Abstract

We sought to investigate the relationships between occupational category, workaholism, and workplace aggression, using a sample of 249 faculty and staff members employed full-time at a large Southeastern university. The focus shifted to exploratory analyses using self-reported categorizations after receiving a low response rate from blue-collar workers, as defined by the current study. Interestingly, self-categorizations were much closer to population estimates. Self-reported occupational category was not significantly related to workaholism nor workplace aggression, however, the self-reports were related to typical factors used to differentiate occupational categories (e.g., percentage of manual labor and level of education). Moreover, workaholism was positively related to and significantly predicted workplace aggression. Findings emphasize the importance of organizational policy targeted at reducing workaholic tendencies, as these policies may also lower workplace aggression incidents. The results also point to a discrepancy between common determinants of true occupational category and one’s self-categorization, a finding that may be of interest to future occupational category research.

1. Introduction

The U.S. Office of Personnel Management ¹ conceptualizes blue- and white-collar jobs into the term occupational categories. These occupational categories have several key defining features—blue-collar work involves activities such as trades (e.g., electrician) and manual labor, while white-collar work has an emphasis on office work, specialized higher education, and the use of analytical or judgement skills. Other researchers have used similar conceptualizations (e.g., ^{2, 3}), thus, occupational category is likely an umbrella term encompassing different situations and job characteristics.

Occupational categories can influence workers and their well-being. Early research suggests white-collar workers value intrinsic job aspects, whereas blue-collar workers were focused on external factors ^{4, 5}. Anjum and Parvez ² found white-collar workers were more satisfied, earned more, were more autonomous, had greater task significance, and exhibited higher self-esteem. In comparison, Wright and colleagues ⁶ reported blue-collar workers had more physical work environment complaints, more health symptoms (mental and physical), less stimulation, less work enjoyment, less autonomy, more monotonous work, feelings of less fit with their job, and lower life satisfaction.

Other work environment factors, such as work stress, can have varying effects on employee well-being. Work stress may lead to counterproductive work behaviors (CWBs; ⁷), which consist of unethical behaviors by an employee intended to harm the organization for which they work ⁸. Anjum and Parvez ² found that, compared to white-collar workers, blue-collar workers are more counterproductive, with the most frequent CWB being abuse. Poor physical working conditions are also related to increased bullying risk ⁹ and are prevalent in blue-collar positions ⁶.

Workplace aggression research has consisted of many terms and conceptualizations involving interpersonal deviance, harassment, CWBs, and more ^{10, 11}. Barling and colleagues ¹⁰ expanded on existing definitions, defining workplace aggression as behaviors enacted by employees intending to harm individuals within the organization or even the organization itself, in which the target is motivated to avoid the aggressive behavior. Examining common CWBs (e.g., abuse) reveals they fit typical definitions of workplace aggression, thus, CWBs will be conceptualized as a form of workplace aggression in the current study. Environmental conditions have been an area of focus for workplace aggression research ^{3, 9}. For instance, role stressors and interpersonal conflicts appear to be triggers of aggressive work behavior ¹². Finstad and colleagues ³ found a lack of autonomy and supportive environment, as well as heavy job demands, could lead to workplace aggression.

2. Current Study

The main objective of the current study was to evaluate the potential moderating influence of workaholism on the occupational category—workplace aggression relationship. Consensus in the literature denotes workaholism is the compulsive need to work excessively hard ¹³. Workaholism positively relates to both work stress and workplace aggression ^{11, 13}. Much research has been conducted on how environmental factors at work, such as workplace aggression (e.g., ^{3, 9}), affect different areas of one’s life, but little has investigated the moderating role of personality factors such as workaholism. Balducci and colleagues ¹¹, after finding a small association between workaholism and workplace aggression, proposed there was potential for workaholism to moderate the relationship between situational factors and workplace aggression.

We looked to add to existing research on the work environment—workplace aggression relationship by focusing on a single organization (i.e., a large Southeastern university), thereby hoping to reduce variation in organizational culture. While there is existing literature showing blue-collar workers exhibit more CWBs ², we used a broader measure of workplace aggression in the current study, instead of focusing on subfactors (e.g., physical, verbal). We also investigated the potential moderating influence of workaholism on the occupational category—workplace aggression relationship, following ideas from prior research ¹¹.

If workaholism moderates the relationship between occupational category and workplace aggression, this will provide more insight on the correlates of workplace aggression. By further understanding these relationships, organizations can better focus training and intervention development efforts to limit this risk. Ultimately, we hope that any interventions developed to reduce aggression risk are tailored to the occupational category at a higher peril for workplace aggression. Accordingly, we strive to contribute to extant literature.

3. Hypotheses

Much research in the domain of workplace aggression has focused on the work environment ³. Specifically, of these work environment factors, work stress and poor physical conditions both relate to higher risk of workplace aggression ^{9, 14} and are commonly portrayed in blue-collar workers ⁶. These poor work environment factors encountered by blue-collar workers ultimately relate to higher levels of workplace aggression incidents. In addition, compared to white-collar work, blue-collar work is more related to CWBs, a form of workplace aggression ². The conservation of resources (COR; ¹⁵) theory can be used to explain this relationship. Based on the COR theory, individuals seek to protect and expand upon their resources; the absence or removal of such resources relates to distress ¹⁵. This loss of resources could manifest in the form of workplace aggression for blue-collar workers, as they may already be lacking in resources and might be more vulnerable to loss spirals in order to maintain the resources they have left. Accordingly, we proposed the following hypothesis:

Hypothesis 1 (H1): Occupational category will be related to workplace aggression such that workplace aggression will be greater in blue-collar workers than in white-collar workers.

Workaholism has commonly been related to an internal compulsion to work ¹⁶. Occupational category research has examined the role of motivators, whereby blue-collar workers value external factors (e.g., rewards) more than white-collar workers ^{4, 5}. Opportunity also plays a role in excessive working found in workaholics. White-collar positions often involve technology and computer usage, which can be provided by an employee’s company ¹⁷. Given their internal motivation, similar to descriptions of an internal compulsion for workaholism ¹³, and the access to technology throughout the day (even after hours), white-collar workers are provided with more reason and opportunity to work excessively hard. According to the COR theory, acquisition of resources enhances an employee's ability to work ¹⁵. However, depletion or loss of these resources may increase the prevalence of stress and the desire to recover lost resources (and gain new ones) through extensive work and workaholic tendencies ¹⁸. This might be especially true of white-collar workers, as their jobs are characterized by task identity, task significance, and autonomy ². Therefore, the following hypothesis is presented:

Hypothesis 2 (H2): Workaholism will be related to occupational category such that workaholism will be greater in white-collar workers than in blue-collar workers.

Many outcomes of workaholism and correlates of workplace aggression seem to overlap. For example, higher levels of workaholism have been linked to CWBs, a form of workplace aggression, along with work stress ^{13, 19}. Furthermore, work stress is associated with increased CWBs and other forms of aggression ^{7, 14, 12}. Shimazu and colleagues ²⁰ found workaholism to be positively related to emotional discharge, defined by Carver et al. ²¹ as openly venting one’s negative feelings to others. Balducci and colleagues ¹¹ theorized emotional discharge could build into aggressive behavior in workaholics. Additionally, as described earlier, in accordance with the COR theory, individuals are focused on protecting and obtaining mental and physical resources ²². Workaholism involves excessive work and constant work-related thoughts ¹³, which drains resources. In addition, workaholism may not lead to improved performance ¹³; in line with COR theory, increased performance refers to resource gain. According to the fourth principle of COR theory, individuals enter a defensive mode when their resources are stretched or depleted, thereby leading to aggressive or irrational responses to cope with the resource depletion ²². Hence, we proposed the following hypothesis:

Hypothesis 3 (H3): Workaholism will be positively related to workplace aggression.

In accordance with COR theory, the potential increased time spent working outside of the office for white-collar workers, as well as the decreased autonomy and poor working conditions faced by blue-collar workers, could be seen as a loss of resources. Of note, in the current study, time spent working outside of the office refers to working in evening and/or weekend. The addition of workaholism, a construct related to work stress ¹⁴, could be perceived across categories as a further depletion of resources (e.g., time and effort) as individuals have a compulsion that drives them to incessantly work. With the exhaustion of resources, principle four of COR theory (i.e., lack of resources will lead to defensive attempts to conserve the remaining resources) would suggest the possibility for aggressive and irrational behavior ²². Thus, the following hypothesis is presented:

Hypothesis 4 (H4): Workaholism will moderate the relationship between occupational category and workplace aggression, such that the higher the levels of workaholism, the stronger the relationship between occupational category and workplace aggression.

4. Methods

Participants

After data were cleaned (e.g., cases were removed for: failing validity check items, not completing study measures, and not meeting the study requirements of full-time employees at least 18 years or older), the final sample consisted of 249 full-time faculty and staff members employed at a large Southeastern university. Demographic information is presented in Table 1.

Procedure

Participants were recruited via email to a random selection of faculty and staff and asked to complete a survey using Qualtrics. They were provided informed consent in which they were assured of the confidentiality and anonymity of the survey, and verified their participation was voluntary. Participants then responded to demographic questions, items about their current position, a workaholism measure, and a workplace aggression measure.

Measures

Occupational Category. The U.S. Office of Personnel Management’s ¹ classifications of blue- and white-collar were utilized. Respondents were also asked to self-categorize as blue-collar or white-collar based on an item on the survey that queried how they would categorize their current position. Additional items were added to the demographics section to further understand each respondent’s position. These items included: “Which description applies to your current position more? I work with my hands and tools OR I work with information;” and “Which skills are more often required by your position? Trade and/or Manual Labor OR Analytical and/or Judgmental?”

Workaholism. The 29-item Workaholism Analysis Questionnaire (WAQ; ²³) was used to measure workaholism. A McDonald’s omega of .924 and Cronbach’s alpha of .921 were obtained in the current study.

Workplace Aggression. The 29-item Aggression Questionnaire ²⁴ was used to measure workplace aggression. As this measure was not initially developed for workplace aggression research, minor changes were made to the instructions and the wording of some items to ensure respondents answered in the context of their work. A McDonald’s omega of .869 and Cronbach’s alpha of .861 were obtained.

5. Results

Transformations were conducted on four variables to account for deviations from normality, including: a rank transformation on the hours per week variable, a log transformation on the position tenure and manual labor percentage variables, and a square root transformation on the remote hours per week variable. Skewness and kurtosis for each of these variables are in Table 2. All analyses containing these variables were completed once with the transformed variables and once with the original variables. Although we present both the transformed and non-transformed models below, the results were, notably, the same when tested both ways.

Table 3 includes the McDonald’s omegas, Pearson correlations, and descriptive statistics for the study, demographic, and working condition variables. Workaholism was positively related to workplace aggression, thereby supporting H3. Due to a lack of blue-collar participants (N = 14 , 5.6%) based on the U.S. Office of Personnel Management’s classifications of blue- and white-collar, exploratory analyses were undertaken to investigate relationships with self-categorizations of occupational category. While the percentage of individuals self-identifying as blue-collar was still low (N = 34, 13.7%), it was much closer to more recent population estimates (13.9%; ²⁵) and hence used for the analyses.

For self-categorization, neither the correlations nor the t-tests were significant for workaholism or aggression, thus refuting H1 and H2. That said, though not significant, the mean scores for both workaholism and aggression were slightly higher for those who self-categorized as blue-collar and the correlations were positive (white-collar = 0, blue-collar =1). Furthermore, self-categorization of occupational category was significantly related to level of education. Additionally, self-categorized white-collar individuals worked significantly more remote hours per week (M = 8.84, SD = 5.91, n = 207) than self-categorized blue-collar individuals (M = 1.30, SD = 3.59, n = 34), t(52.66) = 5.012, p < .001, d = 0.84, 95% CI [0.44, 1.24]. This difference was also significant when using the non-transformed remote hours variable, t(63.57) = 4.971, p < .001, d = 0.77, 95% CI [0.38, 1.15]. Self-categorized blue-collar workers reported a significantly higher percentage of manual labor (M = 16.62, SD = 4.50, n = 34) than did self-categorized white-collar workers (M = 2.50, SD = 3.35, n = 207), t(40.30) = -6.986, p < .001, d = 1.39, 95% CI [0.91, 1.86]. This difference was also significant when using the non-transformed manual labor variable, t(34.19) = -5.399, p < .001, d = 1.26, 95% CI [0.78, 1.72].

Two multiple regression analyses were run for predicting workplace aggression, one using transformed variables and the other using non-transformed variables. For exploratory purposes, the models were fit predicting workplace aggression from workaholism, hours per week, self-categorization of occupational category, whether work hours were fixed or variable, remote hours per week, age, gender, education, and manual labor percentage. Both the transformed model (R² = .26, F(9, 224) = 8.670, p < .001, 90% CI [.15, .32]) and the non-transformed model (R² = .26, F(9, 224) = 8.896, p < .001, 90% CI [16,32] ^16, ) significantly predicted workplace aggression, whereby workaholism and hours worked were the only significant predictors; this provides further support for H3. Slopes, standard errors, p-values, and semi-partial r for the transformed model are in Table 4 (Model 1) and for the non-transformed model in Table 5 (Model 3). We also reran the model predicting workplace aggression (see Model 5) and the interaction (Workaholism X Self-Category) model (see Model 6) with only the variables of interest (i.e., workaholism and self-categorizations); nothing changed in terms of significance (see Table 6), thereby refuting H4. However, workaholism was the only significant predictor of workplace aggression in both models, which again supports H3.

Two additional models were fit including the interaction term for workaholism and self-categorization (Model 2 using transformed data and Model 4 using non-transformed data). Slopes, standard errors, p-values, semi-partial r, and change in R² for the transformed model, including interaction term, are in Table 4, while results for the non-transformed model are in Table 5. The interaction term was not significant, thus, we again have no evidence of moderation and H4 was not supported; only the number of weekly hours was a significant predictor.

6. Discussion

In this study, we sought to investigate the occupational category—workplace aggression relationship, as well as the potential moderating influence of workaholism. The number of blue-collar participants, as defined by the U.S. Office of Personnel Management’s classifications of blue- and white-collar, was much lower than the national average ²⁵. Thus, self-categorizations of occupational category were utilized accordingly. Nonetheless, the relationship between workaholism and workplace aggression could still be investigated. Notably, we found a medium positive workaholism—workplace aggression relationship. Moreover, workaholism significantly predicted aggression in all the models we tested, with a small to moderate amount of variance found to be attributed to workaholism. These results support and expand previous findings by Balducci et al. ¹¹.

Exploratory analyses revealed self-categorization of occupational category was not significantly related to either workplace aggression or workaholism. Self-categorization was, however, significantly related to typical components used to differentiate occupational categories, such as percentage of manual labor and level of education. These results suggest individuals may be aware of and have some insight on generally acknowledged occupational category differences. This result is somewhat surprising given its discrepancy with what was found when investigating participants’ job roles.

Multiple regression models were fit predicting workplace aggression from workaholism and other relevant work and demographic variables. Surprisingly, hours worked had a negative slope despite its positive zero-order correlation with workplace aggression. This result may indicate hours worked is suppressing some variance in workaholism not related to workplace aggression, as its slope exceeded its zero-order correlation. Additional regression models were fit containing the interaction term of workaholism and self-categorized occupational category in order to investigate a potential moderating influence. Analyses revealed the interaction was not significant, nor did the inclusion of the interaction term result in a substantial R² change. This result provided no significant evidence to further probe a moderating effect.

Study Limitations

The current study is not without limitations. The original intent of the study could not be fully examined due to a lack of blue-collar participants, despite attempts to oversample from potentially high blue-collar departments. One potential reason for the low response rate by blue-collar workers, in spite of oversampling, may be a lack of flexibility in work tasks. For example, blue-collars workers may not have the same amount of time as white-collar workers to complete computer-based surveys at work. In the exploratory analyses that followed, self-categorization of occupational category was investigated, however, the discrepancies between self-categorization and actual categorization compared to the study definitions raises questions about whether the initial research goal was adequately achieved.

Another study limitation involves the sample being drawn solely from workers at a large Southeastern university. Several factors, such as racial and gender distribution, as well as education levels, might likely differ from population totals. Hence, caution should be taken when generalizing results outside of a large Southeastern university. Future researchers should take this into account and make additional attempts to achieve a more representative sample.

Organizational Implications and Future Research Directions

Our research findings highlight that workaholism is related to increased workplace aggression, thereby providing an additional reason for organizations to discourage workaholic tendencies and enhance work-life balance initiatives. Furthermore, in the current study, remote work was found to relate to workaholism. While technology can make out-of-office work easier, workaholism still relates to negative outcomes including worse health and work stress, and the increase in work does not necessarily equate to better performance ^{13, 19, 26, 27}. Additionally, the increased aggressive behavior caused by workaholism is related to negative organizational outcomes such as theft ^{11, 14}. Accordingly, our findings highlight the importance of investing in efforts aimed at reducing workaholism and, consequently, workplace aggression.

While workaholism did not have a moderating influence between occupational category and workplace aggression, future researchers should continue to investigate potential moderating factors. The challenge of sample size will continue to exist in occupational category research, however, this should not discourage the investigation of occupational category differences. Future researchers should attempt to sample from large organizations that may provide better representative samples of both blue-and-white-collar occupational categories (e.g., internet and cable industries which have a hybrid of physical on-site and remote computer-based work).

The current study also revealed there may be a difference between how individuals perceive their occupational category and their true occupational category. This difference may be of importance to future research on the subject of occupational categories. Does one’s perception of their occupational category interact with their true occupational category to influence their behavior and well-being? Future research on occupational category differences should consider investigating this question further.

Furthermore, although Buss & Perry’s ²⁴ measure contains four types of workplace aggression (i.e., physical, verbal, anger, and hostility), the average of the total items was measured in the current study. Perhaps there are differences in the relationships we observed (or did not observe) based on the type of aggression. For instance, feasibly most participants did not endorse physical aggression items, whereas verbal aggression, anger, or hostility items may have been more commonly endorsed. Theoretically, there could also be a case made for a stronger relationship between workaholism and anger, in particular. We provide this suggestion for a potential future research direction. Also, in addition to the literature on deviance and aggression, there is also research on positive deviance (e.g., ²⁸). It would be interesting to examine if workaholics are more or less likely to engage in positive deviance compared to their peers. For example, they may use certain forms of constructive deviance (e.g., extra-role behaviors, taking charge, etc.) as a way to engage in their workaholism.

ACKNOWLEDGEMENTS

We gratefully acknowledge Macie Yao for reviewing earlier drafts of this paper.

Conflicts of Interest

On behalf of all authors I, the corresponding author, declare no conflicts of interest.

References