How Does Guilt, Influence and Attitudes Effect the Role We Play in Bullying? The Self-Perception Measure

Abstract

Variations in perceived feelings of guilt, influence, and attitudes can alter a person’s behavior. The following article focused on the development and evaluation of a measure that explored how these self-perceptions affect the behaviour of the various participant roles involved in bullying situations. The participant roles explored included the bully, assistant, reinforcer, victim, defender, and outsider. The initial measure started with 30-items; 10-items for each measure (guilt, influence, and attitudes). The principal component analysis helped reduced the total number of items to 15 with guilt, influence, and attitudes all broken up into two components. Internal guilt measured the respondent’s guilt based on their own actions, external guilt measured the level of guilt based on the presence of others. Internal influence measured the respondent’s perceived influence on others and external influence measured the influence of others on the respondent’s role. Internal attitudes measured a person’s attitudes towards bullying and external attitudes measured a person’s perceived disassociation between their attitudes and their role. The results showed acceptable to good reliability on all measures except internal influence. Future researchers exploring participant roles associated with bullying can use this measure to better understand the motives behind specific behaviors.

Bullying in its traditional form can be defined by repetitive behaviour that is aggressive, intentional, unprovoked, and where there is a clear imbalance of power favoring the perpetrator (Younan 2018). Traditional forms of bullying differ from cyberbullying in that the latter does not always consist of a power imbalance in favor of the perpetrator and the incident may be a one off (O'Moore and Minton 2011). Bullying is an increasingly important area for researchers because of the universal prevalence (Due et al. 2005). One of the largest studies examining the prevalence of bullying among various countries was done by Craig et al. (2009). Their study compared the prevalence of bullying and victimization amongst both boys and girls of different ages in 40 countries (N = 202,056). These findings were able to show that bullying is a universal problem with countries varying in prevalence. The prevalence of bullying varied across countries ranging from 8.6% in Sweden to 45.2% in Lithuania among boys, and from 4.8% in Sweden to 35.8% in Lithuania among girls. Bullying was also higher for boys in all 40 countries, and victimization was higher for girls in 29 of the 40 countries. These findings show bullying to be a universal problem that needs to be addressed as such.

The origins of research into bullying can be traced back to the 1960’s where bullying was referred to as ‘mobbing’ and observed as a collective form of aggression against others of the same species (Lorenz 1966). Researchers study bullying by examining the relationship between the bully and victim (Olweus 1989) or exploring a variety of participant roles and peer relationships associated within a bullying incident (Huitsing and Veenstra 2012; Salmivalli et al. 1996). Researchers have begun to recognize the significance of the peer group and how it may influence contexts (Owens et al. 2000; Stauffacher and DeHart 2006). This recognition has allowed researchers to understand that bullying is not a dyadic role that is engaged in by just bullies and victims. Rather it is a group process where other roles also contribute in various ways (Gini 2006; Salmivalli et al. 1996).

Although still overshadowed by research examining bullying through the (dyadic) relationship between the bully and victim, emerging research has begun to recognize the critical influence the social group has on bullying incidents (Huitsing and Veenstra 2012; Salmivalli et al. 1996). These studies have found that other roles could have a significant effect on the behavior of the bully. Salmivalli et al. (1996) examined bullying incidents in classroom settings by having students complete a questionnaire asking them to evaluate each member within their classroom and themselves on items relating to particular bullying roles. Their results showed that, besides the usual role of the bully and the victim, 87% of the time students were able to identify peers who could be recognized as a reinforcer to the bully. Reinforcers usually offer positive feedback to the bully by laughing, giving encouraging gestures or by being audience members. Some students were also an assistant to the bully (eagerly joins in the bullying once someone else has started it), a defender (comforts the victim or actively tries and make the bullying stop) and an outsider (those that withdraw from bullying situations, without taking sides with anyone).

Through an examination of the diversity of the literature, direct and indirect results have shown that guilt, influence, and attitudes dictate how a person behaves in certain situations (Boster et al. 1999; Chen et al. 2013; Menesini and Camodeca 2008; Nickerson et al. 2008; Olthof 2012; Pozzoli and Gini 2013; Roland and Galloway 2002; Roos et al. 2014; Whatley et al. 1999). What has not been explored is how much guilt each person felt, what influenced them to take a particular role, and their attitudes towards bullying while playing a particular role.

Exploring people’s perceptions and their behavior could help uncover why people engage in specific roles. Understanding this would help with creating a profile of each role involved in bullying incidents. Assessing these factors within a social group setting using the various participant roles established by Salmivalli et al. (1996) is a new way of exploring bullying that should also help to create a deeper understanding of these roles while aiding in the creation of future interventions aimed at reducing bullying. In the following paragraphs, a detailed background of these three measures is provided.

Guilt suggests that a person has done something wrong and indicates that they feel at fault about being involved (Ferguson 2005; Olthof et al. 2004). Guilt appears to decline among bullies compared to other roles associated with a bullying situation (Menesini and Camodeca 2008). While this is evident despite the severity or type of bullying perpetrated by bullies, susceptibility to guilt has been shown to increase prosocial behavior (Chapman et al. 1987; Chen et al. 2013; Menesini and Camodeca 2008; Olthof 2012; Roos et al. 2014). A possible explanation for why bullies feel less guilt could be diminished moral compassion (Gini et al. 2011). Studies examining gender differences in guilt have also noted that females tended to feel more guilt than males when engaging in bullying behavior (Roos et al. 2014).

Measures designed to examine guilt have consistently been developed by providing participants with fictional situations and asking them to gauge their level of guilt (Menesini and Camodeca 2008; Olthof et al. 2000; Roos et al. 2014). The shame and guilt questionnaire developed by Olthof et al. (2000) was designed to identify the moral emotions of shame and guilt in children. The questionnaire included questions relating to shame and guilt and provided participants with ten situations relating to both. Tagney and Dearing’s (2002) test of self-conscious effect for children (TOSCA-C) was another measure that included fictional situations. Participants were provided with 15 scenarios (ten negative and five positive) relating to everyday situations (e.g. “You trip in the cafeteria and you spill your friend’s milk”), and their responses were rated on a five-point Likert scale (from not at all likely to very likely).

A number of studies have shown that bullies feel less guilt compared to victims, outsiders and defenders (Menesini and Camodeca 2008; Olthof 2012) with suggestions that bullies show a diminished moral disengagement (Gini et al. 2011; Menesini et al. 2015; Thornberg et al. 2016). However, from a social perspective, other explanations could be justified for why bullies may feel less guilty. For example, in social environments where bullying is prevalent, the bully is often considered popular despite not being well-liked (Garandeau et al. 2011; Prinstein and Cillessen 2003). Therefore, other goals (i.e. to become popular) may outweigh a person’s feelings of guilt. Based on the previous research showing how social situations change behavior the 10 items developed for the guilt measure within this study was set out to try and understand how social situations may alter levels of guilt.

Responsibility and guilt have been shown to be related in that the more responsibility a person feels, the more guilt they feel (Olweus 1999). In social situations, some factors make people feel more or less responsible for taking action. Pozzoli and Gini (2013) examined the role of passive bystanders and active defenders and found that for people to change roles from passive bystander to active defender they had to hold a positive attitude toward the victim. This then leads to feelings of personal responsibility. From a social perspective, the level of responsibility one feels can be attributed to bystander apathy, which is based on other people choosing not to be involved in the situation (Darley and Latane 1968). Bullies have been found to be more favored in cases where there is a high level of bullying (Sentse et al. 2007). This could also impact the level of responsibility a person feels when deciding how to act. A critical area that has been not explored in much depth is how feelings of responsibility can influence people to behave in antisocial ways when taking various roles within a bullying incident. Rather, the research has focused on the responsibility of individuals to take action and defend the victim or try to stop the bullying (Menesini et al. 2003; Morrow and Downey 2013; Ttofi and Farrington 2011).

Humans are fundamentality motivated to create and maintain meaningful social relationships with others, and they are also motivated to form accurate perceptions of reality and react accordingly to develop and preserve meaningful social relationships and maintain a favorable self-concept (Cialdini and Goldstein 2004). This can help explain why people are obedient to authority or conform to social norms despite having an opposing attitude to their actual behavior (Asch and Guetzkow 1951; Milgram 1974). Whatley et al. (1999) demonstrated this by finding that individuals avoided feelings of shame and fear via diffusion of responsibility. Bandura’s (1999) socio-cognitive theory stated that personal knowledge was based on observing others’ behaviors within certain environments. Therefore, moral disengagement can be identified as a cognitive process allowing an individual to justify and rationalize negative actions (Burns et al. 2010). Research from Baldry (2004) has also shown that people hold a positive attitude toward victims. However, maintaining a positive attitude toward a victim is also not enough to determine whether or not a person will defend a victim (Nickerson et al. 2008; Salmivalli and Voeten 2004). Other factors mentioned in the previous paragraphs all play a part in how a person behaves.

Bullies have been shown to approve of bullying more than other roles (Salmivalli et al. 1998; Salmivalli and Voeten 2004). This could be due to bullies having a higher desire to be more dominant than others (Olthof et al. 2011). Salmivalli and Voeten (2004) found that defending the bullied victims and staying outside of bullying situations were related to anti-bullying attitudes and moral disapproval of bullying, while the opposite actions were found in those who participated in bullying others. Salmivalli and Voeten (2004) also found that classroom norms could strongly affect the attitude of females when engaging in bullying incidents, while male behavior was strongly affected by their attitude toward bullying. Females have also been shown to have stronger attitudes against bullying than males (Pozzoli and Gini 2013).

Baldry (2004) stated, “to learn about attitudes towards bullies and victims, measurements should refer to specific bullying episodes” (p. 592). The measure presented in this study was designed to specifically explore one incident of bullying although goes beyond just exploring attitudes to also looking at guilt and influence.

When social groups are formed, people begin to assimilate by developing attitudes and behaviors that favor that group (Harris and Pinker 2011). These behaviors are so powerful that, even when individuals understand that their attitude or behavior may not be socially appropriate, the need to fit in within particular groups may prevent them from going against it (Nesdale et al. 2004). These behavioral patterns (e.g. contributing to bullying others) are not exclusive to groups of friends in which its members chose each individual. In classroom settings, for example, students are reported to be less likely to intervene and help the victim due to the risk of also being targeted by others (Sentse et al. 2007). Further, the social environment may change attitudes and behavior related to diffusion of responsibility whereby if no one outwardly helps the victim, bystanders feel it is not their responsibility to intervene (Darley and Latane 1968). If people see that no one is helping the victim, then people might feel that it is not their responsibility to intervene and help. To explore these previous findings further the newly created measure included 10 items asking participants about their attitudes towards bullying based on situational factors.

The aim of this study was to develop and evaluate a measure that could examine the self-perception differences in guilt, influence, and attitudes of different participant roles. The development of this measure would aid in exploring bullying from a social perspective by focusing on the relationship between peers within bullying incidents. Understanding how bullying can be affected by a person’s social group would also aid in developing and implementing successful intervention programs.

Method

Participants

Initially, a total of 200 participants responded, however, 50 were excluded from any analysis. A total of 18 participants were excluded because they were under the required age (18 years old), and 32 did not provide adequate responses (either clearly identifying participant roles or because of response bias). This, therefore, gave a 75% response rate comprising 150 adults (M = 29.39, range 18–66 years) from 25 different countries. The majority of the participants were from Australia (45%) and the United States of America (24%). Recruitment was done through the use of social networking websites, flyers pinned up around the University and through word of mouth. At the beginning of the questionnaire, participants were asked whether they suffered from any memory, visual, or attention impairments. Those who did were prevented from completing the questionnaire.

Item Selection

Items for the Self-Perception Measure (SPM) were created by first reviewing the previous literature and their findings on factors relating to guilt, influence, and attitudes amongst bullying incidents. This resulted in a total of 30 items (Table 1). In order to obtain face and content validity a panel of three academics judged and approved the relevance of the measures and all the items. The steps used to develop the SPM were designed to ensure that all relevant dimensions of guilt, influence, and attitudes towards bullying were measured. Guilt, influence, and attitudes each consisted of 10 items. Construct validity was examined by subjecting the items to an exploratory factor analysis. Internal consistency of each item was measured by conducting a reliability analysis. The findings of these tests were detailed in the results section of this study.

Table 1.

The original 30-item set

Guilt: Observing guilt based on the social situation a person was in and social group role

1. I did not feel any guilty for my actions

2. Other people should have felt more guilt than me

3. I felt guilty after the incident

4. I felt guilty despite continuing my actions

5. I believed that the bully felt guilty for their actions

6. The roles of others made me feel less guilty

7. The more people that were involved made me feel less guilty

8. The more people that were involved made me feel guiltier

9. The role I played should have made others feel guiltier

10. I believed that I did not feel any guilt because others did not feel guilty

Influence: Observe the influences that could have had an effect on oneself

1. I felt pressure from peers to act in a certain way

2. The role I played had an influence on others

3. I believed that anything I did would impact the person after the bullying stopped

4. There was no influence on my decisions to act in a certain way

5. I believed that the role I played would make me more popular

6. The influence of others made me act that way

7. I believed that I could have changed my role if I wanted to

8. The role I played made more people join in the bullying

9. I believed that anything I did would not impact the person after the bullying stopped

10. The role I played influenced others to get involved in the incident

Attitudes: Observe people’s attitudes towards bullying based on the participant roles

1. I believed it was someone else’s problem to intervene and stop the bullying

2. I believed it was necessary to behave in the way I did

3. I believed that If students get bullied, they deserved it

4. I felt that I was the most dominant in that incident

5. I understood that what I was doing was wrong despite my actions

6. I believed that the situation I was in had a serious impact on others

7. I approved of bullying

8. I disapproved of bullying despite my role

9. My attitudes were different to the role I played

10. There was a clear social group process to the bullying incident

Procedure

The questionnaire was advertised on social networking sites and flyers around the University. The questionnaire began with a series of demographic questions (e.g. gender, age, country of residence and occupation). Participants were randomly assigned to one of six participant roles (i.e. bully, assistant, reinforcer, victim, defender or outsider) and asked if they could reflect on an incident from their past in which they played that particular role. Participants unable to reflect on an incident from their past using that assigned role were given the ability to choose one of the other roles. Participants were also instructed to report in the questionnaire whether the incident was a physical or verbal form of bullying (which reflects traditional forms of bullying). A definition of each role was also provided so that participants understood clearly what the role entailed (Salmivalli et al. 1996). This was then followed by the 30 items the participant’s self-perceived guilt, influence, and attitudes. The items were responded to using a three-point Likert scale: (1 = disagree, 2 = neither agree nor disagree, 3 = agree) in which participants had to rate whether they agreed or disagreed with each item.

Results

Analysis for the SPM consisted of both a factor and reliability analysis, both of which were conducted on all measures. Before performing any analysis, the suitability of data for factor analysis was assessed. Inspection of the correlation matrix revealed the presence of many coefficients of .3 and above.

Guilt

A factor analysis was first conducted for guilt. The Kaiser-Meyer-Oklin (KMO) value was .61, just exceeding the recommended value of .6 (Kaiser 1970, 1974). The Bartlett’s Test of Sphericity (Bartlett 1954) reached statistical significance (p < .05), supporting the factorability of the correlation matrix. The Principal Components analysis revealed the presence of two components with eigenvalues exceeding 1, explaining 41.92% of the variance. An inspection of the scree plot also revealed a clear break after the second component. Using Cattell's (1966) scree test, it was therefore decided to retain the two components. A Varimax rotation was then performed (presented in Table 2). Both components showed some strong loadings with the majority of variables loading substantially on one of the two components. Component 1, respectfully labeled internal guilt contributed to 22.71% and component 2, respectfully titled external guilt, contributed to 19.21%. A reliability analysis showed that the internal guilt measure (“I felt guilty after the incident” and “I felt guilty despite continuing my actions”) had good internal consistency with a Cronbach alpha coefficient reported of .86 and for external guilt (“the more people that were involved made me feel less guilty,” “the roles of others made me feel less guilty”) had acceptable internal consistency, with a Cronbach alpha coefficient reported of .64.

Table 2.

Factor loadings based on a principal components analysis with a varimax rotation for items for the guilt measure of the SPM (N = 150)

Item	Internal guilt	External guilt
1. I felt guilty after the incident	.85
2. I felt guilty despite continuing my actions	.84
3. The role I played should have made others feel guiltier	−.46
4. Other people should have felt more guilt than me	−.46
5. I did not feel any guilt for my actions	−.38
6. The more people that were involved made me feel less guilty		.84
7. The roles of others made me feel less guilty		.70
8. I believed that I did not feel guilty because others did not feel guilty		.60
9. The more people that were involved made me feel guiltier	.41	−.50
10. I believed that the bully felt guilty for their actions

Influence

The factor analysis for the influence measure revealed the KMO value to be .7, exceeding the recommended value of .6 (Kaiser 1970, 1974). The Bartlett’s Test of Sphericity (Bartlett, 1954) reached statistical significance, supporting the factorability of the correlation matrix. The Principal Components analysis revealed the presence of two components with eigenvalues exceeding one, explaining 44.06% of the variance. An inspection of the scree plot also revealed a clear break after the second component. Using Cattell’s (1966) scree test, it was therefore decided to retain the two components. A Varimax rotation was then performed (presented in Table 3). Both components showed some strong loadings with a majority of the variables loading substantially on one of the two components. Component one (internal influence) contributed 29.23% and component two (external influence) contributed 14.83%. The Influence subscale indicated a poor internal consistency for internal influence (“the role I played had an influence on others” and “the role I played influenced others to get involved in the incident”) with a Cronbach alpha coefficient reported of 0.52; external influence (“the influence of others made me act that way,” “I felt pressure from peers to act in a certain way,” “I believed that the role I played would make me more popular,” and “the role I played made more people join in the bullying”) had good internal consistency, with a Cronbach alpha coefficient reported of 0.79.

Table 3.

Factor loadings based on a principal components analysis with a varimax rotation for items for the influence measure of the SPM (N = 150)

	Internal influence	External influence
1. The influence of others made me act that way		.82
2. Felt pressure from peers to act in a certain way		.76
3. I believed that the role I played would make me more popular		.74
4. The role I played made more people join in the bullying		.71
5. I believed that I could have changed my role if I wanted to		.37
6. There was no influence on my decision to act in a certain way		.33
7. I believed that anything I did would not impact the person after the bullying stopped	.71
8. The role I played had an influence on others	.65
9. The role I played influenced others to get involved in the incident	.59	.31
10. I believed that anything I did would not impact the person after the bullying stopped	−.47	.33

Attitudes

The factor analysis for Attitudes showed a KMO value of .63, just exceeding the recommended value of .6 (Kaiser 1970, 1974). The Bartlett’s Test of Sphericity (Bartlett, 1954) reached statistical significance, supporting the factorability of the correlation matrix. The Principal Components analysis revealed the presence of one component with eigenvalues exceeding one, explaining 64.62% of the variance. An inspection of the scree plot also revealed a clear break after the second component. Using Cattell’s (1966) scree test, it was therefore decided to retain the two components. A Varimax rotation was then performed (presented in Table 4). Both components showed some strong loadings with the majority of variables loading substantially on one of the two components. Component one (external attitudes) contributed 33.14%, and component two (internal attitudes) contributed 31.47%. A reliability analysis identified good internal consistency for both measures with a Cronbach alpha of .76 for external attitudes (“I disapproved of bullying,” “I approved of bullying,” and “ I believed that if students got bullied, they deserved it”) and .70 for internal attitudes (“I understood what I was doing despite my action” and “my attitudes were different to the role I played”. The item “I disapproved of bullying despite my role” was reverse coded).

Table 4.

Factor loadings based on a principal components analysis with a varimax rotation for items for the attitudes measure of the SPM (N = 150)

Attitudes	Internal attitudes	External attitudes
1. I understood that what I was doing was wrong despite my actions		.88
2. My attitudes were different to the role I played		.82
3. I believed it was necessary to behave in the way that I did		.70
4. I disapproved of bullying despite my role	.82
5. I approved of bullying	.78
6. I believed that if students got bullied, then they deserved it	.75

A correlation matrix was computed to assess the relationship between the six measures of the SPM; the results are presented in Table 5. The results showed a moderate correlation between internal guilt and external influence (r = .55, n = 150, p = <.01), internal guilt and external attitudes (r = .58, n = 150, p = <.01), external influence and external attitudes (r = .50, n = 150, p = <.01), and a weak negative correlation between external influence and personal attitudes (r = −.34, n = 150, p = <.01).

Table 5.

Pearson product correlation matrix for the SPM

	1	2	3	4	5	6
1. Internal guilt	1	.07	.10	.55**	−.12	.58**
2. External guilt		1	.22**	.18*	−.21**	.17*
3. Internal influence			1	.28**	−.24**	.11
4. External influence				1	−.34**	−.50**
5. External attitudes					1	−.19*
6. Internal attitudes						1

*p < 0.05, **p < 0.01

The relationships between all items of the SPM were investigated using Pearson product-moment correlation coefficients. Preliminary analyses were performed to ensure no violation of the assumptions of normality, linearity, and homoscedasticity. Inter-item correlations are presented in Table 6 and the variable names for each item are presented in Table 7. The results showed that was a strong, positive inter-item correlations between the items “I understood that what I was doing was wrong despite my actions” and “I felt guilty after the incident” (r = .50, n = 150, p < .001); “I felt pressure from peers to act in a certain way” and “I felt guilty after the incident” (r = .50, n = 150, p < .001); “I felt guilty despite continuing my actions” and “I felt guilty after the incident” (r = .75, n = 150, p < .001); “I felt pressure from peers to act in a certain way” and “I felt guilty despite continuing my actions” (r = .53, n = 150, p < .001); “I understood that what I was doing was wrong despite my actions” and “I felt guilty despite continuing my actions” (r = .52, n = 150, p < .001); “My attitudes were different to the role I played” and “I felt guilty despite continuing my actions” (r = .50, n = 150, p < .001); “I believed that the role I played would make me more popular” and “I felt pressure from peers to act in a certain way” (r = .55, n = 150, p < .001); and “My attitudes were different to the role I played” and “I understood that what I was doing was wrong despite my actions” (r = .62, n = 150, p < .001).

Table 6.

Inter-correlations between all items

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21	22	23	24	25	26	27	28	29	30
1.	1
2.	.15	1
3.	−.17	−.21	1
4.	−.20	−.21	.75	1
5.	−.10	−.08	−.14	−.12	1
6.	−.01	−.07	.06	.02	.11	1
7.	−.13	−.08	.30	.27	−.03	−.06	1
8.	.02	−.11	.11	.04	.14	.48	−.33	1
9.	.12	.21	−.23	−.23	.09	.01	−.06	−.05	1
10.	.01	−.00	−.00	.02	.03	.25	−.20	.36	−.01	1
11.	−.04	−.26	.50	.53	−.09	.17	.29	.03	−.16	.10	1
12.	−.06	−.27	.00	.08	.23	.16	.01	.25	.10	.09	.11	1
13.	−.19	−.15	.05	.17	.13	−.01	.15	−.01	−.01	.05	.15	.23	1
14.	−.16	−.13	.22	.17	.15	.01	.16	−.08	−.03	−.07	.23	.12	−.04	1
15.	−.09	−.26	.35	.36	.08	.22	.13	.13	−.09	.11	.55	.11	.05	.12	1
16.	−.06	−.14	.44	.43	−.00	.16	.22	.05	−.12	.18	.55	.14	−.01	.27	.52	1
17.	−.00	−.14	.20	.21	.04	.20	.06	.21	−.23	.18	.16	.13	.07	.06	.22	.26	1
18.	−.09	−.12	.27	.36	.04	.09	.12	.12	−.07	.32	.40	.12	.05	.15	.42	.46	.22	1
19.	.15	.18	−.04	−.01	−.11	.04	−.00	.02	.02	.26	.15	−.03	−.23	−.10	.02	.21	.01	.20	1
20	.05	−.19	.06	.17	.09	.09	.12	.11	−.01	.1	.13	.36	.17	−.07	.30	.14	.09	.37	.01	1
21	.01	.04	.05	.06	−.12	.07	−.00	.06	.02	.22	.26	−.08	−.09	−.00	.17	.11	.03	.12	.09	−.04	1
22	−.00	−.14	.46	.35	−.03	.17	.03	.15	−.19	−.02	.28	.03	−.12	.24	.24	.30	.29	.23	−.02	−.02	−.10	1
23	−.09	−.15	.15	.05	.02	.24	−.12	.32	−.01	.31	.21	.14	−.14	−.02	.32	.19	.16	.30	.08	.21	.20	.12	1
24	−.20	−.22	−.05	−.09	.26	.10	.00	.12	.03	.20	−.00	.30	.13	.06	.16	−.00	.04	.16	−.02	.34	.02	.03	.32	1
25	−.11	−.14	.49	.52	−.04	.21	.06	.15	−.23	.27	.36	.06	−.00	.14	.38	.37	.32	.29	−.06	.12	.09	.42	.12	.06	1
26	−.09	.05	.11	.17	.00	.02	.10	.03	.09	.07	.12	.10	.19	.05	.13	.06	−.01	.15	−.17	.20	−.14	−.07	−.02	.21	.21	1
27	.01	.08	.00	−.01	−.04	.11	.03	.10	−.01	.32	.14	.13	.05	−.08	.17	.15	.15	.15	.24	.13	.28	.12	.39	.36	.05	−.01	1
28	.13	.04	.01	−.13	−.01	.13	−.08	.14	−.01	.17	.03	.03	−.07	−.10	.08	.07	.05	.00	.18	.04	.14	.06	.44	.02	−.12	−.18	.49	1
29	−.14	−.21	.46	.50	−.02	.11	.09	.05	−.19	.09	.42	.06	.07	.21	.33	.42	.26	.22	−.00	.07	.05	.31	.02	.06	.62	.10	−.01	−.17	1
30	.09	.01	.11	.09	−.03	.09	−.00	.11	.01	.10	.14	.12	−.03	.06	.23	.20	−.04	.10	.11	.06	.07	−.11	.08	−.09	.06	.16	−.04	.01	.11	1

Note. Significant correlations bolded (p < .05)

Table 7.

Items for the Inter-item correlations

Items:

1. I did not feel any guilt for my actions

2. Other people should have felt more guilt than me

3. I felt guilty after the incident

4. I felt guilty despite continuing my actions

5. I believed that the bully felt guilty for their actions

6. The roles of others made me feel less guilty

7. The more people that were involved made me feel less guilty

8. The more people that were involved made me feel guiltier

9. The role I played should have made others feel guiltier

10. I believed that I did not feel guilty because others did not feel guilty

11. I felt pressure from peers to act in a certain way

12.The role I played had an influence on others

13. I believed that anything I did would impact the person after the bullying stopped

14. There was no influence on my decision to act in a certain way

15. I believed that the role I played would make me more popular

16. The influence of others made me act that way

17. I believed that I could have changed my role if I wanted to

18. The role I played made more people join in the bullying

19. I believed that anything I did would not impact the person after the bullying stopped

20. The role I played influenced others to get involved in the incident

21. I believed it was someone else’s problem to intervene and stop the bullying

22. I believed it was necessary to behave in the way that I did

23. I believed that if students got bullied, they deserved it

24. I felt that I was the most dominant person in that incident

25. I understood that what I was doing was wrong despite my actions

26. I believed that the situation I was in had a serious impact on others

27. I approved of bullying

28. I disapproved of bullying despite my role

29. My attitudes were different to the role I played

30. There was a clear social group process to the bullying incident

Discussion

This study set out to develop and evaluate a new measure for examining different self-perceptions (guilt, influence, and attitudes) toward bullying from various participant roles in the hope that it would allow for an examination of the social group and the ways in which people’s perceptions of others’ involvement influenced their attitudes and behaviors.

Guilt was divided into two forms, namely internal and external guilt. Both internal guilt and external guilt consisted of two items. Both forms are important because they show possible differences in guilt due to internal and external factors (Ferguson 2005; Olthof et al. 2004). Internal guilt examined a person’s guilt while playing a particular role while external guilt examined whether there were differences in the level of guilt based on peer presence. As peers are present in a majority of bullying situations, it would be expected that there might be a decreased level of guilt among the participant roles in such cases (Huitsing and Veenstra 2012; Salmivalli et al., 1996). The level of guilt associated with a particular role could also be attributed to the level of responsibility a person feels. Research has shown that feeling a high level of responsibility to intervene is attributed to having a high level of guilt (Boster et al. 1999). It would also be expected that as more people are involved, the less responsibility to intervene might be felt by the bully group (bully, assistant, and reinforcer) or outsiders (Boster et al. 1999). The presence of others would also increase the level of external guilt (decreasing the level of guilt), as people would believe that it would not be their responsibility alone to help and intervene (Grissinger 2012).

A common measure that has been used to examine guilt within the bullying literature is the shame and guilt questionnaire, constructed to examine shame and guilt based on 10 hypothetical scenarios (Olthof et al. 2000). The SPM is different from the shame and guilt questionnaire in a number of respects. Firstly, the SPM explores real incidents of bullying, while the shame and guilt questionnaire explores overall shame and guilt and not just that relating to bullying (e.g. Tommy and his mother are sitting around the table. Tommy drinks his juice and his mother is writing a letter. Then Tommy calls for the cat and the cat jumps on to the table, knocking over Tommy’s glass. Tommy’s juice spills all over his mother’s letter). Secondly, it explores both internal and external factors affecting the level of guilt. Previous research using the SAG questionnaire has shown that bullies felt less guilt compared to victims, outsiders, and defenders (Menesini and Camodeca 2008; Olthof 2012). Possible explanations for bullies feeling less guilt include that bullies may show greater moral disengagement (Gini et al. 2011; Menesini et al. 2015; Thornberg et al. 2016).

Internal influence consisted of four items and examined whether a person’s participant role influenced others to get involved. External influence consisted of two items and focused on influences that would lead a person to behave in a certain way. Both types of influences are important because to maintain meaningful social relationships, people could play a particular role (Cialdini and Goldstein 2004). Therefore, the assistant and reinforcer could perceive themselves to have high levels of external influence to conform to social norms while victims may feel external influence from the bully. Previous research has stated that those in high power or those high in social status may have more control over influencing people’s behavior (Juvonen and Galvan 2008). If this is found to be true, then the roles of the bully, assistant, and reinforcer should have high levels of internal influence. Bullies may also feel external influence to continue bullying others in order to maintain their power and status (Vaillancourt et al. 2003). The socio-cognitive theory also states that knowledge is gained from observing others in different social interactions and experiences (Bandura 1999). Thus, the behavior of others could allow people to mimic similar roles, whether that be bullying or defending. While previous studies have shown behaviors to be influenced by friendships and peer groups (Lodder et al. 2016), no measure has been developed to explore whether a person believed they influenced others’ behaviors or whether others influenced their behavior.

The internal attitudes measure consisted of two items that explored whether there was a dissociation between attitudes and behaviors. Previous studies exploring attitudes toward bullying have used similar measures to those in the SPM. Boulton, Trueman, and Flemington’s (2002) attitudes measure had participants rate whether they agreed, neither agreed nor disagreed, or disagreed with 12 items (e.g. “students should be allowed to bully others who deserve it”). The higher the score, the more positive their attitude toward bullying. Using this measure, they found that bullies were more likely to hold positive attitudes toward bullying. Similarly, Salmivalli and Voeten’s (2004) measure had participants rate on a five-point Likert scale 10 items relating to bullying (e.g., “one should help the bullied victim”), with a higher score indicating a more negative attitude toward bullying. All three attitude measures explored the same constructs, with the only difference being the number of items. Salmivalli and Voeten (2004), using their attitude measure, found the role of the bully to have the highest approval of bullying, followed by the assistant and then the reinforcer. Despite the level of approval from the assistant and reinforcer, their decision to continue to act antisocially could be due to their fear of possibly being excluded from the social group or perhaps targeted themselves if they do not conform (Atlas and Pepler 1998; O'Connell et al. 1999; Osumi et al. 2016; Pepler and Craig 1995; Turner 1999). Therefore, this measure is also expected to show different results based on different participant roles.

Research has contested that having a negative attitude towards bullying is known to inhibit bullying behavior (Rigby 2005) although it is not an essential component to warrant defending a victim (Nickerson et al. 2008; Salmivalli and Voeten 2004). Other factors could play a potential role, including not wanting to stray from social norms if bullying is common (especially within social groups), power differentials, and the possible fear of being targeted themselves (Dijkstra et al. 2008; Pozzoli and Gini 2010; Pozzoli et al. 2012; Pronk et al. 2013). These findings provide support for the exploration of the external attitudes measure.

The external attitudes measure consisted of three items. Previous researchers exploring attitudes toward bullying have only created measures focusing on a person’s approval or disapproval of bullying (Boulton et al., 2002; Rigby & Slee, 1991: Salmivalli and Voeten 2004). This could be because the majority of previous research on bullying has focused on characterizing specific participant roles using dispositional traits instead of exploring how attitudes and behaviors may change due to external factors like the peer group.

The research within this article was able to develop a measure that could examine the social perspective of bullying through the examination of self-perceptions of different participant roles. The established measure consisted of internal guilt, external guilt, internal influence, external influence, personal attitudes, and external attitudes. This measure can be used to examine how people fulfilling each bullying role perceived themselves while participating in a bullying incident and engaging in that role. Future researchers could use this measure on all those involved in a bullying incident to compare differences; this would allow for a greater understanding of how the social ecology affected behaviors and attitudes. This would then aid in the development of more successful intervention programs.

Compliance with Ethical Standards

Disclosure of Interest

Authors have no conflicts of interest to declare.

Ethical Standards and Informed Consent

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation [RMIT University]. Informed consent was obtained from all patients for being included in the study.