Workplace bullying and turnover intentions among workers: a systematic review and meta-analysis

Abstract

Background

Workplace bullying occurs in various professions worldwide and significantly impacts employees and organizations. Researchers have reported that workplace bullying may be associated with an increased intention to leave one’s job. However, the conclusions regarding this relationship remain inconsistent.

Methods

We searched the Cochrane Library, PubMed, Web of Science, Embase, Scopus, PsycInfo, and ProQuest databases from inception to September 20, 2024. Two authors independently screened the studies, assessed the quality of the included studies, and extracted data. Any disagreements were resolved through discussion with a third author. A meta-analysis was performed to combine the regression coefficient (B) or odds ratios (ORs) with their confidence intervals. The I² statistic was used to quantitatively evaluate the degree of heterogeneity. Stata version 18.0 was used to conduct the meta-analysis, sensitivity analysis, and evaluation of publication bias risk.

Results

This systematic review and meta-analysis included 27 studies, including 3 prospective and 24 cross-sectional studies. In the cross-sectional studies, as combined effect sizes, the B and OR values revealed a significant association between the experience of workplace bullying and an increased intention to leave (B = 0.25, 95% CI [0.19, 0.31]) (OR = 1.30, 95% CI [1.17, 1.44]). The prospective studies indicated that experiencing workplace bullying is associated with increased turnover intentions over time (B = 0.09, 95% CI [0.01, 0.17]).

Conclusion

Experiencing workplace bullying is positively correlated with employees’ increased intentions to leave. This result suggests that to reduce employee turnover in organizations, it is crucial to focus on and prevent workplace bullying.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-23339-2.

Introduction

Definition of workplace bullying and turnover intentions

Workplace bullying is a systematic form of interpersonal harm [1]. Victimized employees experience prolonged mistreatment by superiors, colleagues, or subordinates within the work environment, including constant abuse, offensive remarks or teasing, ridicule, the assignment of unreasonable tasks, or social exclusion, and victims of workplace bullying struggle to protect themselves [1]. Compared with other negative workplace behaviors, workplace bullying is distinguished by its recurring nature, sustained duration, and presence of power imbalances [2]. These power imbalances refer not only to job positions but also to factors such as age, job tenure, and even gender [3].

Turnover, which is a common phenomenon in organizational behavior research, is a withdrawal behavior exhibited by individuals following unsatisfactory experiences at work [4]. Turnover intention is the conscious and deliberate willingness of an employee to leave an organization; it is considered a crucial direct predictor of actual turnover behavior [5, 6]. Employees’ departure behaviors following intentions to leave fall within the scope of voluntary turnover. Compared with involuntary turnover, such as layoffs, voluntary turnover is unpredictable and undesirable for organizations, often resulting in more negative impacts on organizational development [7].

The magnitude and importance of the study variables

Workplace bullying is a global phenomenon, and a meta-analysis of prevalence rates indicated that approximately 15% of the global workforce experiences various forms of workplace bullying [8]. A study on the prevalence of workplace bullying among employees in Hong Kong, China, revealed annual and lifetime incidence rates of 39.1% and 58.9%, respectively [9]. In the United Kingdom, a survey revealed that 10.6% of employees reported experiencing workplace bullying and harassment in the past year [10]. In South Africa, a study showed that 31.1% of the sample had experienced workplace bullying [11].

The negative impacts of workplace turnover intentions have been evidenced in studies across different organizations and social contexts. Workplace bullying violates employees’ fundamental rights and can harm both their mental and physical health. Research indicates that workplace bullying increases the likelihood of psychological stress responses, depression and anxiety, sleep disorders, and emotional exhaustion and reduces subjective well-being [12–14]. It may even contribute to the development of physical health conditions such as cardiovascular diseases and type 2 diabetes [15, 16]. Moreover, workplace bullying has significant negative effects on organizations. For example, it damages interpersonal relationships, worsens the work environment, erodes employees’ trust in the organization, reduces work efficiency and quality, and hinders organizational innovation and improvement [17, 18]. Therefore, workplace bullying is not only an individual issue but also a critical factor contributing to the deterioration of organizational environments.

As a critical precursor to actual turnover, turnover intentions pose a significant challenge for organizations [19]. They can lead to reduced employee loyalty and emotional attachment to the organization, diminished commitment to employees’ work, and decreased motivation, with employees potentially biding their time until they can leave [20, 21]. If employees eventually decide to leave, this decision results in turnover and incurs costs related to training, recruitment, replacement, and operational disruptions [4]. Even if employees do not actually depart, those with turnover intentions are unlikely to make meaningful contributions to their organizations [22]. Therefore, it is important to understand the determinants of turnover intentions.

The necessity of conducting this study

Numerous studies have examined the association between workplace bullying and the intention to leave. Among the consequences of workplace bullying, turnover intentions are especially significant because they directly correlate with employee turnover [19]. As turnover intentions precede actual turnover, accurately predicting them allows employers and policy-makers to intervene in time and prevent actual turnover [23]. Turnover intentions have a profound effect on the stability and development of organizations.

Several original studies from various countries and industries have demonstrated the correlation between workplace bullying and employees’ turnover intentions [24–27]. Workplace bullying can directly influence the intention to leave and can also indirectly influence it through mediating variables, such as job burnout [25, 28]. Workplace bullying behavior intensifies employees’ disinterest in their jobs, leading to burnout and increased turnover intentions.

While workplace bullying may lead to greater tendencies to leave, there is currently no consistent conclusion on this matter. Some studies have not supported this view. For example, a survey conducted in Denmark targeting permanent hospital staff revealed that, after controlling for self-rated health, age, and other factors, experiencing workplace bullying in the past year did not significantly influence intentions to leave [29]. Similarly, a study of Korean nurses revealed that, with the exception of face-to-face bullying, other types of bullying behaviors were not significantly associated with intentions to leave [30]. Differences in regional social background and variations in occupation may be potential underlying factors contributing to the inconsistency in the relationship between workplace bullying and turnover intentions [31].

The practical purpose of this study

This systematic review and meta-analysis aim to conduct a statistical integration and synthesis of the existing evidence. The primary objectives are to answer the following two questions: (1) Does experiencing workplace bullying increase employees’ intentions to leave? (2) Is the impact of workplace bullying on turnover intentions consistent across studies in different regions and occupations?

Materials and methods

Protocol and registration

The protocol of the present study was registered with the International Prospective Register of Systematic Reviews (PROSPERO, Registration No. CRD42024582412). The study was designed and reported in accordance with the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines, which are the standard for meta-analyses of observational research [32].

Literature search

We searched the literature on workplace bullying and turnover intentions across seven databases: the Cochrane Library, PubMed, Web of Science, Embase, Scopus, PsycInfo, and ProQuest. The search period spanned from the establishment of these databases to September 20, 2024. Additionally, we performed a manual search using literature references that were previously retrieved. The workplace-related keywords included “workplace,” “work,” “occupation*,” “job,” “colleague,” “organization,” and “employee”; the bullying-related keywords included “bull*,” “mobbing,” and “harassment”; the turnover intention-related keywords included “turnover intention,” “turnover*,” “demission,” “quit,” “departure,” and “leave intention.” The search strategies were appropriately modified in accordance with the characteristics of each database, and there were no language restrictions on the retrieved literature. The retrieval results for each specific database are presented in Table S1.

Study selection

The inclusion criteria for the literature were as follows: (1) quantitative studies; (2) study population: employees from any profession, regardless of age, ethnicity, nationality, or gender; and (3) studies with sufficient information to calculate effect sizes, such as reported regression coefficients (B) or odds ratios (ORs) relating workplace bullying to turnover intentions. The exclusion criteria were as follows: (1) qualitative studies or theoretical discussion papers; (2) articles or conference abstracts lacking data or where the authors failed to provide sufficient data; (3) duplicate literature; (4) studies with effect sizes other than OR and B values; and (5) studies with multiple forms of workplace bullying but no total score for workplace bullying.

Two independent reviewers (Sun and Chen) screened the titles, abstracts, and full-text articles for eligibility. To assess the interrater reliability between the two reviewers, we used the kappa statistic to determine consistent study selection and data extraction. The results revealed that the kappa value for data extraction was 0.897(95% confidence interval (CI): 0.809–0.985). Disagreements were resolved by a third reviewer (Tung).

Characteristics of included studies

We retrieved 2,763 articles from 7 databases. After 946 duplicates were removed, 1,817 articles were screened. After the titles and abstracts were read, 176 studies were identified. After the exclusion of 11 studies whose full texts could not be retrieved, 165 studies underwent full-text evaluation. During the full-text evaluation, 5 conference abstracts, 6 qualitative studies or theoretical discussion papers, 91 studies that failed to provide sufficient data, 32 studies excluding effect sizes other than OR and B values, and 4 studies that, although involving multiple forms of workplace bullying (e.g., person-related, intimidation-related, or work-related workplace bullying), did not report the association between overall workplace bullying scores and turnover intentions were excluded. After screening the titles, abstracts, and full texts of the articles, 27 studies were included [12–38]. Figure S1 shows the PRISMA diagram of the specific screening process.

We used a data extraction form to extract the author, publication year, sample size, study type, region where the sample was collected, age, percentage of female participants, instruments used to measure workplace bullying and turnover intentions, and data reflecting the association between the two variables from the included literature (Table 1). We evaluated the quality of the included literature using the Newcastle–Ottawa Scale (NOS), which is appropriate for cohort studies or cross-sectional studies, and we scored them in terms of three aspects: selection, comparability, and outcome.

Table 1.

Characteristics of selected studies

Study Type	Sample size	female(%)	age(mean ± SD,n(%))	occupation	Measure of Workplace Bullying	Result	NOS
A 3-month prospective study	422	40.70%	54.62 ± 9.14	majority (72%) of participants worked in management and administration or in sales and customer service	A short version of the Negative Acts Questionnaire–Revised (NAQ-R;Einarsen, Hoel & Notelaers, 2009)	B = 0.102, SE = 0.055 adjusted for tyrannical leadership, coworker support	6
A 1-year prospective study	2037	43.70%	42.8 ± 11.4	employees in the public sector	the 22-item Negative Acts Questionnaire-Revised (NAQ-R)	Division-level workplace bullying B = 0.169, SE = 0.058 adjusted for gender, educational status, marital status, occupation, employment status, shift work, and chronic condition	7
A 6-month prospective study	741	14.10%	44 ± 8.9	offshore workers in the North Sea	NAQ-R(Revised version of the Negative Acts Questionnaire,Hauge et al., 2007)	B = 0.04, SE = 0.014 adjusted for age,gender,job insecurity, intention to leave(baseline)	6
Cross-sectional	585	86.80%	29.63 ± 5.68	Nurses	Negative Acts Questionnaire (Chinese version; NAQ)	B = 0.096, SE = 0.009 adjusted for organizational commitment	8
Cross-sectional	318	46.20%	25–35 year 31.1% 36–45 years 45.9% 46–55 years 23.0%	university academic and general staff	The Negative Acts Questionnaire-Revised (NAQ-R)	β = 0.279, SE = 0.058 adjusted for workplace mistreatment, workplace incivility, transformational leadership	8
Cross-sectional	678	74.40%	39 ± 10.91	employees working in the child welfare service	employees working in the child welfare service	B = 1.35, SE = 0.34	8
Cross-sectional	168	94%	45 to 53 years and 54 years or older (68%)	Nurses	22 items (NAQ-R)	B = 0.20,SE = 0.09 adjusted for exhaustion, organizational commitment, job satisfaction	7
Cross-sectional	163	92.6	28.1 ± 4.8	Nurses	Development and testing of a nurse turnover intention scale(NTIS).(Yeun et al.,2013)	B = 0.09 SE = 0.04 adjusted for age, experience in current unit, structural empowerment, psychological empowerment	7
Cross-sectional	867	64.82%	18—65 +	66% worked in large organizations 20.4% worked in medium sized organizations 13.8% working in small organizations	the Short Negative Acts Questionnaire(Notelaers et al., 2019)	B = 0.804, SE = 0.239 adjusted for iclusion, psychological distress,psychosocial safety climate	8
Cross-sectional	126	100%	26.5 ± 3.6	Nurses	The Negative Acts Questionnaire, NAQ	B = 0.15,SE = 0.04 adjusted for age, religion, marital status, ward department, shift work, monthly income	7
Cross-sectional	1024	13.50%	48.48 ± 10.48	Nurses	the Negative Acts Questionnaire-Revised	B = 0.386,SE = 0.122 adjusted for age, injustice perceptions	8
Cross-sectional	1008	92.50%	27.42 ± 6.36	Nurses	A shortened (3-item) version of the NAQ developed by Einarsen and Hoel (2001)	B = 0.303,SE = 0.053 adjusted for structural empowerment, age, sex, education	8
Cross-sectional	178	95.50%	31.5 ± 7.50	Nurses	Moon (2016) and Kim Kang (2017)	B = 0.20, SE = 0.18 adjusted for age, marriage status, highest educational degree held, total experience as a nurse, designation, health status, hospital size, bully,self-esteem	7
Cross-sectional	133	88.70%	25.6 ± 4.4	Nurses	NAQ-R(Revised version of the Negative Acts Questionnaire,Hauge et al., 2007)	B = 0.183,SE = 0.087 adjusted for education,total clinical experience,social support	7
Cross-sectional	197	60.40%	45.2 ± 12.6	Veterinarian	the 22-item negative acts questionnaire (Einarsen et al. 2009)	B = 0.25, SE = 0.10 adjusted for destructive leadership, team conflict, physical health	8
Cross-sectional	373	63.00%	44 ± 8.47	individuals employed in various South African organizations	The 22-item Negative Act QuestionnaireRevised (NAQ-R) et al. (2009)	B = 0.25,SE = 0.04 adjusted for work engagement,vigor,dedication	7
Cross-sectional	2846	56.30%	42 ± 10.6( 18–84)	the occupations of finance/insurance, information, management, and education	the Negative Acts Questionnaire (Einarsen and Hoel, 2001)	B = 0.02,SE = 0.001 adjusted for age, sex,organizational virtuousness, and positive individual treatment	8
Cross-sectional	708	98.20%	30.17 ± 7.15	Nurses	Negative Acts QuestionnaireRevises, NAQ-R	B = 0.14, SE = 0.01 adjusted for have religious belief, formal employee, the department	8
Cross-sectional	1023	13.40%	48.5( 21–69)	Bus drivers	the Negative Acts Questionnaire Revised (NAQ-R; Einarsen, Hoel& Notelaers, 2009)	B = 0.76, SE = 0.10 adjusted for trait anger	7
Cross-sectional	462	13.90%	45 ± 11.77	employees in a maritime transportation company	the Negative Acts Questionnaires (NAQ;Einarsen & Raknes, 1997)	B = 0.74, SE = 0.14	7
Cross-sectional	649	——	——	a large USA-based Multi-National Corporation	——	B = 0.51, SE = 0.10 adjusted for job satisfaction, absenteeism, anxiety	6
Cross-sectional	450	86.70%	39.1 ± 10.2	Nurses	22-item Negative Acts Questionnaire-Revised (NAQ-R)	aOR 1.057 (1.043–1.071) adjusted for sex,age,clinical experience,shift work,understaffed department	7
Cross-sectional	8800	89.90%	< 25 4.9% 25–29 9.5% 30–39 19.6% 40–49 26.2% 50–59 28.3% 60–65 10.4% 66–69 1.0%	Nurses	Have you been exposed to bullying at your workplace during the last 12 months? The response options were ‘No’, ‘Yes, daily’, ‘Yes, weekly’, ‘Yes, monthly’, and ‘Yes, a few times’	aOR No 1 Yes, a few times 2.94[2.501–3.445] adjusted for gender, age group, and type of service	9
Cross-sectional	345	89%	46.6 ± 12.2[22, 73]	Nurses	the Negative Acts QuestionnaireRevised (NAQR)	aOR 1.070[1.044, 1.097] adjusted for age, gender,race/ethnicity, educational degree, years in current position, PSS stress total score, RS-25 resilience total score, hospital vs. non-hospital employee status, working at a magnet hospital status, working full-time vs. otherwise, and shift worked most frequently (night vs. day, other vs. day)	8
Cross-sectional	312	94.23%	23.7(20–24)	Nurses	the Copenhagen Psychosocial Questionnaire (COPSOQ II)	aOR 2.20[1.25, 3.89] adjusted for patient,patient's family,nurse colleague,nurse manager,physician	8
Cross-sectional	242	87%	45 ± 9.7(22–63)	special school staff	the Copenhagen Psychosocial Questionnaire (COPSOQ II)	aOR 1.04[1.00–1.08] adjusted for age,quantitative demands,emotional demands,equipment	8
Cross-sectional	1782	75%	Would quit 44.9 ± 11.1 Would not quit 42.6 ± 11.3	hospital employees	——	aOR 1.47 [0.98–2.17] adjusted for gender; influence on work tasks; equal treatment independent of job function, gender, age, ethnic background, etc.; communication from closest superior	7

Table 1 presents the characteristics of the included studies. All studies originated from 5 continents and 13 countries. Among them, there were 3 longitudinal studies with a total sample size of 3,200 participants and 24 cross-sectional studies with a total sample size of 23,437 participants. The occupations involved in the studies included employees in the public sector, child welfare services, schools, hospitals, companies, offshore workers, and bus drivers.

Analysis

Stata 18 was used for the meta-analysis and subgroup analysis. Regression coefficients (B) or odds ratios (ORs), along with their 95% confidence intervals, were used as effect sizes for integration. The I² statistic was employed to assess heterogeneity. An I² value of ≥ 50% indicated that there was heterogeneity among studies rather than random variability. For issues with heterogeneity, this study utilized a random effects model. Conduct subgroup analyses to identify the sources of heterogeneity. Publication bias was assessed using funnel plots and Egger’s test. Funnel plots are used to visually assess whether the data points are symmetrically distributed; if there is clear asymmetry, then the plot indicates significant publication bias. If Egger’s test results in a P value less than 0.05, then the result suggests significant publication bias. If significant publication bias is present, then we further use the trim-and-fill method to account for studies that may not have been published owing to publication bias, thus adjusting the estimate of the effect size to reflect a more accurate research outcome. If the adjusted effect size differs significantly from the original estimate, then publication bias may have substantially influenced the research conclusions.

Results

The meta-analysis of cross-sectional studies

Eighteen cross-sectional studies used regression coefficients as the effect size. The heterogeneity test results show I² = 96.28%, indicating heterogeneity among the studies; therefore, a random effects model was used. The meta-analysis results in Fig. 1 indicate that workplace bullying is significantly associated with employee turnover intentions (B = 0.25, 95% CI [0.19, 0.31]).

Fig. 1.

Forest plot of cross-sectional studies with regression coefficient (B) as the effect size

The publication bias funnel plot of the correlation indicators between workplace bullying and turnover intentions, along with the results of Egger’s test (t = 5.67, P < 0.001), suggests that there was publication bias in this study (Figure S2). Further correction of the results using the trim-and-fill method indicates that after data from 10 hypothetical studies were included and the meta-analysis of all studies was repeated, the combined results still show a positive correlation between workplace bullying and turnover intentions.

Using the stepwise exclusion method, after one study was excluded, the remaining studies were combined, and all the results consistently show a significant positive correlation between workplace bullying and turnover intentions, which aligns with the meta-analysis results (Figure S3). The combined results are robust.

Five cross-sectional studies used odds ratios as the effect size, with I² = 96.99%, and there was strong heterogeneity among the studies. Hence, a random effects model was adopted. The meta-analysis results shown in Fig. 2 show that workplace bullying is significantly associated with the intention to leave (OR = 1.30, 95% CI [1.17, 1.44]). Due to the limited number of studies using odds ratio as the effect size measure, publication bias assessment was not conducted.

Fig. 2.

Forest plot of cross-sectional studies with Odds ratio (OR) as the effect size

Sensitivity analysis of the combined results reveals that after any one study was excluded, the combined results of the remaining studies still have statistical significance. This finding is consistent with the meta-analysis results, indicating that the combined results are robust (Figure S4).

The meta-analysis of prospective studies

Among the included studies, three had a prospective design, with all of them using regression coefficients as the effect size. The heterogeneity test results show I² = 64.37%. Therefore, a random effects model was adopted. In the meta-analysis of prospective studies shown in Fig. 3, workplace bullying has a significant effect on turnover intentions (B = 0.09, 95% CI [0.01, 0.17]). Owing to the limited number of longitudinal studies, no publication bias tests or sensitivity analyses were conducted.

Fig. 3.

Forest plot of prospective studies

Subgroup analysis

Subgroup analyses were conducted by occupation and region to examine whether this relationship differs across professions and regional cultures. As shown in Fig. 4a and Fig. 4b, workplace bullying is significantly associated with turnover intentions in both Asian countries and countries on other continents. However, based on regression coefficients, the magnitude of the association between workplace bullying and turnover intentions is smaller in studies from Asian regions. Categorization was performed according to occupation, which was divided into two categories, nurses and nonnurses, as nurses accounted for a large proportion of the occupations included in the studies for this review, and the rest of the occupations were more dispersed. Workplace bullying is significantly associated with turnover intentions across all occupational categories. Notably, the strength of this association is weaker in the nurse population than in the nonnurse population.

Fig. 4.

a Subgroup analysis by region. b Subgroup analysis by occupation

Discussion

Main findings

This article conducted a systematic review and meta-analysis of all published studies on the correlation between workplace bullying and turnover intentions among employees across seven commonly used databases. In cross-sectional studies, regardless of whether regression coefficients or odds ratios were used as the effect size, the meta-analysis consistently indicated an association between workplace bullying and turnover intentions. The meta-analysis results of three longitudinal studies showed that experiencing workplace bullying significantly increases future intentions to leave.

The meta-analysis results of the cross-sectional studies are consistent with the findings of Islam’s literature review, indicating a positive correlation between workplace bullying and turnover intentions [33]. A qualitative systematic review targeting nurses indicated that for many nurses experiencing workplace bullying, resignation is the only solution for coping with bullying. Research has indicated that the impact of workplace bullying on turnover intentions occurs primarily through certain mediating variables. Examples include organizational commitment, job dissatisfaction, work engagement and emotional experiences [24, 27, 34, 35].

Conservation of resources theory can effectively explain the impact of workplace bullying on employees’ turnover intentions [36]. According to conservation of resources theory, individuals tend to strive to acquire, maintain, cultivate, and protect the resources that they cherish. Exposure to bullying clearly threatens personal resources, damaging aspects such as job security, needs, self-esteem, and health [37]. Employees’ motivation to avoid further losses of resources may lead them to choose to leave their current work environment and seek other employment opportunities.

Some studies also explain the association between workplace bullying and turnover intentions from the perspectives of psychological contracts and organizational justice theory [38]. When employees believe that they have not received the expected returns (e.g., being subjected to workplace bullying), they feel that their psychological contract has been violated, which in turn leads to turnover intentions. Organizational justice refers to the fairness perceived by employees within an organization, and it includes three forms: procedural, distributive, and interactional justice [39]. When employees are bullied, all forms of fairness are undoubtedly compromised, and their mental and physical health may suffer as a result. Ultimately, they may choose to resign.

Subgroup analyses revealed that workplace bullying is significantly associated with turnover intentions across countries and occupations. However, when regression coefficients are examined, the association between workplace bullying and turnover intentions is found to be weaker in Eastern countries and among nursing professionals. Cultural differences might explain why the link between workplace bullying and turnover intentions varies across regions. Research has shown that employees from cultures that accept power differences as a result of structural factors (high power distance countries, e.g., Eastern countries) exhibit weaker negative reactions to workplace bullying than do employees from societies that are less inclined to accept such power differences (low power distance countries, e.g., Western countries) [31]. Differences among professions may arise for the following reasons. First, compared with individuals working in private companies, nurses have a more stable work environment. Second, nursing is predominantly a female profession, and women generally have slightly lower levels of self-esteem than men do [40]. Self-esteem can moderate the impact of workplace bullying on employees’ intentions to leave, and individuals with lower self-esteem may be less likely to develop intentions to leave when facing such situations [41]. Furthermore, research has shown that women have greater resilience in handling stressful situations, which makes them less likely to develop intentions to leave [42]. To further explore the association between workplace bullying and turnover intentions across different regions and professions, future research is needed.

Theoretical and practical implications

Theoretical Implications: This review integrates existing studies, and through meta-analysis, it validates the result that workplace bullying increases employees’ turnover intentions across different professions and regions. This finding suggests that workplace bullying plays a significant role in research related to employee retention and turnover. Furthermore, the literature included in this review employed various theories to explain the relationship between workplace bullying and turnover intentions. These contributions enhance our comprehensive understanding of the phenomenon and provide a theoretical foundation for future research and practice in this field.

Practical Implications: The results of this review indicate that to reduce employee turnover in organizations, it is crucial to focus on and prevent workplace bullying. Organizations should implement interventions to reduce the incidence of workplace bullying, such as clearly defining bullying behaviors, emphasizing their serious consequences, and providing channels to address bullying incidents. Efforts should also be made to create a healthy work environment, including regular training for employees and managers to identify and respond to bullying behaviors. Moreover, given the differences in the association between workplace bullying and intentions to leave across different professions and regions, different handling strategies should be adopted for various professions and regions. In addition, organizations should pay attention to the mediating variables identified in existing research that link workplace bullying to intentions to leave, such as negative emotions, burnout, and reduced work engagement. When employees exhibit these signs, attention should be paid to whether there is bullying in their current work environment and their intentions to leave.

Heterogeneity and publication bias in the meta-analysis

Heterogeneity is a factor that needs to be considered in meta-analyses, and excessive heterogeneity may lead to unreliable results [43]. The I² statistic is commonly used to reflect the proportion of heterogeneity among included studies in the total variation. The studies included in this review exhibited high heterogeneity, with 96.28% and 96.99% heterogeneity for cross-sectional studies using B values and OR values as effect sizes, respectively, and 64.37% heterogeneity for prospective studies. Subgroup analysis was further conducted to explore the sources of heterogeneity, and it was found that heterogeneity between groups decreased after grouping by study region and occupation. However, the heterogeneity of the results remained unavoidable.

Publication bias is defined as the failure to publish the results of a study on the basis of the direction or strength of the study findings [44]. This bias refers to statistically significant positive results being more likely to be published than negative or nonsignificant results, resulting in a reporting bias. This study utilized funnel plots and Egger’s test to assess publication bias and identified significant publication bias in studies employing the regression coefficient B as the effect size. However, following correction with the trim-and-fill method, a positive correlation between workplace bullying and turnover intentions persisted. This finding indicates that publication bias has a limited impact on the results of the meta-analysis.

Limitations

This meta-analysis has certain limitations: (1) The literature related to the topic retrieved in this article was primarily cross-sectional, with only three longitudinal studies. This makes the results of merged longitudinal studies less robust. Therefore, it is not possible to draw an accurate conclusion regarding the causal relationship between workplace bullying and turnover intentions. (2) Most of the included studies controlled for potential confounding factors, but the controlled confounding factors differed among studies. Some studies did not control for confounding factors that have been proven to be associated with turnover intentions, which may have affected the accuracy of the results. (3) There was high heterogeneity among the studies, which may be due to inconsistent measurements of workplace bullying and turnover intentions in the included studies.

Conclusion

Across both cross-sectional and prospective studies, the meta-analysis results consistently revealed a positive correlation between workplace bullying and increased employee turnover intentions. However, to validate the predictive effect of workplace bullying on turnover intentions, further prospective studies are needed. Future research should also consider the impact of factors such as regional culture and occupation on the relationship between workplace bullying and turnover intentions.

Abbreviations

ORs

Odds Ratios

CI

Confidence Interval

PROSPERO

The International Prospective Register of Systematic Reviews

MOOSE

The Meta-analysis of Observational Studies in Epidemiology

PRISMA

The Preferred Reporting Items for Systematic reviews and Meta-Analyses

NOS

Newcastle–Ottawa Scale

Authors’ contributions

S. S. and T. T. conceived this review. S. S., H. C., F. Y., and Y. H. contributed to studies search and screening, data extraction and analysis, and draft writing. Y. Y. contributed to the methodology. H. C. and T. T. supervised and reviewed.

Funding

This work was supported by the National Natural Science Foundation of China [funding ID 72374157].

Data availability

All data generated or analysed during this study are included in this article.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.