
|
.139 |
|
.095 |
|
.077 |
|
.124 |
Step 4 |
Disenchantment |
.179 |
2.93** |
.116 |
1.852 |
.194 |
3.00** |
.154 |
2.30* |
.036 |
.525 |
.212 |
3.26** |
F Change |
|
|
8.61** |
|
3.50 |
|
8.97** |
|
5.30** |
|
.275 |
|
10.63** |
R2 |
|
|
.300 |
|
.264 |
|
.213 |
|
.159 |
|
.124 |
|
.205 |
AdjR2 |
|
|
.260 |
|
.222 |
|
.168 |
|
.112 |
|
.075 |
|
.160 |
*p < .05; **p < .01; ***p < .001.
The final set of regressions looked at the CWB-Checklist and its five types of CWB. The CWB-Checklist was predicted by gender in the first step (β = −.189; p = .003), with males being more likely to commit CWBs. In the second step, which accounted for 7.8% of the variance, Extraversion (β = .138; p = .035) and Neuroticism (β = .146; p = .031) were positive predictors of CWB, whilst Agreeableness was a negative predictor (β = −.238; p < .001). In the third step, all previous significant predictors became non-significant, with Psychopathy being the only significant predictor (β = .421; p < .001); this accounted for 23.8% of the variance. In the fourth step, moral disengagement was not a significant predictor of CWB, explaining no additional variance. The fifth step found that employee disenchantment was a significant positive predictor of CWBs (β = .179; p = .004). The multivariate regression was significant (F(13, 229) = 7.55; p < .001) and explained 26% of the variance in CWBs, with only Psychopathy and Disenchantment being positive predictors in the final model. Disenchantment was found to explain significant additional variance for product deviance, sabotage, and withdrawal after controlling for demographics, personality, and moral disengagement. .
5. Discussion
In this study, we were particularly interested in the extent of employee disenchantment’s instrumentality in predicting CWB. The results from this study provide support the hypothesis that employee disenchantment, a cognitive-affective response to unjust organisational processes and management practices, is an important predictor of CWB. Disenchantment significantly correlated with both organisational and interpersonal deviance and different methods of engaging in CWB, including abuse, product deviance, sabotage, and withdrawal. The borderline non-significance in correlating with theft is partially supported by the literature; previous research has shown theft to often be instrumental in nature, and not contingent on situation effects (Spector et al., 2006). Higher levels of disenchantment were related to employees engaging in specific methods of CWB —i.e. abuse, product deviance, sabotage, and withdrawal (hypothesis 1a)—as well as different targets of CWB—i.e. both organisational and individual deviance (hypothesis 1b and c). All this suggests that assessing disenchantment may be as useful as assessing engagement in organisations.
The significance of the results for multiple forms of CWB also indicates that there are potentially multiple pathways that disenchantment can impact CWBs. Previous research has shown that perceived injustice causes employees to maximise their own self-interest at work, reducing their awareness or compliance with proper conduct (Johnson & Lord, 2010). Disenchantment therefore motivates employees to engage in CWBs as a method of “balancing the scales” at work.
These results support this previous literature, finding that increased disenchantment is linked to retributive and restorative forms of CWB, namely interpersonal abuse focused on the source of mistreatment (Hershcovis et al., 2007). Additionally, these findings could also occur as disenchantment depletes an individual’s psychological and self-regulatory resources, support ego depletion theory, causing the individual to be more likely to have an uninhibited reaction to the source of the injustice. There is further support for the ego depletion theory of the link between disenchantment and CWBs, as the results found increased disenchantment to be predictive of withdrawal—indicating the employees no longer have the resources to psychologically cope at work—and product deviance withdrawal—indicating indicating that disenchantment depletes the needed self-regulatory resources needed to follow proper work protocol.
However, perceiving the workplace, its organisational processes, or management practises as unjust does not always lead to deviate behaviour. For example, Howard and Cordes (2010) found that perceived injustice increased an individual’s likelihood of withdrawing from an organisation (e.g. absenteeism and turnover intentions) as well as increasing their risk of experiencing burnout. Instead, there has been now a growing interest by psychologists researching workplace injustice to understand the role of individual differences in how an employee behaviourally, emotionally, and 76 cognitively reacts to injustice (Karremans, Van Lange, & Holland, 2005). What is unclear is the extent to which ego depletion theory explains additional consequential behaviours (or lack thereof) of disenchantment.
Previous research has called for the combination of cognitive and affective models to better understand how situational factors influence CWB (Colquitt et al., 2013). The nature of disenchantment provides insight into how a monistic approach, combining cognitive and affective responses to injustice, is instrumental in predicting CWB. However, CWBs are best thought of as a combination between individual differences and situational factors (Matta et al., 2015). Confirming hypothesis 4, disenchantment was found to explain additional unique variance in most CWBs after controlling for demographics, bright and dark personality, and moral disengagement.
The study did highlight various sex differences consistently showing that women scored much lower on CWBs than men, though there was little differences in Eds. This suggests that whilst men and women may be likely to experience disengagement in similar ways, they may act of these experiences rather differently.
Another interesting finding was the positive association between trait Neuroticism and both Anti-Social Behaviour and Interpersonal Deviance in the regressions. It was the only bright-side factor positively associated with the CWB measures and warrants further investigation. Also across the tables, Agreeableness notably shows a negative and often significant relationship with various adverse outcomes (e.g., CWB Checklist). This is to be expected because of the nature of Agreeable people who are more prone to forgiveness and reconciliation that revenge and retribution (Furnham, 2017). Similarly, Honesty-Humility’s negative relationship with most negative outcomes supports its protective role.
The results of this study are in stark contrast to what was found and recommended by O’Neill et al. (2011). In their paper, it was argued that organisations should focus on leveraging personality characteristics as a way of identifying and screening out these employees. However, this study extended these previous findings in several ways. Firstly, this study utilised an injustice measure, rather than looking at the absence of justice, which could explain why this study was able to show that disenchantment acted as a significant additional predictor. Additionally, this study was novel in looking at the interaction effects between disenchantment and personality. For instance, disenchantment was found to moderate the relationship between Narcissism and organisational deviance, Psychopathy and sabotage, Honesty-Humility and theft, 77 and Machiavellianism with sabotage and theft. Disenchantment was noted to be a less powerful direct predictor of the specific methods in which employees engage in CWB (i.e. abuse vs. theft). Instead, personality traits (particularly “dark” variables) explain greater variance in these types of CWB.
These results support previous research that has demonstrated the dark triad’s role in “misbehaving” at work (Furnham & Taylor, 2011). The aggressive, impulsive, and malevolent tendencies associated with these dark traits facilitate employees engaging in CWB. Low conscientiousness has also been associated with increased CWB, as these employees are less likely to believe that they will be caught as well as having lower motivation to comply with normal organisational conduct (Ryan & Sackett, 1987). Hypothesis 5a attempted to replicate previous research by testing the direct link of MD on increased CWB. This hypothesis was not supported, finding that MD was not a direct significant predictor of any CWB after controlling for bright and dark personality. This is in large contradiction to previous research that has argued that MD acts both directly effects (Moore et al., 2012) and mediates (Fida et al., 2015) the instance of employee deviance. One potential reason for the difference in results is that previous research has struggled to properly conceptualise and distinguish MD from dark personality in CWB research. Whilst it is defined as a socio-cognitive dissociation from deviant acts (Fida et al., 2015), it is not properly clarified as to whether this represents a form of state-based, maladaptive processing or a stable characteristic of individual difference (e.g. Roeser et al., 2016).
The lack of predictive validity of MD on CWB after including such a range of bright and dark personality traits seems to suggest that MD is best characterised as a personality trait. For instance, it is possible that previous research neglected the co-occurrence of dark triad traits and MD, as MD offered no incremental validity once these traits were accounted for. Additionally, hypothesis 3b was partially supported, finding that MD partially mediated the effect of disenchantment on CWBs. These results suggested that employees who are highly disenchanted are less likely to engage in CWBs if they possess a low propensity to morally disengage. MD has previously been related to lower self-regulatory efficacy providing a heuristic to rationalise negative behaviour. MD appears to be triggered in response to self-regulatory resources in the link between disenchantment and CWB; as disenchantment increases and depletes psychological resources, this increases an employee’s likelihood to morally disengage and commit CWBs. Further research is needed on MD to fully understand whether it should be measured as a stable, individual difference or whether it is artificially measuring related dark personality traits (e.g. psychopathy) that better explain the relationship with workplace behaviours due to its measurement scale.
6. Limitations
This study is not without its limitations. We had a somewhat limited and restricted sample of working people and it would be most desirable to replicate this study on a much larger population group. The cross-sectional nature of this study only provides a snapshot into how disenchantment facilitates the explanation of employee deviance. Cross-sectional study design has been criticised with regards to predicting related outcomes for not being able to fully verify causality (Wright, Gardner, & Moynihan, 2003; Meier & Spector, 2013). Future research into disenchantment would need to assess the longitudinal impact this negative effect has on the proclivity to harm an organisation or its employees. Also, the data in this study was collected solely through self-report measures. Future research should aim to address this limitation by collecting observational or behavioural data. Other-report data on disenchantment and CWBs would provide more insight into the relationship between these variables. Finally, we focused on individual difference variance and did not pay attention to possible environmental contributors to disenchantment such as corporate culture.
Data Availability
This is obtainable from the first author on request.
Author Contribution
LT collected the data and did the statistical analyses, AF wrote the paper.