This study examines the impact of abusive supervision on withdrawal behavior among bank employees at the staff level in Sri Lanka's Northern Province. The study focuses on physical and psychological withdrawal behavior to determine how supervisory behavior influences worker disengagement.
Primary data were collected by a guided questionnaire from 131 employees of licensed commercial banks operating in the Northern Province. The research uses linear regression analysis through the SPSS (Version 29) to explore the linkage between withdrawal behaviour and abusive supervision. Descriptive statistics and correlation matrices were also used to examine associations with demographic variables.
This study investigates the impact of abusive supervision on physical and psychological withdrawal behaviors among bank employees in Sri Lanka's Northern Province, a context rarely explored in organizational literature. Using linear regression and SEM, abusive supervision was found to significantly predict withdrawal behaviors (β = 0.649 for physical, β = 0.594 for psychological; p < 0.001). Grounded in Conservation of Resources theory, the findings highlight emotional resource depletion as a mechanism driving disengagement, uniquely framed in a post-conflict, culturally distinct banking sector.
This study uniquely investigates the dual impact of abusive supervision on physical and psychological withdrawal behaviors among bank staff in Sri Lanka's post-conflict Northern Province, an underexplored context in organizational literature. Using Conservation of Resources (COR) theory, it frames withdrawal as a response to emotional resource loss. By incorporating context-specific factors like sector and marital status, the study adds meaningful, theory-driven insights to the global conversation on workplace abuse.
1. Introduction
Workplace dynamics significantly influence employees' attitudes, behaviour, and job satisfaction. The employee-supervisor relationship is crucial in this social environment. Effective leadership can inspire and drive productivity, while poor leadership can have detrimental effects. Employees who show withdrawal behaviour should be studied to retain them. The prevalence of abusive supervision has increased globally in recent decades and may contribute to withdrawal behaviour. In the US, nearly one-third of employees have experienced abusive supervision (Tepper, 2000).
Employee abuse has been a considerable problem over the years in Sri Lanka. Employees are assets of the firms. Contributions made by them to the organization are never taken for granted. However, many organizations are seeing more and more withdrawal behaviour from their employees, and this seems to be affecting the performance and sustainability of the firms. Employees have withdrawn from jobs for many reasons over the years. It has increased in the last few years in Sri Lanka, especially in the banking sector [1].
Abusive supervision and employee withdrawal behaviour underpin this study, especially in the Sri Lankan context. Many previous studies have focused on abusive supervision and withdrawal behaviour (Chi and Liang, 2013; Huang et al., 2020; Tepper et al., 2017; Whitman et al., 2014). However, no previous studies have examined the impact of abusive supervision on withdrawal behaviour in the context of Sri Lanka, particularly in the banking sector and the Northern Province.
The problem tackled in this research is that withdrawal behaviour is largely expressed by employees in any field. Employee withdrawal behaviour has been attributed to several issues, including organizational climate and structure. Employees react strongly to management's attitude toward them. The obvious reaction to a bad attitude is disengagement, which harms organizational performance.
The universality of abusive supervision underscores its significance as a global workplace issue. Its consequences extend beyond individual suffering to impact organizations at large. High turnover, decreased productivity, and increased absenteeism pose serious challenges for organizations. Employees subjected to such treatment often experience high stress (Tepper, 2007), decline in job satisfaction (Tepper, 2000), decreased organizational commitment (Liu et al., 2012), and impaired job performance (Hoobler et al., 2011).
Previous studies have shown that abusive supervision lowers self-efficacy (Harvey et al., 2007), increases psychological pressure such as tension and emotional exhaustion (Duffy et al., 2002), and negatively affects job behaviour and attitudes (Duffy et al., 2002; Tepper et al., 2008). According to Lee et al. while abusive supervision drains employee resources, it may also increase resources like inventiveness, potentially leading to better team performance. Withdrawal behaviour results in inefficient operations, emotional exhaustion, and burnout, affecting coworkers' well-being (Roche and Haar, 2013).
Banking is a crucial sector driving economic progress by providing essential services. Banks are in a competitive environment, and employee engagement is crucial for sustainability. The use of technology has revolutionized banks, and underperforming employees could cost the organization significantly.
Nowadays, employees in the banking sector are leaving due to high stress, mostly caused by abusive supervision (Ahmad and Begum, 2020). The banking sector is known for its high-stress environment, stringent deadlines, customer demands, and financial targets. Additionally, strict performance metrics and regulatory frameworks can prompt abusive tactics from supervisors. Pressure to achieve targets impacts both employee well-being and customer satisfaction.
Selecting the Northern Province for research on abusive supervision and withdrawal behaviour is justified by the region's socio-economic and linguistic context. Post-war expansion of banks in the region has created a competitive environment where managers face pressure to meet targets. Many employees come from rural backgrounds and may struggle to adapt to the demands of the industry, increasing stress. The predominance of the Tamil language may lead to communication challenges and conflict. Studying this context offers insights into organizational dynamics, regional development, and cultural sensitivity in management practices.
In Sri Lanka, especially in licensed commercial banks in the Northern Province, no studies have investigated employee withdrawal behaviour and abusive supervision. To fill this gap, this research sought to answer the following questions:
Does abusive supervision impact the physical withdrawal behaviour of employees in licensed commercial banks?
Does abusive supervision impact the psychological withdrawal behaviour of employees in licensed commercial banks?
According to the research questions, the objectives of the study are:
Examine the impact of abusive supervision on the physical withdrawal behaviour of employees in licensed commercial banks.
Investigate the influence of abusive supervision on the psychological withdrawal behavior of staff-level employees in the banking sector.
2. Literature review
2.1 Theoretical review
Abusive supervision has attracted significant academic attention over the past twenty years due to its very serious consequences for employees' well-being and organizational performance. Tepper (2000) works with a conceptualization of abusive supervision as followers' experiences of persistent hostile verbal and nonverbal acts by their supervisors without involving physical contact. These include ridicule, public humiliation, and intentional undermining (Tepper, 2007). These are not only inimical to professional respect but also influence emotional well-being and organizational functioning. For instance, Wu (2008) established a positive relationship between emotional labor and perceptions of abusive supervision, capturing the psychological toll the employee endures in such mistreatment.
Various theoretical models have been used in previous research to explain processes involved in employees' reactions to abusive supervision. For instance, Ferguson et al. (2011) drew on the crossover and spillover theories to study supervisory abuse's effects on the employees as well as their families. Their findings indicated that workplace abuse reinforces work-to-family conflict and strain in the relationship, ultimately decreasing family satisfaction for both the subordinate and spouse. Similarly, Ali et al. (2022) applied uncertainty management theory and social exchange theory (SET) to highlight how perceptions of unfairness and political processes, evoked by abusive leadership, generate workplace deviance and incivility. These models describe the ways in which breaches of trust and breaches of perceived equity result in acts of retaliation.
Recent empirical studies continue to reinforce the harmful effects of abusive supervision through a Conservation of Resources (COR) theory lens. For example, The effect of abusive supervision variability on work–family conflict: The role of psychological detachment and optimism (Wang et al., 2023) demonstrates that variability in supervisory abuse — when some subordinates experience more abuse than others — significantly increases work–family conflict, mediated by lower psychological detachment.
Although these models have been insightful, the present research adopts Conservation of Resources (COR) Theory (Hobfoll, 1989) as its widest theoretical framework on the grounds that it provides a more integrative and resource-based account of withdrawal acts. COR theory proposes that employees work to acquire, retain, and protect limited resources, time, energy, self-worth, and social capital. Fear arises when these resources are threatened, lost, or insufficiently restored. In abusive supervision contexts, the persistent affective weight and psychological tension serve as potent resource-depleting forces. As a consequence, employees may use withdrawal behaviors, both physical (e.g. absenteeism, tardiness) and psychological (e.g. disengagement, reduced effort), as defensive tactics.
Contemporary empirical studies establish the universality of COR theory in this application. Akram et al. (2022) demonstrated that emotional exhaustion mediates abusive supervision and counterproductive work behavior, particularly under high job demands.
But the link is not always linear or automatic. Mawritz et al. (2012) found that the influence of perceived abuse on withdrawal behavior is moderated by factors such as self-efficacy and perceived organizational justice. These boundary conditions suggest that not all people respond to abusive supervision by withdrawing, some may be resilient or have protection through institutional mechanisms.
By making COR theory its foundation, this study aims to move beyond relational justice and reciprocity-based theories like SET. Instead, it positions withdrawal behaviors as coping strategies in response to resource depletion, hence offering a more psychological explanation for disengagement on the part of employees. The theory further explains both hypotheses of the study, as physical and psychological withdrawal behaviors can be conceptualized as differential yet related reactions to losing valuable personal and professional resources.
2.2 Empirical review
2.2.1 Abusive supervision and physical withdrawal behaviour
Abusive supervision, characterized by the misuse of authority and negative behaviour by supervisors, has been recognized as a significant predictor of employee withdrawal behaviour in various organizational contexts. Research by Tepper (2000) provided early evidence of the detrimental effects of abusive supervision on employee outcomes, including physical withdrawal behavior such as absenteeism and lateness. Tepper's findings suggested that employees who experience abusive supervision are more likely to engage in behaviour aimed at distancing themselves from the workplace. Similarly, Mitchell and Ambrose (2007) found that employees subjected to abusive supervision are more likely to engage in behaviour aimed at distancing themselves physically from the workplace. Similarly, a study by Harris et al. (2018) found that abusive supervision was positively associated with physical withdrawal behaviour among employees, including increased absenteeism and reduced effort on tasks. The study highlighted the role of abusive supervision in creating a hostile work environment that prompts employees to disengage from their work duties.
Further research by Park and Shaw (2013) corroborated these findings, demonstrating a significant relationship between abusive supervision and physical withdrawal behaviour, such as decreased job performance and increased turnover intentions. The study underscored the impact of abusive supervision on employee morale and motivation, leading to withdrawal behaviour aimed at avoiding further negative interactions with supervisors. Several behavioral and attitudinal employee outcomes, including job performance (Hoobler and Hu, 2013), creativity (Liu et al., 2012), job satisfaction (Palanski et al., 2014), organizational support (Kernan et al., 2011), in-role job performance (Xu et al., 2012), corporate citizenship behavior (Zellars et al., 2002), psychological well-being (Schyns and Schilling, 2013), affective commitment (Yu et al., 2016), etc., are negatively correlated with abusive supervision.
Recent empirical evidence continues to support the harmful effects of abusive supervision. For instance, Liu et al. (2025) found that abusive supervision negatively affects employee job performance, with the effects moderated by employment contract type. Khan et al. (2023) demonstrated that abusive supervision significantly reduces service employees' well-being, increasing stress and emotional exhaustion.
The following hypothesis is developed and examined in light of the studies mentioned earlier:
There is a significant impact of abusive supervision on the physical withdrawal behaviour of employees.
2.2.2 Abusive supervision and psychological withdrawal behaviour
Abusive supervision has been consistently linked to various forms of negative employee outcomes, including psychological withdrawal behaviour such as disengagement and reduced job satisfaction. It encompasses behaviours like open denigration (Tepper, 2000), invasions of privacy, taking credit for subordinates' work, placing blame, being rude, and yelling or humiliating employees (Tepper et al., 2006; Mitchell and Ambrose, 2007). It is seen as a hierarchical stressor that might overwhelm a worker's assets (Restubog et al., 2011).
Research supports that abusive supervisors are linked to mental distress and outcomes like dejection, vulnerability, strong well-being protests, and decreased self-confidence (Burton and Hoobler, 2006). Studies (Martinko et al., 2013; Tepper et al., 2017) have shown strong correlations between abusive supervision and increased stress, emotional weariness, turnover, and work-life imbalance. Lin et al. (2013) also noted poorer psychological well-being.
According to Akram et al. (2022), abusive supervision negatively affects employee creativity through perceived psychological distress. Distributive and procedural fairness can reduce these effects. Gilbreath and Benson (2004) noted that supervisor attitudes significantly impact employees' psychological health. Subordinates should recognize the supervisor's role in fostering well-being.
Zellars et al. (2002) describe abusive management as persistent emotional maltreatment, such as public ridicule, withholding information, and coercive tactics. Hoobler et al. (2011) linked abusive supervision with reduced effort and disengagement. Liu et al. (2012) found it leads to emotional exhaustion and reduced organizational commitment.
Duffy et al. (2002) provided early evidence that abusive supervision causes stress and emotional exhaustion, leading to psychological withdrawal. Tepper et al. (2008) also found a significant relationship, with abusive supervision reducing commitment and increasing turnover intentions. Supervisory abuse is linked to job dissatisfaction, emotional strain, and interpersonal conflict (Ashforth, 1997; Tepper, 2000).
Mackey et al. (2017) found that employees in the banking sector exposed to abusive supervision reported higher job dissatisfaction and psychological withdrawal. Empirical data show abusive supervision increases work-family conflict (Carlson et al., 2012), depression (Mackey, 2016), stress (Schyns and Schilling, 2013), emotional exhaustion (Wheeler et al., 2013), and workplace deviance (Wang et al., 2015). Harvey et al. (2007) similarly reported stress and emotional tiredness. From a stress perspective, abusive supervision causes strain reactions like poor mental health and job dissatisfaction.
Nevertheless, studies like Aryee et al. (2007) point out that not all psychologically affected employees disengage; some channel emotional strain into job-focused coping efforts, indicating alternative behavioral responses. Based on the above literature, the following hypothesis is formulated and tested in this study.
There is a significant impact of abusive supervision on the psychological withdrawal behavior of employees.
3. Research method
3.1 Sample and data
The population is all staff-level employees from Licensed Commercial Banks in Northern Province, Sri Lanka. The sample size is 300 staff-level employees from Licensed Commercial Banks in Northern Province, Sri Lanka. Utilizing a quantitative research paradigm, the study employs a survey research design, with data gathered via a structured questionnaire distributed to licensed commercial bank employees at the staff level in the Northern Province. This developed survey covers two main sections. Part A is data related to respondents' demographic profiles, and Part B illustrates research data. 300 questionnaires were printed and issued to respondents, of which 137 were returned. However, 6 responses were rejected due to incomplete questionnaires.
3.2 Respondents profile
The demographic distribution of the sample is shown in Table 1. A significant number of individuals fall within the 25–26 age range, indicating a younger group. Female participants (80) outnumber male participants (51), showing gender diversity. Most participants (81) are employed in government entities, highlighting a strong public sector presence. All the participants are at staff level. Interestingly, 74% have 6–10 years of experience, which is a mid-career group. The majority (92) are married, which suggests a balance between work and family life. Educational levels vary, with more having degrees (56) than diplomas (29) or secondary education (28). The majority (125) earn above 55,000, which suggests financial stability. This demographic profile provides useful insight into the research participants.
Demographic profile of respondents
| Demographic variables | Categories | N | Percentage |
|---|---|---|---|
| Age | Below 25 | 10 | 7.6 |
| 25–26 | 79 | 60.3 | |
| 36–45 | 41 | 31.3 | |
| Above 45 | 1 | 0.8 | |
| Gender | Male | 51 | 38.9 |
| Female | 80 | 61.1 | |
| Sector | Private | 50 | 38.2 |
| Government | 81 | 61.8 | |
| Years of experience | Less 3 | 1 | 0.8 |
| 3–5 | 25 | 19.1 | |
| 6–10 | 74 | 56.5 | |
| Above 10 | 31 | 23.7 | |
| Education | Secondary | 28 | 21.4 |
| Diploma | 29 | 22.1 | |
| Degree | 56 | 42.7 | |
| Postgraduate | 14 | 10.7 | |
| Others | 4 | 3.1 | |
| Level of income | 30,001–45,000 | 2 | 1.5 |
| 45,001–55,000 | 4 | 3.1 | |
| Above 55,000 | 125 | 95.4 | |
| Marital status | Married | 92 | 70.2 |
| Single | 39 | 29.8 |
3.3 Conceptual model
Figure 1 illustrates the relationship between the independent and dependent variables of the study. Abusive supervision is conceptualized as the independent variable, whereas physical and psychological withdrawal behaviours are treated as the dependent variables.
The conceptual model contains a rectangle positioned on the left labeled “Abusive Supervision”. Two rightward arrows extend from this rectangle and point to rectangles positioned on the right. The upper arrow is labeled “H 1” and points to a rectangle labeled “Physical Withdrawal Behaviour”. The lower arrow is labeled “H 2” and points to a rectangle labeled “Psychological Withdrawal Behaviour”.Conceptual model
The conceptual model contains a rectangle positioned on the left labeled “Abusive Supervision”. Two rightward arrows extend from this rectangle and point to rectangles positioned on the right. The upper arrow is labeled “H 1” and points to a rectangle labeled “Physical Withdrawal Behaviour”. The lower arrow is labeled “H 2” and points to a rectangle labeled “Psychological Withdrawal Behaviour”.Conceptual model
3.4 Measures of the study
Table 2 presents the measurement scales used in this study. Abusive supervision was measured using eight (8) items taken directly from Tepper (2000), and withdrawal behavior was assessed using eight (8) items developed by Lehman and Simpson (1992), four items for physical withdrawal and four for psychological withdrawal.
Operationalization
| Concept | Dimensions | Indicators | Measurement |
|---|---|---|---|
| Abusive supervision | Hostile Verbal Behaviors | Public criticism, yelling, ridicule, or humiliation |
|
| Hostile Non-Verbal Behaviors | Glaring, ignoring, excluding from meetings, or giving the silent treatment |
| |
| Withdrawal Behavior | Physical Withdrawal | Absenteeism, tardiness, taking longer breaks, leaving work early |
|
| Psychological Withdrawal | Daydreaming, presenteeism, reduced effort, lack of concentration |
|
These scales were not modified or adapted, and no pilot study or psychometric validation was conducted in the current study context. While these instruments are well-established in prior literature, the absence of local validation is acknowledged as a limitation. Future studies should consider conducting reliability and validity checks (e.g. Cronbach's alpha, factor analysis) to confirm the robustness of the scales in different cultural and organizational contexts.
3.5 Statistical models
Two statistical models are formulated and tested in this study, which are:
Where:
PWB: Withdrawal behavior physical
PWB: Withdrawal behavior psychological
AS: Abusive supervision
AGE: Age
GEN: Gender
SEC: Sector (private or public)
EXP: Experience
MS: Marital status
EDU: Education
INC: Income
3.6 Structural equation model (SEM)
SEM was employed to assess the overall model fit. Abusive supervision was measured using 8 items, while both physical and psychological withdrawal behaviors were measured using 4 items each. As presented in Figure 2, the Default model showed a Normed Fit Index (NFI) of 0.720, indicating that 72% of the variance is explained. Other fit indices like IFI, TLI, and CFI were around 0.755, suggesting moderate model fit.
The structural model begins with a circle labeled “A S” positioned at the left. Multiple leftward arrows extend from “A S” to rectangles arranged vertically on its left side, from top to bottom labeled “A S 1” through “A S 8”. The arrows are labeled “0.879”, “0.832”, “0.782”, “0.857”, “0.842”, “0.873”, “0.900”, and “0.837”, respectively. A diagonal rightward arrow labeled “0.671” extends from “A S” to the circle positioned at the upper right labeled “W B P”, with the value “0.450” inside the circle. Individually, four rightward arrows extend from “W B P” to rectangles arranged vertically on its right side from top to bottom labeled “W B P 1” through “W B P 4”. The arrows are labeled “0.895”, “0.932”, “0.943”, and “0.908”, respectively. A second diagonal rightward arrow labeled “0.659” extends from “A S” to the circle positioned at the lower right labeled “W B S”, with the value “0.435” inside the circle. Individually, four rightward arrows extend from “W B S” to rectangles arranged vertically on its right side from top to bottom labeled “W B S 1” through “W B S 4”. The arrows are labeled “0.887”, “0.846”, “0.901”, and “0.883”, respectively.Structural equation model
The structural model begins with a circle labeled “A S” positioned at the left. Multiple leftward arrows extend from “A S” to rectangles arranged vertically on its left side, from top to bottom labeled “A S 1” through “A S 8”. The arrows are labeled “0.879”, “0.832”, “0.782”, “0.857”, “0.842”, “0.873”, “0.900”, and “0.837”, respectively. A diagonal rightward arrow labeled “0.671” extends from “A S” to the circle positioned at the upper right labeled “W B P”, with the value “0.450” inside the circle. Individually, four rightward arrows extend from “W B P” to rectangles arranged vertically on its right side from top to bottom labeled “W B P 1” through “W B P 4”. The arrows are labeled “0.895”, “0.932”, “0.943”, and “0.908”, respectively. A second diagonal rightward arrow labeled “0.659” extends from “A S” to the circle positioned at the lower right labeled “W B S”, with the value “0.435” inside the circle. Individually, four rightward arrows extend from “W B S” to rectangles arranged vertically on its right side from top to bottom labeled “W B S 1” through “W B S 4”. The arrows are labeled “0.887”, “0.846”, “0.901”, and “0.883”, respectively.Structural equation model
Although the Saturated model showed a perfect fit and the Independence model showed poor fit, the default model's indices suggest the relationships among abusive supervision and withdrawal behaviors are reasonably well represented.
3.7 Data analysis
The SPSS is used to analyze the acquired data. Descriptive statistics are used to highlight important dataset characteristics, and inferential statistics are used to test hypotheses. Ethical considerations include obtaining approval, ensuring participant confidentiality, and storing data securely for research purposes only.
4. Results
4.1 Descriptive
Table 3 shows descriptive statistics. In terms of abusive supervision (AS1- AS8), the mean scores indicate a moderate level of reported abusive behaviour, with some variability across different questions, as reflected by the standard deviations. Similarly, for physical withdrawal behaviour (WBP1-WBP4), participants reported relatively lower levels compared to abusive supervision, yet with notable variability in responses. The mean scores for psychological withdrawal behaviour (WBS1-WBS4) indicate a slightly higher level, also with variability in responses. Overall, the mean total scores for abusive supervision, physical withdrawal behaviour, and psychological withdrawal behaviour provide a comprehensive overview of participants' perceptions across all questions, highlighting the complex interplay between abusive supervision and withdrawal behaviour in the workplace.
Descriptive statistics
| Mean | Std. deviation | |
|---|---|---|
| AS1 | 2.05 | 1.341 |
| AS2 | 2.28 | 1.332 |
| AS3 | 2.21 | 1.222 |
| AS4 | 2.24 | 1.313 |
| AS5 | 2.24 | 1.323 |
| AS6 | 2.07 | 1.229 |
| AS7 | 2.09 | 1.224 |
| AS8 | 2.10 | 1.182 |
| WBP1 | 1.85 | 1.268 |
| WBP2 | 1.79 | 1.209 |
| WBP3 | 1.67 | 1.180 |
| WBP4 | 1.69 | 1.227 |
| WBS1 | 1.79 | 1.183 |
| WBS2 | 1.95 | 1.211 |
| WBS3 | 1.85 | 1.190 |
| WBS4 | 1.91 | 1.255 |
| Total Abusive Supervision (AS) | 2.16 | 1.08 |
| Total Physical Withdrawal Behaviour (WBP) | 1.75 | 1.12 |
| Total Psychological Withdrawal Behaviour (WBS) | 1.75 | 1.06 |
4.2 Reliability
Tables 4 show reliability statistics. The reliability of each construct was assessed using Cronbach's alpha. The Abusive Supervision scale (8 items) demonstrated high internal consistency (α = 0.957). Physical Withdrawal Behaviour (4 items) and Psychological Withdrawal Behaviour (4 items) also showed strong reliability (α = 0.956 and α = 0.957, respectively). The overall scale reliability for all 16 items was excellent (α = 0.959), indicating that the instrument is highly consistent.
4.3 Correlation analysis
Table 5 shows the correlation matrix among the tested variables in this study. Importantly, abusive supervision is positively correlated to physical and psychological withdrawal behaviors. Also, marital status is positively correlated with physical and psychological withdrawal behaviors. On the other hand, Age, gender, sector, and level of income are negatively correlated with physical and psychological withdrawal behaviors. Interestingly, experience and education are positively correlated with physical withdrawal behaviors, but these are negatively correlated with psychological withdrawal behaviors.
Correlation matrix
| Variables | (1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | (9) | (10) |
|---|---|---|---|---|---|---|---|---|---|---|
| AS (1) | 1 | 0.66 | 0.65 | −0.02 | −0.18 | −0.14 | 0.04 | 0.01 | 0.12 | −0.05 |
| WBP (2) | 0.66 | 1 | 0.89 | −0.12 | −0.25 | −0.26 | −0.01 | 0.08 | −0.01 | −0.04 |
| WBS (3) | 0.65 | 0.89 | 1 | −0.10 | −0.26 | −0.23 | 0.03 | 0.03 | 0.06 | −0.03 |
| AGE (4) | −0.02 | −0.13 | −0.10 | 1 | −0.11 | 0.17 | 0.38 | −0.38 | 0.12 | 0.17 |
| GEN (5) | −0.18 | −0.25 | −0.26 | −0.11 | 1 | 0.21 | 0.04 | 0.02 | −0.16 | −0.01 |
| SEC (6) | −0.14 | −0.26 | −0.23 | 0.17 | 0.21 | 1 | 0.17 | −0.25 | 0.36 | −0.00 |
| EXP (7) | 0.04 | −0.01 | 0.03 | 0.38 | 0.04 | 0.17 | 1 | −0.29 | 0.14 | 0.12 |
| MS (8) | 0.01 | 0.08 | 0.03 | −0.38 | 0.02 | −0.25 | −0.29 | 1 | −0.24 | −0.25 |
| EDU (9) | 0.12 | −0.01 | 0.06 | 0.12 | −0.16 | 0.36 | 0.14 | −0.24 | 1 | 0.00 |
| INC (10) | −0.05 | −0.04 | −0.03 | 0.17 | −0.01 | −0.00 | 0.12 | −0.25 | 0.00 | 1 |
| N | 131 | 131 | 131 | 131 | 131 | 131 | 131 | 131 | 131 | 131 |
Note(s): Figures in italic indicate significant at 1%, and figures in underline indicate significant at 5%
4.4 Regression analysis
H1: There is a significant impact of abusive supervision on the physical withdrawal behaviour of employees (Model 1)
As shown in Table 6, the model summary indicates a strong correlation between AS and WBP, with an R value of 0.701. The R2 value of 0.492 suggests that 49.2% of the variance in WBP is explained by AS.
Model summary
| Model | R | R Square | Adjusted R Square | Std. error of the estimate |
|---|---|---|---|---|
| 1 | 0.701a | 0.492 | 0.458 | 0.82598 |
Predictors: (Constant), Independent variable (AS), and Control variables (AGE, GEN, MS, EDU, EXP, SEC, INC)
As presented in Table 7, the ANOVA results indicate that the model is statistically significant (F = 14.752, p < 0.001).
Table 8 shows that AS has a coefficient of 0.649 with a p-value of less than 0.001. This indicates that AS is a significant predictor of WBP. Specifically, a one-unit increase in AS results in a 0.649 unit increase in WBP, demonstrating a strong positive relationship. Thus, H1 is accepted.
Coefficienta
| Model 1 | Unstandardized coefficients | Standardized coefficients | t | Sig | Collinearity statistics | ||
|---|---|---|---|---|---|---|---|
| B | Std. error | Beta | Tolerance | VIF | |||
| (Constant) | 1.658 | 1.233 | 1.345 | 0.181 | |||
| AS | 0.649 | 0.070 | 0.625 | 9.325 | 0.000 | 0.928 | 1.077 |
| AGE | −0.210 | 0.139 | −0.112 | −1.506 | 0.135 | 0.752 | 1.331 |
| GEN | −0.325 | 0.160 | −0.142 | −2.035 | 0.044 | 0.857 | 1.167 |
| SEC | −0.250 | 0.174 | −0.109 | −1.440 | 0.152 | 0.733 | 1.364 |
| EXP | 0.055 | 0.118 | 0.033 | 0.461 | 0.646 | 0.814 | 1.229 |
| MS | 0.010 | 0.175 | 0.004 | 0.056 | 0.956 | 0.751 | 1.331 |
| EDU | −0.060 | 0.079 | −0.056 | −0.762 | 0.448 | 0.770 | 1.299 |
| INC | 0.017 | 0.254 | 0.004 | 0.067 | 0.947 | 0.921 | 1.086 |
Note(s): aDependent Variable: WBP
H2: There is a significant impact of abusive supervision on the psychological withdrawal behaviour of employees (Model 2)
Table 9 shows that the model summary reveals that AS is significantly correlated with WBS among employees in licensed commercial banks in the Northern Province of Sri Lanka. The strong correlation (R = 0.686) indicates that higher levels of abusive supervision tend to coincide with increased physical withdrawal behaviour.
Model summary
| Model | R | R Square | Adjusted R Square | Std. error of the estimate |
|---|---|---|---|---|
| 2 | 0.686a | 0.470 | 0.435 | 0.79955 |
Predictors: (Constant), Independent variable (AS), and Control variables (AGE, GEN, MS, EDU, EXP, SEC, INC)
Table 10 shows that the ANOVA table shows that the model is statistically significant with an F value of 13.527 and a p-value of less than 0.001.
Table 11 shows that the constant term (B = 1.622, p = 0.177) represents the intercept of the regression equation, which is not statistically significant. The key predictor, AS, has a coefficient of 0.594 with a p-value of less than 0.001. This indicates that AS is a significant predictor of WBS. Therefore, H2 is accepted.
Coefficienta
| Model 2 | Unstandardized coefficients | Standardized coefficients | T | Sig | Collinearity statistics | ||
|---|---|---|---|---|---|---|---|
| B | Std. error | Beta | Tolerance | VIF | |||
| (Constant) | 1.622 | 1.193 | 1.359 | 0.177 | |||
| AS | 0.594 | 0.067 | 0.604 | 8.826 | 0.000 | 0.928 | 1.077 |
| AGE | −0.212 | 0.135 | −0.119 | −1.568 | 0.119 | 0.752 | 1.331 |
| GEN | −0.317 | 0.155 | −0.146 | −2.047 | 0.043 | 0.857 | 1.167 |
| SEC | −0.241 | 0.168 | −0.111 | −1.435 | 0.154 | 0.733 | 1.364 |
| EXP | 0.104 | 0.115 | 0.066 | 0.906 | 0.366 | 0.814 | 1.229 |
| MS | −0.040 | 0.169 | −0.018 | −0.237 | 0.813 | 0.751 | 1.331 |
| EDU | 0.007 | 0.077 | 0.007 | 0.091 | 0.928 | 0.770 | 1.299 |
| INC | 0.017 | 0.246 | 0.005 | 0.068 | 0.946 | 0.921 | 1.086 |
Note(s): aDependent Variable: WBS
5. Discussion
The results of this investigation revealed the significant impact of abusive supervision on both physical and psychological withdrawal behaviour among workers in the banking industry. The results indicate a clear positive association between supervisors' abusive behaviour and employees' withdrawal behaviour, suggesting that as instances of abusive supervision increase, Workers are more likely to stop participating in their work physically and psychologically. This is consistent with earlier studies showing the negative impacts of abusive management techniques on employees' intentions to leave the company and their commitment. To preserve a positive and healthy work environment, companies must acknowledge the negative effects of abusive supervision and take proactive steps to confront and avoid such practices.
According to Ma et al. (2020), their research on hotels revealed that abusive supervision lowers service personnel's sense of organizational identity, which in turn affects how they feel. In light of this, my investigation revealed a clear correlation between employees' withdrawal behaviour and abusive supervision.
To the best of the reviewer's knowledge, no previous research has precisely quantified the effect of an abusive supervisor's actions on workers' withdrawal behaviour. The current study's findings, however, are consistent with the themes of other earlier investigations (e.g. Decoster et al., 2013; Hoobler and Hu, 2013; Palanski et al., 2014; Yu et al., 2016). Research has demonstrated that employees' attitudes and behaviour (such as organizational support, intention to leave, and psychological well-being) are negatively impacted by abusive supervision. The generalizability of the theory that addresses abusive supervision as a detrimental construct of employees' behavioural and attitudinal outcomes is indicated by consistency of the findings.
This study discovered a favorable correlation between the abusive behaviour of supervisors and the psychological and physical withdrawal behaviour of employees. This correlation implies that when supervisors behave abusively, employees withdraw more psychologically and physically, and when supervisors behave less abusively, employees withdraw less psychologically and physically. When a supervisor behaves abusively in the workplace, staff members withdraw from the position and the company. At the same time, researchers (Liu et al., 2012; Tepper, 2000) have discovered that employees' intentions to leave an organization are positively impacted by abusive management. Therefore, increasing abusive supervision inside an organization has an impact on employee turnover intentions, additionally, this will make workers want to quit the company and start hunting for new work. This study supports the findings of Khan (2015) by demonstrating how abusive supervisory behaviour influences employee turnover.
6. Conclusion
This study demonstrates a significant positive relationship between abusive supervision and both physical and psychological withdrawal behaviours among employees in the banking sector, particularly in Sri Lanka's Northern Province. In line with Conservation of Resources (COR) theory, employees engage in withdrawal behaviours to conserve depleted emotional and cognitive resources. The findings highlight the importance of addressing abusive supervisory practices to reduce turnover intentions, absenteeism, and disengagement, and to foster a healthier, more productive work environment. Practically, interventions such as supervisor training, supportive policies, and stress management initiatives are recommended. This research contributes to the human resource management literature by providing context-specific insights into abusive supervision and employee withdrawal.
6.1 Practical implications
The study underscores the detrimental impact of abusive supervision on withdrawal behaviors among banking staff, encompassing both physical and psychological manifestations. To mitigate this, firms should prioritize interventions such as supervisor training in stress management and conflict resolution. Creating a supportive workplace environment through policies that discourage abusive behaviors and promote respectful communication is crucial. HR strategies should focus on enhancing supervisor-subordinate relationships and fostering a culture of trust and fairness. By addressing these issues proactively, organizations can reduce employee turnover, enhance job satisfaction, and cultivate a more productive workforce in the banking sector of Sri Lanka.
In addition to interpersonal interventions, organizations should integrate formal policies that institutionalize a zero-tolerance approach to abusive supervision. This may include implementing anonymous reporting systems, routine supervisor evaluations tied to respectful leadership indicators, and clear protocols for corrective action when abuse is reported. Such policy-level changes ensure that respectful leadership is embedded into the organizational culture and actively enforced, thereby reducing the likelihood of employee withdrawal and turnover.
6.2 Limitations and future directions
Individual response measurements may pose issues with common method variance. The validity of the data is dependent on the respondents' attitudes, and the dimensions employed in our study are difficult to measure outside of self-response. The sample size is limited to 131 respondents from a specific region in Sri Lanka, which may restrict the generalizability of the findings to other regions or sectors. Moreover, the study lacks robustness checks such as bootstrapping or cross-validation, which could have strengthened the reliability and statistical stability of the results.
Data collection exclusively employed a Likert scale questionnaire, omitting the richer insights that could emerge from interview-based methods. Future research could incorporate qualitative methods, such as interviews or focus groups, to better understand the nuanced experiences of abusive supervision and withdrawal behaviors.
In addition to methodological improvements, future research could explore moderating variables such as emotional intelligence, perceptions of organizational justice, or employee resilience, which may buffer the negative impact of abusive supervision. Longitudinal research designs would provide stronger causal inferences and allow tracking of how supervisory abuse and withdrawal behaviors evolve over time. Additionally, conducting cross-industry or cross-cultural studies could enhance the generalizability of findings and illuminate how cultural norms shape supervisory conduct and employee coping responses. Although the study used established scales for abusive supervision and withdrawal behaviour, these instruments have not been validated in the Sri Lankan context. Therefore, cultural and contextual differences may affect how respondents interpret the items, which could influence the results.
Notes
Sri Lanka's banking sector may have seen about 10,000 employees leaving in 2023 out of which a significant number has migrated, as banks resort to massive recruitment drives to fill in the vacancies, a top banker said (Imesh, 2023).
