Examining Age-Dependent Patterns in Academic Bullying Behaviors =============================================================== * Sherry E. Moss * Morteza Mahmoudi ## Abstract Academic bullying is a pervasive and longstanding issue that negatively affects all fields of science. While influential factors such as scientific discipline, sex, and ethnicity have been investigated, the impact of perpetrators’ age on the contextual behaviors of academic bullying has not been thoroughly studied. Our cross-sectional global survey of academic bullying, involving 2,390 participants, provides evidence that age significantly influences the contextual behaviors of academic bullying. Our results revealed that the youngest perpetrators (aged 25-35) exhibit significantly less abusive behavior compared to other age categories (36 and older). Additionally, the lowest percentage of abusive behavior was observed in the youngest group, while the highest was found in the older groups (aged 56-65). Our comprehensive analysis on the potential influence of the perpetrator’s academic position, ethnicity, or sex on their age group revealed no significant impact. These findings have strong implications for the development of more customized academic bullying training, policy development, and monitoring strategies, ultimately benefiting the scientific community at large in many aspects including evolution of science and improving the mental health of the academic workforce. ## Introduction Academic bullying has garnered significant interest in recent years, with scientific communities across various disciplines highlighting the importance of addressing this pervasive and longstanding issue.1-3 Researchers have explored its negative effects not only on the immediate targets but also on the wider community and the progression of scientific research.4 Various influential factors, such as gender and ethnicity of both targets and perpetrators, have been investigated to better understand the root causes of academic bullying and identify vulnerable populations and likely perpetrators.5,6 However, the impact of perpetrators’ age on the contextual behaviors of academic bullying has not been thoroughly examined or systematically studied. Age is a significant factor influencing a wide range of prosocial and socioemotional behaviors and actions.7 Emerging evidence suggests that older adults tend to be more prosocial than younger adults, as demonstrated by their increased propensity to learn about rewards for others and engage in effortful actions for the benefit of others.8,9 Theoretical accounts of lifespan development, such as socioemotional selectivity theory, propose that this increase in prosocial behavior is driven by age-related shifts in goals and priorities.9,10 As individuals age, their motivation for socially and emotionally meaningful behaviors grows, driven by a focus on immediate emotional satisfaction and meaningful social interactions. This theory suggests that older adults prioritize positive emotional experiences and meaningful connections, which in turn enhances their prosocial behaviors. Despite extensive research on various factors contributing to academic bullying (including our own results5,11), the specific influence of perpetrators’ age on their bullying behaviors has not been thoroughly investigated.6,12-16 Understanding whether aging has similar socioemotional effects on academic bullying and harassment as it does on general prosocial behavior is crucial. This paper aims to investigate the hypothesis that aging has comparable socioemotional effects on academic bullying and harassment, potentially influencing the nature and extent of bullying behaviors exhibited by older versus younger perpetrators. Exploring the relationship between age and academic bullying behaviors can provide valuable insights into the dynamics of bullying in academic settings. It can help identify whether older individuals are more likely to engage in or refrain from bullying behaviors due to their socioemotional development. Furthermore, this research can inform the development of age-specific interventions and policies aimed at mitigating academic bullying, thereby fostering a healthier and more inclusive academic environment. ## Results and discussion To investigate the potential role of perpetrators’ age in academic bullying, we analyzed data from our cross-sectional global survey on the topic. Full details about IRB approval [Wake Forest University (IRB00023594) and Michigan State University (STUDY00003215)], consent, and the declaration of informed consent for data usage are provided in the survey documentation (see our earlier publication5 for details). Data were collected from 2,390 individuals, solicited through various channels including advertorial pieces and third-party emails in Science17 and Nature magazines, as well as online panel advertisements and third-party emails from the American Chemical Society. Because targets are unlikely to know the exact chronological age of their perpetrators, we asked them to estimate their perpetrators’ age group using the following 5 categories: 25-35, 36-45, 46-55, 56-65, and > 65. We first removed cases that had missing data for the question “What is the approximate age of the perpetrator?” in order to run the one-way ANOVA, Kruskal-Wallis, and Chi-Square Test of Independence analyses. **Table 1** shows the descriptive statistics and analyses for this reduced dataset (n=1,412). View this table: [Table 1.](http://medrxiv.org/content/early/2024/09/25/2024.09.24.24314297/T1) Table 1. An overview of the dataset for the variables of interest. Age of perpetrator and academic bullying behaviors, defined as items on the Tepper Scale and Checklist of Bullying Behaviors (CBB). The survey questions details are available in **the Supporting Information**. (“P” stands for “Percentile”) We used the 15-item abusive supervision scale developed by Tepper18 to assess general bullying behaviors. Sample items included statements like “my supervisor ridicules me” and “my supervisor puts me down in front of others.” Participants who perceived themselves as having been bullied responded to these items using a 5-point Likert scale ranging from 1 (“I cannot remember him/her ever using this behavior with me”) to 5 (“He/she uses this behavior very often with me”). In addition to the Tepper scale, we utilized our contextual bullying behavior (CBB) checklist, which consists of 10 items, and was developed through the analysis of publicized stories and narratives collected over time by the authors.5 These behaviors include abusing authorship or violating intellectual property rights19; threatening to cancel funding, positions, or visas6; and damaging the reputations of budding scientists through negative recommendations or speaking poorly about them to others12. It is noteworthy that we linked the study of what is colloquially known as “bullying” in academic science with the established science of abusive supervision from the organizational literature.5 To determine whether the age of perpetrators affects their level of abusive supervision, we investigated the relationship between age categories from the survey and scores on the Tepper18 scale, which measures abusive supervision, together with the influence of age on the contextual behaviors associated with abusive actions. We ran analyses looking at the entire Tepper scale (15 items). The Tepper scale score distribution across the age groups was similar enough to use a one-way ANOVA analysis to compare the total Tepper scale score by perpetrator age. 155 responses were removed from the analysis due to missing data. We did not find a significant difference among any of the perpetrator age groups and total Tepper scale score (**Figure 1a-d**). Full data analysis of each individual behavior with perpetrator age group is available on the **SI**. ![Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/09/25/2024.09.24.24314297/F1.medium.gif) [Figure 1.](http://medrxiv.org/content/early/2024/09/25/2024.09.24.24314297/F1) Figure 1. Tepper scale analysis of the age of the perpetrators. Frequency (a) and counts (b) of the respondents of the total scores on the Tepper scale; (c) box plots showing the distribution of the Tepper scale scores by the reported perpetrator ages; (d) one-way ANOVA results comparing perpetrator age group to Tepper Scale score. ![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/09/25/2024.09.24.24314297/F2.medium.gif) [Figure 2.](http://medrxiv.org/content/early/2024/09/25/2024.09.24.24314297/F2) Figure 2. Contextual bullying behavior (CBB) checklist analysis of the age of the perpetrators. Frequency (a) and counts (b) of the respondents of the total scores on the CBB; (c) box plots showing the distribution of the number of abusive behaviors by the reported perpetrator ages; and (d) pairwise comparisons using Wilcoxon rank sun test with continuity correction. Age group “25-35” is significantly different than all other age groups. We then investigated whether the perpetrator’s academic position, ethnicity, and sex influenced the observed age-related results. Our analysis revealed no significant interactions between the perpetrator’s academic position, ethnicity, or sex with their age group. Specifically, the professional positions of the perpetrators—Principal Investigator (n=722), Lab Supervisor (n=75), Group Leader (n=155), and Department Chair (n=239)—showed no significant correlation with the perpetrator’s age groups. Furthermore, there was no additive effect between the perpetrator’s age group and their professional role, as indicated by the non-significant results from the unbalanced two-way ANOVA. Similarly, no significant interactions were found between the perpetrator’s age group and their race/ethnicity—White (n=860), Asian (n=133), Middle Eastern (n=65), and East Indian (n=40). However, an additive relationship between the perpetrator’s age group and their race/ethnicity was observed, with the unbalanced two-way ANOVA indicating a significant result (p<0.01). Tukey’s HSD test identified one significant relationship between the “White” and “Asian” groups. Nonetheless, post-hoc analyses revealed that the ANOVA normality assumption was not met (p<0.001), as the Shapiro-Wilk normality test result was below the threshold of p<0.05. Additionally, there were no interactions between the perpetrator’s age group and sex. An additive relationship was found between the perpetrator’s age group and sex—male (n=751) and female (n=420)—with the unbalanced two-way ANOVA showing a significant result (p<0.01). However, post-hoc analyses revealed that ANOVA normality assumption was not met (p<0.001). To be met, the Shapiro-Wilk normality test must not be p<0.05. Overall, these analyses suggest that the perpetrator’s age group results were not significantly influenced by their academic position, ethnicity, or sex. Next, we ran analyses looking at the summed checklist of binary yes/no academic bullying behaviors (10 items) as well as the individual items in relation to the reported age of the perpetrator (**Table 2**). The distribution of the CBB scores across the age groups were not similar enough to use a one-way ANOVA analysis. Because these were skewed, we used a Kruskal-Wallis test to compare the total CBB scale score by perpetrator age. 150 responses were removed from the analysis due to missing data. We found a significant difference among the lowest perpetrator age groups (25-35) and total CBB score when compared to all other perpetrator age groups (p<0.01). View this table: [Table 2.](http://medrxiv.org/content/early/2024/09/25/2024.09.24.24314297/T2) Table 2. Percentage of perpetrators engaging in contextual bullying behaviors across age groups We finally looked at the individual items on the CBB in relation to the perpetrator’s age group. We ran chi-square tests of independence to analyze these items, as both the predictor variable (perpetrator age) and outcome variable (CBB item) were categorical variables (**Table 3** and **Figure 3;** full details of the statistical analysis are available in the **SI**). We found five items with a significant relationship between the bullying behavior and perpetrator age group (p<0.05) and 3 items with less significant differences (0.05