The Role of Self-Control in Cyberbullying Bystander Behavior

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study is potentially important and certainly interesting. I have serious concerns related to the measured variables of the study and the analytic decisions taken which actually make it difficult for me to interpret the results. Thus, before I can offer a proper evaluation of the quality and contribution of the study, I would like to have the following answered properly:

1. I would like to see the measurement model and a figure of the bifactor model. I would like to know if the bifactor model is interpretable, if the specific factors are still in existence and so for. Also, the reliabilities of both. Bifactor models almost always are statistically superior to less parameterized models, so, unless I see exactly what was fit and what was actually found I cannot take it as granted that the bifactor model was the optimal model with these data.

2. For any CFA type of model I need to see the chi-square test values and the degrees of freedom. So the authors decided not to include any global statistics. Why?

3. What is the power for the CFA models?

4. Did the authors test for mutlivariate kurtosis? I believe AMOS presents Mardia's coefficient. The univariate skew/kurtosis estimates do not add anything given that items were modeled as indicators of latent constructs. However, min/max values are needed in Table 1.

5. Was there any missing data? if yes, how were they treated?

6. There is nothing controlling about the control variables as they were used in the present study. What did they control for? it would be much more informative if the measurement models would be tested for invariance for gender and age groups.

7. Similarly, it would be extremely more valuable if the figure 1 sem model is fit for both males and females with tests of invariance. I believe there is ample evidence that there are gender differences in bullying. So aggregate estimates likely wash out what is actually going on.

Again, I do believe the work is potentially interesting, but right now, there are too many methodological issues that prevent me from being more positive. If the authors provide a proper revision I would be happy to reconsider.

Author Response

Please see attached.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript deals with the important topic of the role of low self-control in bystander behaviors. I consider it extremely positive to check the applicability of classical theories that explain risky behaviors and breaking the law for risky behaviors in the digital world. What I find challenging in the introductory part, although the rationale is clearly laid out, is the lack of recent sources. Namely, although it is logical that for the general theory of crime whose beginnings go back to 1969, sources that are several decades old are used, for the field that refers to the interaction of an individual with intensively changing technology, sources that are almost 20 years old are simply outdated. Although it is clear to me that the authors refer to Suler, who, although starting from a psychodynamic perspective, emphasized disinhibition that can be linked to some key concepts of the General Theory of Crime, I find it challenging to use the definition of cyberbullying by Smith et al . Namely, they started from a simple transfer of Olweus's definition to the digital environment. After their definition, many authors pointed out significant differences (e.g. Menesini) - for example, the question of frequency from the position of victims is quite different, because by sharing some content, one incident in the digital world can hurt the victim multiple times (and bystanders often play a significant role in this ). The question of power has also been shifted from the difference in physical strength or superior status, which is more important for bullying, to anonymity and technical knowledge that will allow attacking the victim in cyberbullying. A large part of the terror for victims of cyberbullying comes from not knowing who the attacker is. Therefore, I recommend that the introductory part be supplemented with critically reviewed recent works - authors should take into account that the experience of the Covid-19 pandemic had a significant effect on cyberbullying, so it would be important to focus on works from the last 2-3 years. The introductory part also talks about the types of bystanders, while the measure itself (which I consider more appropriate) is based on data on three types of bystanders ' behavior (measured for each participant), and not on the categorization of participants into three groups of bystanders. I consider this kind of analysis more appropriate because it can be expected that young people in different cyberbullying incidents show different bystanders ' behavior. I suggest that this be emphasized in the introductory part, that is, that the hypothesis that a certain component of low self-control will be related to a certain behavior be commented on in view of this suggestion.

Figure 1 is folded into two pages and cannot be easily followed.

While reading the discussion, I wondered why participant’s experience with cyberbullying ( bullies, victims, uninvolved.) was not used as a control or perhaps moderator variable. I assume, since I assume it is a part larger study, that the authors also have this data, so I suggest that they consider this kind of processing as well. Now that I'm reading the restrictions, I see that the authors stated that they didn't monitor it, but I'm not sure if the restriction only applies to offline or online behaviors as well.

 

I would like to congratulate the authors on an interesting manuscript and I hope that my comments will be helpful.

Author Response

Please see attached. 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for attending to my comments and suggestions. There are few more things that I don't believe were addressed properly.

1. What are the between factor correlations in the CFA model?

2. Is the AIC in Table 4 for the second order model an error? it is 1610.7 should it be 161.07?

3. Given the significance in Mardia's coefficients did the authors use MLR or MLMV estimators? I believe some of these are available in AMOS.

4. Were gender and age regressed on the latent factor variable only?

5. In figure 1, this sem model is not a bifactor model. If is its a bifactor model, the general factor should contain factor loadings for all measured items. The LSC has only three paths and those with no arrow so we cannot even tell if they are loadings or covariances. This is very confusing, please describe the bifactor model properly and plot it properly. 

6. Why is bullying regressed on gender only and not age? why is the general factor regressed on age and not gender?

7. Statistical power was not addressed.

Author Response

Please see attached.

Author Response File: Author Response.pdf