پیش بینی ارتکاب مزاحمت سایبری در میان دانشجویان: استفاده از تئوری منطقی عمل
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|30376||2014||9 صفحه PDF||سفارش دهید||5985 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Computers in Human Behavior, Volume 36, July 2014, Pages 154–162
The present study tested the Theory of Reasoned Action (TRA) as an explanation for cyberbullying perpetration among 375 (128 male, 246 female) college students. Empathy toward cyberbullying victims was also included in the models. Participants completed the cyberbullying perpetration scale of the Cyberbullying Experiences Survey (Doane, Kelley, Chiang, & Padilla, 2013) that assesses four types of cyberbullying (deception, malice, public humiliation, and unwanted contact). Across all four models, results showed that lower empathy toward cyberbullying victims predicted more favorable attitudes toward cyberbullying perpetration, more favorable attitudes toward cyberbullying predicted higher intentions to cyberbully, and higher cyberbullying intentions predicted more frequent perpetration of cyberbullying behaviors. Injunctive norms regarding cyberbullying (e.g., perception of peers’ approval of cyberbullying perpetration) predicted intentions to engage in malice and unwanted contact behaviors. The results demonstrate that the TRA is a useful framework for understanding cyberbullying perpetration.
A growing body of literature shows that victims and perpetrators of cyberbullying are at greater risk for experiencing a myriad of mental health problems including depressive symptoms (Bonanno & Hymel, 2013), suicidal ideation (Bonanno and Hymel, 2013 and Hinduja and Patchin, 2010), and suicide attempts (Hinduja & Patchin, 2010). Despite awareness of the mental health risks associated with cyberbullying, few studies have applied a theoretical framework to understanding the perpetration of cyberbullying. To inform prevention/intervention of cyberbullying behaviors, we applied the Theory of Reasoned Action to explain cyberbullying perpetration among college students. 1.1. Cyberbullying prevalence Obtaining accurate estimates of the rates of cyberbullying is difficult due to variation in the definition of cyberbullying and discrepancies in its measurement (see Rivers & Noret, 2010, for a discussion). Across studies, the assessment windows (i.e., time frames over which the behaviors occurred), modes of communication included (e.g., cell phones, computer, e-mail), and specific types of behaviors assessed have been inconsistent. With measurement limitations in mind, in a review article, Tokunaga (2010) found 20–40% of youth reported that they had been cyberbullied. Although the percentages have varied, a number of studies have reported those who have been both a victim and a perpetrator of cyberbullying (e.g., about 10%, Hempbill, Tollit, & Kotevski, 2012; 12%, Hinduja & Patchin, 2009; 7%, Kowalski & Limber, 2007; 26%, Mishna, Khoury-Kassabri, Gadalla, & Daciuk, 2012). The overlap between cyberbullying victimization and perpetration could be in part explained by Hinduja and Patchin’s (2009) study which found that revenge against bullies was the most frequently reported reason for cyberbullying perpetration. Fewer studies have examined college students’ experiences of cyberbullying; however, recent studies have found between 9% and 11% of U.S. college students have been “cyberbullied” (Kraft and Wang, 2010, Schenk and Fremouw, 2012 and Walker et al., 2011) or have experienced repeated harassment, insults, or threats through e-mail or instant messaging (Finn, 2004). Finding slightly higher estimates of cyberbullying victimization, MacDonald and Roberts-Pittman (2010) found 21.9% of college students had been a victim of cyberbullying, whereas 8.6% had been a perpetrator of cyberbullying. In contrast, Aricak (2009) and Dilmaç (2009) found over half (54.4% and 55.3%, respectively) of Turkish college students had been cyberbullied in their lifetime, and approximately one-fifth (19.7% and 22.5%, respectively) had cyberbullied others. Although prevalence rates among college students vary widely, all studies suggest that a substantial portion of college students are victims and/or perpetrators of cyberbullying. 1.2. Theory of Reasoned Action Although several recent studies have examined rates of cyberbullying, few studies have employed established theories to explain cyberbullying behavior. One notable exception was Heirman and Walrave’s (2012) application of the Theory of Planned Behavior (Ajzen, 2012) in a sample of Belgian adolescents. When originally proposed, the Theory of Reasoned Action (TRA) was applied to behaviors for which individuals have complete control (Ajzen, 2012). TRA was later expanded to include the perception of one’s ability to perform a behavior (i.e., perceived behavioral control) and renamed the Theory of Planned Behavior (TPB; Ajzen, 2012). Because college students have access to the Internet and cell phones, nearly all college students have the ability to engage in cyberbullying. Specifically, in the United States, 98% of young adults use the Internet (Pew Internet & American Life Project, 2013), 97% of young adults use their cell phone for texting (Duggan & Rainie, 2012), and cyberbullying can be perpetrated anonymously. Therefore, we believed that the TRA was the most appropriate theory for our purposes. TRA posits that one’s attitude toward a behavior and subjective norms of the behavior influence behavioral intentions, which in turn influence behavior (Ajzen, 1985). 1.2.1. Attitudes toward behavior Attitudes involve how positively or negatively a person evaluates a behavior (Ajzen, 1985). According to Olweus (1993), bullies often have more positive attitudes toward violence and low empathy toward victims. Both studies of childhood (Elledge et al., 2013) and college students (Barlett and Gentile, 2012 and Boulton et al., 2012) have supported this argument. For instance, at both the individual and classroom level, Finnish children who had more positive attitudes toward victims were less likely to report having cyberbullied others (Elledge et al., 2013). Among college students in the United Kingdom, those with less accepting attitudes toward bullying were less likely to report engaging in social networking, text, physical, or verbal bullying (Boulton et al., 2012). In addition, less accepting attitudes toward perpetrators predicted less likelihood of verbal or social exclusion bullying (i.e., purposely excluding someone from friends or activities). Social exclusion bullying was also predicted by feeling sorry for victims. Recently, Barlett and Gentile (2012) found both more accepting attitudes toward strength differential (e.g., higher acceptability of “weaker” and “smaller” people cyberbullying bullies to get even) and more accepting attitudes toward anonymity (e.g., greater comfort level with cyberbullying individuals regardless of whether they know the person) predicted more positive attitudes toward cyberbullying perpetration, which in turn predicted cyberbullying perpetration. 1.2.2. Perceived norms Initially, the term subjective norms (i.e., the degree to which individuals perceive that others apply pressure to engage in the behavior) was used to describe perceived norms in the TRA (Ajzen, 1985). More recently, perceived norms have been expanded to include both subjective norms (i.e., now referred to as injunctive norms, the perception of others’ approval or disapproval of a behavior) and descriptive norms (i.e., the perception that others actually perform the behavior; Fishbein & Ajzen, 2010). Although their definitions of normative beliefs differed from the definitions used in the TRA, previous research has examined normative beliefs concerning cyberbullying (Ang et al., 2011, Werner et al., 2010 and Williams and Guerra, 2007). For example, in a sample of youth, Williams and Guerra found that believing bullying and bystander behavior (i.e., encouraging others to engage in bullying behaviors) is morally acceptable significantly predicted both traditional and Internet bullying. In addition, Barlett and Gentile (2012) found that cyberbullying reinforcement (i.e., positive reinforcement of cyberbullying perpetration) predicted cyberbullying perpetration. To our knowledge, no studies have examined the relationship between descriptive norms regarding cyberbullying (i.e., perceptions of others’ engagement in cyberbullying behavior) and cyberbullying behavior. However, a meta-analysis examining associations between attitudes, subjective norms, descriptive norms, perceived behavioral control, and intentions to engage in a wide range of behaviors found attitudes was the strongest predictor and descriptive norms was the second strongest predictor of intentions to engage in various behaviors (Rivis & Sheeran, 2003). Interestingly, the association between descriptive norms and intentions was stronger for younger (i.e., youth and undergraduate students) vs. older samples. 1.3. Empathy Although empathy is not explicitly included in the TRA or TPB, Ajzen (2011) has indicated that the association between other factors and specific behaviors may be mediated by the TRA/TPB constructs. Empathy appears to be associated with cyberbullying. For instance, as compared to adolescents not involved in cyberbullying, German adolescents who were either victims or perpetrators of cyberbullying reported lower levels of empathy (Schultze-Krumbholz & Scheithauer, 2009). Although Schultze-Krumbholz and Scheithauer measured overall empathy, three studies distinguished between affective and cognitive empathy as predictors of cyberbullying. Specifically, in a sample of Italian adolescents, Renati, Berrone, and Zanetti (2012) found that compared to victims of cyberbullying and compared to those not involved in cyberbullying, perpetrators of cyberbullying were significantly lower on affective empathy (i.e., experiencing others’ emotions). However, no differences in affective empathy were found between cyberbullying perpetrators and those who were both victims and perpetrators of cyberbullying. Cognitive empathy (i.e., understanding others’ emotional perspectives) did not differ significantly between groups. Moreover, among Turkish adolescents, Topcu and Erdur-Baker (2012) found that the combination of affective and cognitive empathy mediated the relationship between gender and cyberbullying perpetration. In a study of Singaporean adolescents, participants with low levels of affective empathy and high levels of cognitive empathy reported less frequent cyberbullying compared to those with low levels of affective empathy and low levels of cognitive empathy (Ang & Goh, 2010). In addition, among boys with high levels of affective empathy, boys with high cognitive empathy reported less frequent cyberbullying perpetration than boys with low cognitive empathy. In contrast, for girls with high levels of affective empathy, there was no difference in cyberbullying perpetration between those with high and low cognitive empathy. In contrast to studies that have assessed general empathy, Steffgen, König, Pfetsch, and Melzer (2011) examined empathy in the context of cyberbullying (e.g., “I find websites that make fun of other people funny/amusing;” p. 645) among adolescents in Luxembourg. Perpetrators of cyberbullying reported significantly lower levels of empathy associated with cyberbullying as compared to victims of cyberbullying and participants not involved in cyberbullying. 1.4. Theory-based explanation for cyberbullying behavior In the only published study to apply the TPB to cyberbullying behavior, Heirman and Walrave (2012) found more favorable cyberbullying attitudes, more positive perceived subjective norms about cyberbullying, and higher perceived behavioral control (i.e., “cyberbullying is easy to perform”) each predicted higher cyberbullying intentions. Importantly, attitudes toward cyberbullying was the strongest predictor of cyberbullying intentions. Moreover, intention to cyberbully predicted cyberbullying perpetration. 1.5. Present study The purpose of the present study was to test whether the TRA explained each of four types of cyberbullying perpetration (deception, malice, public humiliation, and unwanted contact) in a sample of college students. Hypothesis 1: Attitudes toward cyberbullying, injunctive norms, and descriptive norms were expected to predict cyberbullying intentions, which in turn were expected to predict cyberbullying behaviors. Hypothesis 2: The effect of empathy toward cyberbullying victims on cyberbullying intentions was expected to be mediated by the TRA constructs (i.e., attitudes toward cyberbullying, injunctive norms, and descriptive norms).
نتیجه گیری انگلیسی
3. Results 3.1. Descriptives and correlations Data from one participant, who endorsed the most extreme score for every behavior, intention, and norms item, was excluded from all analyses. Descriptive statistics and bivariate correlations between all study variables are shown in Table 1. Table 1. Descriptive statistics and bivariate correlations. Behavior Intentions Attitudes Injunctive norms Descriptive norms M SD α Deception Behavior – .27 .61 .87 Intentions .77⁎⁎ – .11 .50 .95 Attitudes .58⁎⁎ .64⁎⁎ – .21 .53 .91 Injunctive norms .53⁎⁎ .51⁎⁎ .58⁎⁎ – .46 .89 .87 Descriptive norms .46⁎⁎ .43⁎⁎ .46⁎⁎ .57⁎⁎ – .67 1.08 .89 Empathy −.09 −.08 −.25⁎⁎ −.20⁎⁎ −.12⁎ 3.77 1.52 .93 Malice Behavior – .83 .98 .91 Intentions .72⁎⁎ – .40 .86 .92 Attitudes .66⁎⁎ .69⁎⁎ – .43 .79 .96 Injunctive norms .65⁎⁎ .65⁎⁎ .68⁎⁎ – .89 1.17 .94 Descriptive norms .56⁎⁎ .50⁎⁎ .53⁎⁎ .70⁎⁎ – 1.53 1.46 .95 Empathy −.32⁎⁎ −.33⁎⁎ −.40⁎⁎ −.39⁎⁎ −.23⁎⁎ 3.67 1.54 .98 Public humiliation Behavior – .24 .59 .84 Intentions .72⁎⁎ – .13 .56 .94 Attitudes .57⁎⁎ .70⁎⁎ – .26 .64 .86 Injunctive norms .45⁎⁎ .49⁎⁎ .66⁎⁎ – .55 1.02 .88 Descriptive norms .42⁎⁎ .39⁎⁎ .50⁎⁎ .62⁎⁎ – 1.01 1.36 .91 Empathy −.13⁎ −.13⁎ −.28⁎⁎ −.25⁎⁎ −.18⁎⁎ 3.81 1.49 .90 Unwanted contact Behavior – .13 .52 .97 Intentions .93⁎⁎ – .09 .46 .98 Attitudes .70⁎⁎ .72⁎⁎ – .16 .52 .96 Injunctive norms .66⁎⁎ .65⁎⁎ .67⁎⁎ – .28 .65 .94 Descriptive norms .48⁎⁎ .45⁎⁎ .47⁎⁎ .62⁎⁎ – .48 .91 .96 Empathy −.13⁎ −.12⁎ −.24⁎⁎ −.24⁎⁎ −.14⁎⁎ 3.87 1.51 .98 ⁎ p < .05. ⁎⁎ p < .01. Table options 3.2. Path models The direct effects (i.e., standardized regression coefficients) for each path analysis model are summarized in Fig. 2. Across all four models, the TRA-based model accounted for a substantial portion of the variance in both cyberbullying intentions (.465 < R2 < .581) and cyberbullying behaviors (.543 < R2 < .872). The strongest associations were between attitudes toward cyberbullying and cyberbullying intentions (.46 < βs < .68) and cyberbullying intentions and all four types of cyberbullying behaviors (.42 < βs < .85). Full-size image (68 K) Full-size image (26 K) Fig. 2. Shows the observed relationships between empathy, Theory of Reasoned Action constructs, and cyberbullying behaviors for (a) deception, (b) malice, (c) public humiliation, and (d) unwanted contact. All coefficients reflect standardized regression coefficients. All paths based on the Theory of Reasoned Action are shown as well as additional significant effects that were not necessarily predicted by the Theory of Reasoned Action. Significant effects are shown in bold typeface for emphasis. Although gender and age were entered as correlated exogenous variables predicting all other variables in the model, as the effects of demographics were not of primary importance, these paths are not shown for reasons of parsimony. Figure options The total, direct, and indirect effects (i.e., mediated effects) of attitudes toward cyberbullying, injunctive norms, and descriptive norms on cyberbullying behaviors, as well as the effects of empathy on cyberbullying intentions and behaviors are summarized in Table 2. Attitudes toward cyberbullying had a total effect on all four types of cyberbullying behaviors. In addition, for three of the four cyberbullying behaviors (i.e., deception, public humiliation, and unwanted contact), cyberbullying intentions “fully” mediated the cyberbullying attitudes-cyberbullying perpetration association. In contrast to the other models, results revealed both an indirect effect via intentions and a direct effect of attitudes toward cyberbullying on cyberbullying behaviors that involved malice. That is, more positive attitudes toward cyberbullying had a direct positive effect on the perpetration of cyberbullying behaviors that involved malice. Table 2. Summary of total, indirect, and direct effects of empathy, attitudes, injunctive norms, and descriptive norms on cyberbullying behaviors. Deception Malice Public humiliation Unwanted contact β β β β Attitudes → Behavior Total .403 .374 .450 .453 Indirect (Intentions) .329 .193 .415 .438 Direct .074 .181 .035 .015 Injunctive norms → Behavior Total .223 .260 .069 .340 Indirect (Intentions) .101 .123 .037 .274 Direct .122 .137 .033 .066 Descriptive norms → Behavior Total .155 .176 .163 .073 Indirect (Intentions) .065 .017 .018 .028 Direct .090 .159 .145 .044 Empathy → Behavior Total −.082 −.311 −.142 −.140 Indirect (total) −.101 −.300 −.141 −.143 Intentions .055 −.014 .041 .064 Attitudes −.018 −.069 −.010 −.004 Injunctive norms −.023 −.053 −.009 −.016 Descriptive norms −.010 −.037 −.029 −.007 Attitudes → Intentions −.078 −.074 −.120 −.110 Injunctive norms → Intentions −.020 −.048 −.010 −.066 Descriptive norms → Intentions −.008 −.004 −.004 −.004 Direct .019 −.012 −.001 .003 Empathy → Intentions Total −.081 −.336 −.151 −.137 Indirect (total) −.170 −.302 −.217 −.211 Attitudes −.126 −.178 −.196 −.129 Injunctive norms −.032 −.115 −.016 −.077 Descriptive norms −.012 −.010 −.006 −.005 Direct .089 −.033 .066 .075 Note. Significant effects are determined by a 95% Confidence Interval that does not contain zero and are underlined for emphasis. Table options Injunctive norms, that is, perceptions of peers’ approval of cyberbullying, had a total effect on deception, malice, and unwanted contact behaviors (but no significant total effect on public humiliation). For deception behaviors, neither the direct nor indirect effect via intentions reached statistical significance. However, for both malice and unwanted contact, cyberbullying intentions “fully” mediated the predictive effects of injunctive norms on these cyberbullying behaviors. Nonetheless, it should be noted that the size of the direct effect of injunctive norms on malice was of similar magnitude of the indirect effect (even though the direct effect was not significant). Descriptive norms, that is, perceptions of peers’ perpetration of cyberbullying behavior, had total effects on deception, malice, and public humiliation. Specifically, perceiving one’s peers as more likely to engage in these behaviors was associated with one’s own reports of engaging in these behaviors. Descriptive norms did not have a significant total effect on unwanted contact. For cyberbullying behaviors that involved deception, neither the direct nor indirect effect via cyberbullying intentions reached statistical significance. For both malice and public humiliation, only the direct effects of descriptive norms of these cyberbullying behaviors were significant. The pattern of relationships between empathy and TRA constructs was consistent across all models. Empathy predicted attitudes, injunctive norms, and descriptive norms in all four models. The predictive effects of empathy on cyberbullying intentions and cyberbullying behaviors were fully mediated by TRA constructs. Specifically, the predictive effects of empathy on cyberbullying intentions was significantly mediated by attitudes in all four models, and significantly mediated by injunctive norms in the malice and unwanted contact models. Similarly, the double-mediated path of empathy on cyberbullying behaviors via attitudes and intentions was significant across all four models. Other significant indirect effects of empathy on cyberbullying were not consistent across all models and are summarized in Table 2.