Adaptation of Collective Moral Disengagement Scale into Turkish Culture for Adolescents

In this study, reliability and validity are assessed for a Turkish culture adaptation of the Collective Moral Disengagement Scale for Adolescents. The study was carried out in two stages. In the first stage, translation, exploratory factor analysis, internal consistency coefficients, and test-retest method were performed; in the second stage, confirmatory factor analysis, and compliance validation study were conducted. In the first stage of the research, a total of 339 adolescents and in the second stage, 283 adolescents participated. At the end of the exploratory factor analysis performed in the first stage of the research, it is observed that the CMDS (Collective Moral Disengagement Scale for Adolescents) explained 35 percent of total variance. The factor loadings of the scale items ranged from 0.41 to 0.76. In addition, the internal consistency coefficient was found as 0.86 and test-retest reliability coefficient as 0.86 at the end of the analysis made for the CMDS's reliability in the first stage. At the end of confirmatory factor analysis made in the second stage, acceptable fit indices were obtained (χ=288.07, df= 90, χ /df= 3.20, RMSEA=0.08, RMR=0.13, SRMR=0.06, GFI=0.89, AGFI=0.85, CFI=0.94, NNFI=0.93). Within the compliance validity of CMDS, correlation coefficients of Bullying Scale and Bullying Sub-scale (0.41) and the Victim Subscale (0.47) were examined. In conclusion, we can say that the Turkish form of 15-item Collective Moral Disengagement Scale for Adolescents is valid and reliable.


Introduction
The concept of moral disengagement has in recent years begun grabbing attention as being among the reasons of aggressive and violent behaviors.Moral disengagement, which is based on Bandura's Social-Cognitive learning and used to explain ways of justifying individuals' immoral behaviors, takes part in research subjects of different disciplines as it is related to antisocial behaviors that are not desired from school age to adulthood [1].
As of their childhood, individuals develop right and wrong standards that will direct their behaviors.Through self-regulation, human beings tend to behave in accordance with moral standards that will satisfy them and develop their self-respect or tend to avoid negative acts that lessen self-control.In some cases, the self-control system that keeps behaviors within moral standards deteriorates and people generally tend to reveal negative behaviors.Sometimes, individuals activate their moral disengagement mechanisms to avoid negative self-sanction when they act apart from moral standards.Moral disengagement is defined as a cognitive process through which one justifies his detrimental and aggressive behaviors by loosening his inherent self-regulatory mechanism [2].Moral disengagement behaviors is done through reframing destructive behaviors, reducing the state of being charged of the behavior, paying no attention to what could follow it and reducing the negative impacts that could happen by accusing the victim [3].Thanks to these mechanisms, before criminal behavior occurs, the expected negative feelings of individuals can be minimized and thus emotional self-regulation moves freely [2].
It is seen that researches concerning to moral disengagement are frequently conducted on children and adolescents.This term is seen as a key factor in explaining the spread of aggressive and violent behaviors in children and adolescents [1,4,5,6].A positive relation has been suggested between the moral disengagement and antisocial, aggressive and violent behaviors [4,7] as well as peer victimization [8,9].
Research findings include that those who reveal moral disengagement behavior tend to be angrier and take revenge as well as revealing physical and verbal aggression [4,5,10,11].It is stated that the persons who reveal moral disengagement behave less empathic against their victims they harm, and feel less guilty [4,11].In addition, they also perceive the crime they commit at a lower level [12].
A wide literature states that the moral disengagement behavior is not only related to unwanted behaviors of children and adolescents but also to that of adults.Moral disengagement in adults is found related to gambling [13], violent acts against animals [14], and computer crime behaviors [15].Some researches show that the children whose moral disengagement behaviors are high in early adolescence period have high aggressive and violent behaviors and it continues till adolescence [11] and even till adulthood [16].Together with the increase of researches done on individual moral disengagement, researches were also being done in the collective moral disengagement.Collective moral disengagement is associated to persons change in behavior to be able to live with a class or a group and his/her disclosure of feelings related to becoming part of that class or group.A person could go through moral disengagement by not being in charge of the disruptive behaviors and protect him/herself of the potential negative behaviors [3,17].Researches done in the area shown individual and collective disengagement being positively associated with peer aggression and by-standing..As compared to individual moral disengagement taking responsibility and reveling desired behavior is difficult in collective moral disengagement.In the later, responsibility is shared in the group [18,19,20].
As seen, the collective moral disengagement behavior that develops in childhood threatens the behaviors of adolescents, youngsters and adults.As collective moral disengagement is mostly related to unwanted behavior; determining these behaviors in early ages and taking measures necessary to prevent them bear great importance.In Turkey, research about the collective moral disengagement has yet to be conducted when we examine the literature in the country.However, across the world, we can see many researches on moral disengagement about children [8,9,16], adolescents [9,11,21,22] and adults [13,14,15].In studies conducted with adolescents, the relations between bullying and aggressive behaviors, and collective moral disengagement were studied.Despite the importance of collective moral disengagement as being among reasons of antisocial behaviors observed in adolescence, there is not any research about collective moral disengagement in Turkey and a measurement tool to directly measure the collective moral disengagement, either.Absences of a scale that will measure collective moral disengagement, prompt awareness about antisocial behaviors in adolescence and preventing them have inspired the adaptation of this scale.
Collective Moral Disengagement Scale for Adolescents (CMDS) has been developed by Gini, Pozzoli and Bussey [18] to determine collective moral disengagement levels of adolescents.Collective moral disengagement scale has been formed with adaptation of items of individual moral disengagement scale -developed by Bandura et al.'s [4]-into classroom environment.As impact of peer and group is important in adolescents' behaviors, adaptation of moral disengagement scale that enables fast answering and use with other variables has been decided.The original form of the CMDS is composed of 17 items and one dimension.At the scale, a fivefold Likert "1-No one, 5-Everyone" rating is used.No reversely-scored item exists at the scale.Getting a high score at the scale indicates that the level of moral disengagement in a class environment is high, whereas low score means low level collective moral disengagement in the class environment.Factor loadings of the scale vary between 0.40-0.58.At the end of analysis made within the scope of reliability, the internal consistency coefficient is computed as 0.84.In addition, fit indexes were obtained at good level at the end of confirmatory factor analysis [18].
Purpose of this research is to enable adaptation of the CMDS into Turkish culture.Within this scope, in the first stage, exploratory factor analysis was performed to reveal structure of the scale in Turkish culture; and confirmatory factor analysis was performed to test the structure of the scale in Turkish culture in the second stage.With this study, it is thought that a reliable and valid assessment instrument which provides researchers and field experts with psychometric measurement will be available for studies that will determine the collective moral disengagement level of the adolescents.

Method
In the first stage of the research, translation of the scale to Turkish, exploratory factor analysis, internal consistency coefficient, test-retest methods were carried out as part of adaptation of the scale into Turkish culture and some values were obtained.

Participants
Adolescents who study in a province located in Turkey's Middle Black Sea Region form study group of this research.A total of 339 adolescents participated into the first stage of the research.Of those, 173 are girls (51%) and 166 boys (49%).The age of the participants differs between 10 and 17, and the age mean is 13.4.Test-retest was carried out with 64 adolescents.Of those, 30 are girls (46.9%) and 34 boys (53.1%).The ages of the adolescents are between 14 and 17 and the age mean is 15.6.

Process and Analysis of Data
Necessary permission has been taken through an e-mail communication with the writer of the scale for the adaptation of the CMDS into Turkish culture.Then, the translation of the scale -from English into Turkish -has been done by two experts working on psychological consultation.Later, the Turkish scale form was translated into English.Eight experts controlled the original English form and Turkish form for the suitability of the translation.The latest form of the scale was formed in accordance with suggestions by the experts.
In the first stage, construct validity of the CMDS was carried out with exploratory factor analysis (EFA).The reason of applying the EFA is to reveal structure of the CMDS in Turkish culture.In addition, the reliability of the CMDS was computed through Cronbach's alpha internal consistency and test-retest methods.Also, item analysis was made through item-total correlations method.

Findings Exploratory Factor Analysis
Kaiser-Meyer-Olkin (KMO=0.89)coefficient was computed in order to determine suitability of performing factor analysis of the CMDS.In addition, Bartlett Test (χ 2 =1521.645,df=105 (p<0.001) was performed.The exploratory factor analysis (EFA) was also performed to reveal the structure of research that was made with the adolescents.While performing the EFA, it was limited to one factor so as to obtain the structure same with the original one.Also, at the end of the EFA, analysis was renewed after excluding 4 th and 7 th items which are under 0.40 item factor loading value.The item factor loads obtained at the end of factor analysis that was conducted for CMDS are given at Table1.As it is seen at Table 1, the factor loadings of the CMDS change between 0.41-0.76.The eigenvalue of CMDS is 5.30, whereas the explained total variance is 35.31%.

Reliability
Reliability of the CMDS was studied by computing Cronbach's alpha internal consistency coefficient and test-retest reliability coefficient.Cronbach's alpha internal consistency coefficient of the CMDS was found as 0.86.The scale was carried out with 64 adolescents with 15 days of interval for test-retest reliability and retest reliability coefficient between the two applications was found as 0.84 (p<0.001).

Item Analysis
Item-total correlations changes between 0.34-0.69,mean change between 1.76-3.06and standard deviation between 1.13-1.65 at the end of item analysis.The obtained findings are given at Table 2.

Method
Confirmatory factor analysis and compliance validity were performed at the second stage of the research for the adaptation of the scale.

Participants
A total of 283 adolescents constitute the study group at this stage.Of those, 145 are girls (51.2%) and 138 boys (48.8%).The ages of the participants change between 10 and 15, and the age mean is 12.8.

Process and Analysis of Data
In the first research, confirmatory factor analysis (CFA) was carried out to test the confirmation of the structure that emerged at the end of exploratory factor analysis (EFA) which was performed for the CMDS.At this stage, compliance validity method was applied to measure the validity of the CMDS.The relation between the CMDS and Bullying Scale was computed with correlation analysis for the compliance validity.SPSS 21.0 and LİSREL 8.7 programs were used for validity and reliability analyses of the CMDS.

Data Collection Tools
In this section, there is information about Bullying Scale used for the compliance validity of the CMDS which is adapted into Turkish culture.
Bullying Scale: Bullying scale has been developed by Kutlu [23].The scale is composed of three dimensions; bullying, victim and enjoying life.The scale is composed of 19 items, with having no reverse-scored items.Fivefold Likert rating (1-Strongly Disagree, 5-Strongly Agree) was used at the scale.For example, Cronbach alpha's reliability coefficient was computed as 0.83 for bullying sub-scale, 0.86 for victim sub-scale and 0.70 for control items.In addition, three-factor structure of the scale gave goodness-of-fit-indexes at the end of confirmatory factor analysis.

Findings Confirmatory Factor Analysis
Confirmatory factor analysis process was applied to confirm the structure obtained at the end of exploratory factor analysis.When examining the fit index values of one-dimensional model at the confirmatory factor analysis, Chi-square Fit Test value (χ 2 =288.07,df=90, χ 2 /df=3.20,p=0.00) was found meaningful.In addition, other fit indexes were found as follow; RMSEA (Root Mean Square Error of Approximation)=0.08,RMR (Root Mean Squared Residual)=0.13,SRMR=0.06,GFI (Goodness of Fit Index)=0.89,AGFI (Adjusted Goodness of Fit Index)=0.85,CFI (Compared Fit Index)=0.94and NNFI (Non-normed Fit Index)=0.93.The factors taking part at the model are seen meaningful (p<0.05).The path diagram which shows standardized coefficients is given at Figure 1.
At the end of this analysis, the rate of tacit variances for explaining observed variances changes between 0.41-0.74for CMDS.It is seen that all items in the CMDS resulted in meaningful t value in explaining the tacit variances.

Compliance Validity
Bullying Scale and correlation coefficient were computed for the compliance validity of the CMDS.A positive meaningful relation was found at 0.41 level with Bullying Sub-scale of Bullying Scale and at 0.47 with Victim Sub-scale.

Conclusion and Discussion
In this research, the reliability and validity studies were carried out within the adaptation of Collective Moral Disengagement Scale for Adolescents into Turkish culture.While conducting the work of translation of the CMDS into Turkish, the scale adaptation stages were taken into consideration.Firstly, a lingual equivalence was maintained with obtaining views from experts in each step while translating the scale.
Exploratory and confirmatory factor analyses were performed with different study groups in order to determine the factor structures of the CMDS.The factor loadings showed a change between 0.41 and 0.76 at the end of EFA in the first stage of the research.As the fact that factor loading is 0.40 and over is a well-accepted value, the 4th and 7th items taking part in the original scale were extracted out of this scale as they took loading under this value [24].Thus, the CMDS which will be adapted into Turkish culture consists of 15 items.The variance of 15 items proved as 35 percent at the end of the factor analyses.
Although the total variance is above the 30 percent variance explanation, which is a well-accepted value for one-factor structure [24], it proves weak in explaining the total variance of the CMDS.On the other hand, according to the confirmatory factor analysis results, the Turkish form of the scale proved goodness-of-fit on the adolescents and it can be said the adapted form was confirmed.At the end of the analyses made within the reliability of the CMDS, the internal consistency coefficient was found 0.86 and the test-retest reliability coefficient which was performed with a two-week break was found as 0.84.This result is seen to be close to internal consistency coefficient obtained in the original scale.A positive meaningful relation was found at 0.41 level with Bullying Sub-scale of Bullying Scale and at 0.47 with Victim Sub-scale within the scope of compliance validity of the CMDS.When the literature was studied, a positive relation was determined between moral disengagement and bullying in adolescents [9,11,19,21,25,26].In addition, in a meta-analysis work, a positive relation was found between moral disengagement and bullying [19].In another study conducted with adolescents, 64 percent of students agreed the statement of "Bullying is part of childhood", 51 percent of them agreed on "Bullying makes people strong" and 56 percent of them supported "The students bullied deserve it" [9].
In studies conducted with Danish [21] and Italian adolescents [11], a positive relation was found between moral disengagement and bullying.In studies with Australian adolescents, too, a positive relation was found between moral disengagement and bullying, and it was revealed that the moral disengagement was determiner of the bullying [23].
Collective moral disengagement behaviors grounded in social learning theory are learned through interaction of a group of people who share particular interest and modeling them during socialization [27].In addition, when the class is considered as a social system, at classroom level collective moral disengagement could cause establishment of group norms, ways of thinking and behaving.Very specifically in adolescence peer influence is high and the determination of moral disengagement as a group norm can have detrimental effect on the behavior and perceptions of other students in the group [20].
Some studies have shown that high level of collective moral disengagement has become reason for unpleasant atmosphere in the classroom.This is in the form of adoption of aggressive behavior and a change in place of who could be in charge between the victim and persecutor.Passive students in the class have been affected by aggressive behaviors established from high level collective moral disengagement [20,28].
In research in which the moral disengagement levels of both bully and victims were studied, the level of moral disengagement of bullyboys was found higher.In addition, it was found that the victims produced aggrieved grounds and less moral rules.Also in this research, it was seen that the evaluation of moral responsibility among victims decreased and frequency of deviant behaviors increased [29].However, in another research, it was put forward that the mechanisms of moral disengagement did not have any relation to being a victim [22].In this research also it was found that those who are victimized behave as per the behaviors determined by the level of moral engagement in the class.
As a result, parallel to literature findings, this research saw a close relation between moral disengagement and bullying.With reference to obtained findings, it can be said the more the collective moral disengagement level increases in adolescents, the more the level of being both bullying and victim increases.
In conclusion, the researchers and field experts can use the scale in international comparison studies and moral disengagement studies in Turkey.In addition, the use of this scale in studies which will discuss the relationship between moral disengagement in adolescents and different parameters (violence, self-respect, personal characteristics etc.) strengthens the opinion that it will contribute to measuring power of this assessment instrument.

Figure 1 .
Figure 1.Path Diagram and factor loadings about Collective Moral Disengagement