The Positive Expectation of What the Future Holds Analysis

4 attachments

Reviewers/Editor Comments:

Reviewer #1: Introduction/Problem: The research problem of bullying in the US if fairly well specified, although references to world-wide rates are less compelling. The problem in terms of children and adolescents in the US, in schools and other venues, should be the focus for this manuscript. Unfortunately, this problem not described in the context of school nursing, school-based health, or pediatric nurse practitioner practice until the implications. It would be more effective to introduce early on the setting for addressing the problem and who might carry that out, in order to further establish the importance of the problem. This would add to the importance established by the description of concurrent and long-term consequences of bullying and make it more clear what can be done about it.  The Positive Expectation of What the Future Holds Analysis
There is no theoretical framework identified. The various types of bullying are well defined, and include cyber-bullying. Rates of cyber-bullying should be found. The characteristics of bullies are nicely reported as is the nature of the interpersonal dynamics that attend bullying. The central concepts are adequately defined and discussed in relation to the problem, as separate sections in the introduction. Evidence for future expectations and attitudes towards violence in relation to bullying are supported with limited evidence (2 sources each) that these personal characteristics contribute to bullying but the argument is plausible. The introduction would benefit from a review of more literature to speak to interventions that have been developed for bullying and what they target, as well as how effective they have been found. In this way, critique of the literature could be integrated as well. Then to introduce what little is known about future expectations and attitude toward violence serves an argument for the study but suggests how these two characteristics may have a role. In other words, given how little is known, there needs to be a more clear argument for researching this gap to find out how the characteristics relate and perhaps point to an arena for intervention. The purpose and hypotheses flow logically from the introduction. The Positive Expectation of What the Future Holds Analysis

Design/Method: The setting for the study is a middle school setting. Again, the manuscript would be stronger if the problem and introduction addressed the school setting and middle school-aged adolescents as the population of interest, as specific developmental dynamics characterize this age group and an understanding of the phenomenon. The sample and the population from which it was drawn needs description; it is not possible to know how representative this sample is of the middle school. I was struck by there being more female participants than male, although the evidence suggests boys bully more, especially physically. Did more boys refuse to participate? How representative are the proportions of African American and Caucasian middle school students? Were repeated opportunities provided for participation? It would have been better to address this imbalance during the recruitment and data collection phase, perhaps including more schools. Also, the n’s of the various sample characteristics need to be provided either in the text or in the table; it is fine to report proportions for race and socioeconomic status but the n’s need to be reported as well in at least one place in the manuscript. Again, how representative this sample was of the school is not evident and generalization is limited. The Positive Expectation of What the Future Holds Analysis

The procedures are fairly precisely detailed, although it would help to know the range of time it took to complete the survey as well as the average (with SD). The measures and their items and response formats are adequately described. Sources for items are also provided. The internal consistency estimates are reported and adequate, although the attitude toward violence is a little lower. Some estimate of test-retest reliability would have been advisable. The descriptive statistics for the measures were included in the table with the sample characteristics; these should be separated or included in the text describing the measures. The demographic variables are described but may be better described with the sample? How they are coded really does not need to be described here. Maternal education is reported as a variable although not addressed to make it relevant to the topic; in fact, socioeconomic context was not addressed in the introduction.

The description of the approaches to data analysis should be integrated with the results. In other words, integrate what analysis was used before the specific results are reported.

Results: The sample needs to be described in the method, with the table referenced and data included as appropriate in that text specific to the sample OR do not report descriptive data in the method and include it here; a choice needs to be made. Again, n’s are not included in the text or the table; they should be in one or the other. Given there were no gender differences (or at least no association between variables and gender), why not collapse the genders and expand the sample?
The lack of correlation between demographic characteristics and the study variables nicely supports that these personal characteristics are otherwise developed. The multivariate analysis was fairly clearly described.

Discussion: The discussion proceeds nicely although a lot of new literature is introduced. The introduction needs to be expanded by including the literature that is reported in the discussion; doing so may solve the problems of the literature review in the introduction and strengthen the problem. There are suggestions for future research and the methodological limitations are noted (although not the gender imbalance). In the discussion, risk and promotive factors are introduced; this suggests a theoretical framework that is not uncommon in school-based or public health research and could be presented as an organizing framework up front.
Implications: The implications may be better integrated with the discussion of findings so the recommendations or implications for practice are more closely aligned with the findings and related literature. As written separately here, they seem a bit global and not tightly linked—so more speculative.

Summary/Conclusions: This section is missing.

 

 

 

Reviewer #2: This manuscript presents a mediation analysis of attitudes towards violence on the relationship between future expectations and bullying behavior during early adolescence. The manuscript is nicely written, but the topic does not fit the aim and scope of this journal—”to circulate scientific papers in nursing to improve care, alleviate suffering, and advance well-being.” I would like to suggest the authors submit it to an educational, psychological, or behavioral journal. Meanwhile, I would also like to provide a few methodological comments and suggestions for improving the manuscript as follows, by page numbers:

1) Page 4, Lines 20-21: Physical bullying and relational bullying were analyzed separately. Such approach lacks justification. The separate (or univariate) analyses increase the type I error because physical bullying and relational bullying were highly correlated with r = .71 (p < .05), the highest among all the correlations between the study variables (see Table 2). I would recommend using a multivariate approach analyzing all variables simultaneously in one structural equation model.

2) Page 4, Line 6: It is not clear whether a mean or total score was calculated as “a composite score.”

3) Page 4, Line 7: it is not clear whether alpha = .77 is the reliability for your study sample or Wyman et al.’s (1993). Also, you need to report reliability for both your sample and ones published in the literature for all of your key variables.

4) Page 7, Line 19: Here you mentioned that Cohen, Cohen, West, and Aiken’s (2003) method was used to deal with missing data for the variable of mother’s education. But, the whole Cohen, Cohen, West, and Aiken’s (2003) book is about regression/correlation analysis; and it is not clear what method was used and how the missing data were treated. Also, a description of whether or not all other variables had missing data and how the missing data were dealt with is needed. Furthermore, Cohen, Cohen, West, and Aiken’s (2003) is not in the reference list. Please double check the reference citations in the text and make sure all reference citations are listed in the reference list.  The Positive Expectation of What the Future Holds Analysis

5) Page 10, Line 7: The section heading is misleading because the analyses reported in this section are not multivariate, but univariate. See the first comment above.

 

 

Reviewer #3: Overall this is a well-written paper that appears to offer new information to our understanding of bullying, especially as the bulk of the research has focused on the recipients and there has been little research done on perpetuators. There are, however, some major limitations in the study’s design and execution.

A major concern that this reviewer has is the failure to include cyberbullying in their study, especially as it now viewed as (arguably) the most common bullying leading to the most pernicious outcomes. The reason for this exclusion needs to be addressed.
A second concern is the failure to discuss the role socio-economic status plays in future expectations, although there is a significant body of literature that has investigated these relationships. This is important from a theoretical perspective but also considering the sample and demographic variables described.

A third concern is there is conceptual confusion (or at least word selection) with selected terms used throughout the manuscript. Direct bullying goes beyond physical altercations. The terms physical bullying and direct bullying are used interchangeably throughout the paper. Tables 1 & 2 uses the terms relational and physical ‘aggression’ where the text uses bullying. There is of concern as aggression does not encompass all the same components (i.e., power differential) as does bullying. Table 3 and Figure 1 use the terms physical and relational bullying. There needs to be consistency throughout.
Specific comments by section are listed below.

Introduction
* Page 1, line 22—page 2, line 2: Would be specific about the audience when discussing the short- and long-term impacts of bullying. Many of the factors listed affect both the recipient and instigator.
* Page 2, lines 2-12: Should also address age/developmental level as there is interplay between age and gender with types of bullying and their instigators and recipients and age appears in your demographics. Ethnicity was not referenced…was there a reason?

Methods
* More description on the development, preliminary psychometrics, and utility of the total survey is needed as this information is foundational to the rest of the study. The researcher needs to be explicit that there were 5 separate measures included in a single survey and why there is no single measure specific to bullying. It should be specified that the alphas reported for each measures are Cronbach’s alphas. It appears from the references that some of these questions used in selected measures were taken from previous work that is fairly old. As there has been significant work done on bullying measurement over the past decade, further explication is needed.
* A reference as to why mother’s educational attainment alone was considered a good proxy measure of SES is needed; usually a minimum of two factors (ed and employment) is used.

Results
* It is unclear why bullying isn’t initially handled as a single variable as the overall concept being investigated is bullying perpetration (see overall study purpose). This goes back to the intro where the concepts are introduced and where physical and relational are considered bullying subtypes.. The relationships with the overall concept should be explored prior to investigating relationships with among the sub-constructs.
* Under the Multivariate Model heading: Were all of the demographic variables not significant? How was that determined? This information is not in evidence. I question the validity of including gender if this was non-significant in the univariate model (assuming this was first done).

Discussion
* The discussion leads off by describing the community as ‘economically distressed’. There was no information provided earlier that that was the case. In fact, the authors state that the sample was highly diverse.
* The discussion focuses on bullying as a single construct yet this does not match the analysis.
* The authors introduce social skill training in this section although this is not discussed earlier. There is literature to support the lack of social skills is an important component of bullying, particularly in bully-victims, but this discussion is outside of the information presented.
* Another limitation that should be added is that of ‘self-report’. While this is always an issue, it is more so with bullies where self-aggrandizement is common.

References
* The following references were not noted in the reference list:
o Bentler, 1990
o Brown & Cudeck, 1993
o Cohen, Cohen, West & Aiken; 2003
o Kline, 2011
Other
* The word ‘youth’ not ‘youth’s is used for the plural form throughout this paper; although this is coming into more common usage.
* There are some relatively minor departures in APA style throughout.

—

ADDITIONAL COMMENTS FROM THE EDITOR
TITLE
Please revise to include all variables in the mediation model, and to identify the approach as a mediation evaluation. Clarify that bullying perpetration (not victimization by bullying) is the outcome of interest. Then in the paper, it would add clarity to always use the term “bullying perpetration” (physical or relational, as appropriate for context).

RELEVANCE TO NURSING SCIENCE
Please address relevance to nursing science in the background and discussion.

MEDIATION HYPOTHESES
The review of literature defines variables in the mediation system (future expectations, attitude towards violence, bullying perpetration) and cites literature finding associations among the variables. The argument for the mediation hypothesis is weak theoretically and empirically (“Yet, few, if any, researchers have examined the potential mediating role of attitude towards violence to solve problems on the link between adolescents’ future expectations and bullying behavior.”) Essentially, this thought needs to be extended to explain why it makes sense to examine the mediation model. This argument is essential because mediation analysis is designed to “extract information about the causal mechanism(s) by which a predictor affects an outcome” (Preacher, 2015, p. 826; and many others). The strong claim made in any mediation hypothesis deserves a strong foundation. This is especially true when a cross-sectional design is used to suggest an essential feature of a fundamentally time-based process. With that in mind, it would be helpful to explain (on the basis of logic, theory, or existing data) at least how the temporal precedence of the variables can be justified (future expectations precedes attitude towards violence precedes bullying perpetration). Another important issue is to consider alternative explanations, or at least acknowledge that they exist and are important lines of investigation in future studies.

MEDIATION ANALYSIS
Overall, the analysis was well done and clear.

It wasn’t obvious why two separate models were estimated (for bullying perpetration–physical and bullying perpetration–relational). Within the SEM framework, it should be relatively easy to set one model with both outcomes, and incorporate the relationship between the outcomes in the model as well.

Can you explain why gender was used as a covariate, rather than setting up the model as a multiple group analysis? Use of equality constraints could then highlight similarities and differences in the mediation model for girls and boys (essentially, allowing a moderated mediation hypothesis to be examined).

Address the statistical assumptions; especially that errors in prediction for M and Ys are uncorrelated (again see Preacher (2015) (who noted that this issue was not mentioned in Baron & Kenny, 1986).

QUESTIONS ABOUT YOUR REVISION
Please contact me using email via Editorial Manager (“send letter to editorial office”).

References
Preacher, K. J. (2015). Advances in mediation analysis: A survey and synthesis of new developments. Annual Review of Psychology, 66, 825-852. doi: 10.1146/annurev-psych-010814-015258

———————————————————————-
CHECKLIST FOR STYLE

TITLE PAGE —
–Supply credentials and/or highest earned degree for Varela.
–Supply running head of less than 50 characters (no abbreviations).

 

TEXT —
–When your revision has been completed, run a final spell check. Also, proof to be sure that spelling errors not picked up by the spell check have been corrected.

 

REFERENCES —
–When your revision has been completed, verify consistency between references cited in the text and references included on the reference list.

Update REFERENCE LIST using APA 6th Ed. format. In particular:
–The following reference is not cited in the text. Remove from reference list or add to text: Vernberg et al., 1999.

Update IN-TEXT CITATIONS using APA 6th Ed. Format. In particular:
–On first citation of a reference, list all authors if there are less than 6. For subsequent citations, use first author followed by et al. (i.e., Crapanzano 2011).

 

–Add IN-TEXT citations of the following references to the reference list:
— Cohen, Cohen, West & Aiken, 2003
–Brown & Cudeck, 1993
–Kline, 2011

–Please do not introduce sources/citations in the Discussion (e.g., Borntrager et al., 2009). They should be included in the Background initially.

 

TABLES —
–Bolding is not necessary, please remove.