Dellasega, Cheryl PhD, CRNP; Volpe, Rebecca L. PhD; Edmonson, Cole DNP, RN, FACHE, NEA-BC; Hopkins, Margaret MA, MEd
The national dialogue on components of the healthy work environment (WE) in healthcare has taken on new urgency. When a nurse leaves a unit because of an unhealthy WE, it is extremely costly with orientation of a new nurse costing approximately $60,000.1 In addition to financial costs, there are also patient-care concerns. Poor WEs have been shown to increase the odds of patient death and failure to rescue2 and are related to lower patient satisfaction.3
Administrators and managers have begun to consider strategies for promoting a more fulfilling and productive WE, but often feel poorly prepared to address conflict in the workplace.4 Relational aggression (RA) occurs when someone uses a relationship rather than physical means to inflict social harm. Relational aggression is sometimes known as female bullying, incivility, or “mean girls” and is more common among women than among men, especially during the formative adolescent years.5 Although the study of RA began with work in children, recently, authors have begun to examine this construct and its correlates in adults.6 Relational aggression among nurses is of special interest because nursing remains predominately a female profession.7
In general, 3 different roles are associated with RA.8 The aggressor is 1 or more persons in the position of launching an attack; victim(s) are the intended target. Bystanders see the aggression and do not intervene.8 It would be overly simplistic to characterize an individual as 1 of the above because in reality these roles are most often fluid, with yesterday’s victims becoming today’s aggressors. The largest group appears to be bystanders, who witness abuse—possibly encourage it—and fear that if they intervene, they will become the next target.8
Relational aggression is related to a number of similar social phenomena, including bullying, workplace incivility, lateral violence, and horizontal violence.9 Scholars sometimes use these words interchangeably, and a clear demarcation of language is sorely needed. For the present study, we understand RA to be broader than these concepts, because it includes bystander and victim roles in addition to that of aggressor.
Prior research has explored the frequency of workplace mistreatment in nurses. Some scholars find high rates of mistreatment: 1 study found that up to 25% of New York State nurses were often or frequently the victim of horizontal violence,10 and another study found that 38% of nurses witnessed horizontal hostility or bullying behaviors weekly or daily.11 Other research reported low levels of workplace mistreatment. In 1 of the 1st studies on workplace bullying in nurses, Johnson and Rea6 found that only 18% of their sample responded in the affirmative when asked if they had ever been bullied at work. A recent study found that new-graduate nurses experience bullying only every now and then (mean bullying score was 1.57 on a 5-point Likert scale).12 Finally, a study on workplace incivility found that nurses reported very little workplace incivility from their supervisors or their colleagues: the mean supervisor incivility score was 0.66, and the mean colleague incivility score was 0.81—both less than 1 on a 7-point scale.13
These mixed findings point toward a few different interpretations. It is possible that most nurses do not experience workplace mistreatment, but that it is a frequent experience for a small minority. Another possibility is that most nurses do experience workplace mistreatment, but only infrequently. Findings from Lewis and Malecha14 support this 2nd perspective: their study showed that 85% of nurses reported experiencing workplace incivility in the last 12 months.
Scales explicitly measuring RA have been put forth in the literature. The 1st RA instrument for adults was the Peer Assessment of RA and Social Adjustment Scale, which consists of 24 items, including one 7-item subscale designed to measure relationally aggressive behavior.15 This was followed by the Self-report of Aggression and Social Behavior, a 39-item measure that examines physical aggression, RA, relational victimization, and exclusivity.16,17 The Relational Aggression Assessment Scale (RAAS), which was the best available scale for the present study, does overlap in some ways with these other measures. However, in addition to being oriented toward a work setting and toward RA—and being gender neutral—the RAAS includes the “bystander” category, a common and understudied group on the RA spectrum and in work mistreatment in general.18
Because work mistreatment is associated with a number of adverse effects, including decreased nurse retention, productivity and job satisfaction, and increased burnout and emotional exhaustion, further study of the emotional WE is warranted.11-14 Relational aggression has been shown to be associated with adverse physical and mental outcomes in other populations,5,16 and anecdotal evidence suggests that nurses who experience RA suffer from problems such as depression, anxiety, somatic symptoms, hypertension, and other negative health consequences.19,20 Nurses who are involved in RA may leave their jobs prematurely, take extra sick days, and even deliver subpar patient care.21,22 Because job satisfaction and intent to remain on the job can be impacted by RA, it was the purpose of this study to explore these concepts in a large sample of nurses.
Purpose and Research Questions
When just 1 nurse is relationally aggressive, it can change the culture of the entire unit. Creating and maintaining a positive work climate are critically important in every work setting, but are perhaps especially important in nursing, where the stakes can be high, and the environment is high stress. For nurse executives, managing relationships in the workplace can be a challenge. The overarching purpose of this study was to explore RA as 1 component of the WE, with the long-term goal of designing an intervention to improve relationships in the workplace, which will, in turn, improve patient care.
We had 2 primary research questions: (1) What is the prevalence of aggressors, bystanders, and victims, in a sample of nurses employed at an academic medical center? And (2) is RA related to job satisfaction and intent to leave?
The total pool of potential participants included all physicians (n = 751) and nurses (n = 2096) employed at a large academic medical center. Only the nursing data are reported here.
Prior to initiating data collection, the investigators posted fliers and visited departmental meetings to promote the study. Recruitment was also enhanced by use of intranet Listservs and newsletters.
REDcap, a secure, Web-based application designed to support data capture for research studies, was used to collect data over a 1-week period. During this time, study participants were able to log onto the survey site and complete the questionnaires, which were piloted and took approximately 10 minutes to complete. A generic reminder e-mail was sent to all eligible participants halfway through data collection.
Data were next imported and analyzed using SPSS version 21.0 (Armonk, New York). Basic statistics, including frequencies and correlations, were calculated for each variable.
The self-administered survey questionnaire was composed of 4 sections:
* Relational Aggression Assessment Scale. The RAAS scale is best construed as 3 discrete scales that address the roles described previously: the RAAS-Aggressor scale (RAAS-A), the RAAS-Victim scale (RAAS-V), and the RAAS-Bystander scale (RAAS-B). The scale was developed by Neil Montgomery and Cheryl Dellasega and validated using a sample of young women in college. This prior research indicates that the RAAS is correlated with a range of personality tests that we would expect it to be correlated with, such as openness, conscientiousness, extraversion, agreeableness, and neuroticism—providing support for the validity of the scale.23 The RAAS-A, RAAS-V, and RAAS-B are composed of Likert-type questions on a 5-point scale and have 13, 9, and 11 items, respectively. Reverse coding was used on 2 items to avoid response set bias.
For the purpose of this study, we modified items as appropriate to reflect the workplace setting. The mean was calculated to compare scores, which were not mutually exclusive. Therefore, it was possible for a respondent to score high on items from each subscale, indicating use of aggressor, victim, and bystander behaviors.
* Job satisfaction. To evaluate job satisfaction, participants were asked to respond to the following prompt: “How satisfied are you, currently, with your job? Circle the number that best represents your satisfaction, where 1 is extremely dissatisfied and 10 is extremely satisfied.”
* Intent to leave was measured by a single item worded as follows: “Would you like to leave your current position?”24 Participants rated their response on a 5-point Likert scale from very interested (1) to not at all interested (5).
* Demographic profile included age, sex, and years at the institution.
The institutional review board approved the study as exempt research.
Of the n = 2,096 nurses who were invited to participate in the study, n = 842 responded (40%). Most respondents were female (94%; n = 787) between the age of 25 and 55 years (80%; n = 672) with bachelor’s degrees (61%; n = 506) who had worked at the institution for 4 years or more (67%; n = 578) (Table 1).
The reliability coefficients for the RAAS were good. The Cronbach’s α coefficient for RRAS-A is .83, that for RAAS-V is .83, and that for RAAS-B is .83, indicating good internal consistency reliability.
To calculate scores for each category—RAAS-A, RAAS-V, and RAAS-B—a mean was obtained. The mean for RAAS-A was 1.83, the median was 1.85, and the mode was 2.0. The range was 1 to 5, and the SD was 0.47.
The mean for RAAS-B was 1.71, the median was 1.73, and the mode was 2. The range was 1 to 5, and the SD was 0.50. The mean for RAAS-V was 2.29, the median was 2.22, and the mode was 2. The range was 1 to 5, and the SD was 0.64. For every RAAS-V item, at least 9% of respondents agreed or strongly agreed, but 3 items stood out. See Table 2 for the top 3 most agreed-with items in the RAAS-A, RAAS-V, and RAAS-B.
Spearman ρ indicated that RAAS-A, RAAS-V, and RAAS-B were significantly correlated. There was a moderate correlation between RAAS-V and RAAS-B (rs = 0.45, P < .001) and RAAS-A and RAAS-V (rs = 0.53, P < .001) and a strong correlation between RAAS-A and RAAS-B (rs = 0.74, P < .001).
The extent of RAAS-A, RAAS-V, and RAAS-B behaviors did not vary significantly by age, education level, years at the institution, whether the nurse engaged in direct patient care, or whether the nurse worked in an ICU setting. (The latter was examined to explore whether a highly specialized unit is more or less conducive to RA.) The RAAS-A and RAAS-B did vary based on sex: the mean RAAS-A score for women was 1.82, whereas the mean for men was 1.93; the mean RAAS-B score for women was 1.70, whereas for men it was 1.81. This difference did not reach statistical significance when evaluated via t test.
Respondents whose mean on the RAAS-A, RAAS-V, and RAAS-B was higher than 1 SD above the mean were evaluated. This group of respondents used relatively high levels of aggressive, victim, or bystander behaviors when compared with their colleagues. Of the total sample of nurses (n = 842), 14% (n = 115) scored high on the RAAS-A scale, 16% (n = 134) scored high on the RAAS-V scale, and 11% (n = 91) scored high on the RAAS-B scale. Years at the institution was positively correlated with scoring high on the RAAS-V (rs = 0.17, P < .05) and RAAS-B (rs = 0.22, P < .05); however, there was no significant correlation with age, education level, whether the nurse engaged in direct patient care, or whether the nurse worked in an ICU setting.
Job Satisfaction and Intent to Leave
The mean job satisfaction of nurses was 6.90, the median was 7, and the mode was 8 (n = 763). Eleven percent (11%; n = 81) of nurses rated their job satisfaction as a 4 or below, 23% of nurses (n = 175) rated their job satisfaction as a 5 or a 6, and 62% of nurses (n = 507) rated their job satisfaction as 7 or greater. The mean intent to leave for nurses was 3.36, the median was 4, and the mode was 5 (n = 840). Ten percent of respondents (10%, n = 85) were very interested in leaving, 22% (n = 187) were somewhat interested in leaving, 15% (n = 129) were not sure, 26% (n = 216) were mostly not interested in leaving, and 27% (n = 223) were not at all interested in leaving.
The RAAS was weakly, but significantly, correlated with job satisfaction and intent to leave. The RAAS-A was positively correlated with being interested in leaving the institution (rs = 0.26, P < .001) and inversely related to job satisfaction (rs = −0.28, P < .001). The RAAS-V was positively correlated with being interested in leaving the institution (rs = 0.24, P < .001) and inversely related to job satisfaction (rs = −0.25, P < .001). The RAAS-B was positively correlated with being interested in leaving the institution (rs = 0.13, P < .001) and inversely related to job satisfaction (rs = −0.13, P < .001).
Multiple linear regression analysis was used to develop a model for predicting nurses’ job satisfaction from their RAAS scores. Each of the predictor variables had a significant (P < .01) effect in the model. A model with only RAAS-V was a significant predictor of job satisfaction (R2 = 0.055, Finc(1,761) = 45.34, P < .001). In step 2, RAAS-A was added to the model (R2 = 0.062, Finc(1,760) = 7.83, P < .01) (R2 change = 0.010), and in step 3 RAAS-B was added to the model (R2 = 0.087, Finc(1,759) = 20.76, P < .001) (R2 change = 0.025) (n = 761). These results indicate that the variability in victims, aggressors, and bystanders explains 8.7% of the variability in job satisfaction.
Multiple linear regression analysis was used to develop a model for predicting nurses’ intent to leave the institution from their RAAS scores. Each of the predictor variables had a significant (P < .001) effect in the model. A model with only RAAS-A was a significant predictor of intent to leave (R2 = 0.062, Finc(1,838) = 56.24, P < .001). In step 2, RAAS-V was added to the model (R2 = 0.076, Finc(1,837) = 14.27, P < .01) (R2 change = 0.016), and in step 3, RAAS-B was added to the model (R2 = 0.093, Finc(1,836) = 15.92, P < .001) (R2 change = 0.017) (n = 838). These results indicate that the variability in victims, aggressors, and bystanders explains 9.3% of the variability in intent to leave the institution.
In each of the 3 RA scales—aggressor, bystander, and victim—the mean response was less than 3 (neutral) on a 5-point scale. Our findings indicate that overall nurses in this study did not use relationally aggressive, bystander, or victim behaviors frequently. In addition, of the 3 behavior sets, respondents were more likely to use victim-like behaviors than aggressor or bystander behaviors.
A recent study by Montgomery et al23 used the RAAS with 420 female students enrolled in 7 sections of Introductory Psychology at a northeastern college. Their mean RAAS-A score was 2.2 (vs 1.8 in the present study), their mean RAAS-B score was 2.2 (vs 1.7 in the present study), and their mean RAAS-V score was 2.5 (vs 2.3 in the present study). These results mirror our own in 2 important ways: 1st, they found that in every case the mean response is below 3; 2nd, they found that victim-like behaviors are the most readily admitted to.
Together, these findings point toward the possibility that most people do not frequently use relationally aggressive, bystander, or victim-like behaviors. Instead, it may well be that only a small minority of individuals frequently utilize these behaviors and that, as the saying goes, “one bad apple spoils the whole bunch.” Prior research supports this perspective: 1 recent study found that 39% of nurses never or very infrequently observed bullying behaviors, 19% observed it monthly, 23% weekly, and only 15% observed bullying daily.11 It is possible that the 15% who observe bullying daily have a “bad apple” in their WE, and that the rest of the group observes only the periodic flares that might be considered normal in a high-stress environment.
Other research has found that workplace mistreatment is a significant and frequent problem.25 These results clearly paint a very different picture than the present findings; however, the critical difference could be in the questions asked. Namely, we sought to quantify the frequency with which nurses use these behaviors; others have sought to identify whether nurses ever use bullying behaviors (eg, Lewis25 found that 85% of nurses had experienced workplace mistreatment in the last 12 months). Our findings support the idea that most nurses have witnessed or used bullying behaviors, but that very few nurses witness/use the behaviors frequently. In the present study, it was also of interest that on items that represented behaviors that negatively influenced the workplace, a sizable segment of nurses was “unsure” if they used the behavior or not.
Prior research indicates that a toxic WE can lead to job dissatisfaction and intent to leave,9,11,13 and our findings support this conclusion. We found that RA explained 9% of the variability in job satisfaction and 9% of the variability in intent to leave the institution.
The present research indicates that relying on bystander and victim-like behaviors may be related to length of time at the institution. These results are concerning inasmuch as they indicate that longer exposure to the institution results in increased use of less than ideal workplace behaviors. In addition, some authors posit that victimization and aggression are linked in a fundamental way, in that greater exposure to victimization may contribute to an individual’s vulnerability to becoming more aggressive; in short, over time, victims may retaliate with aggression.16 Our findings reporting a strong correlation between victims and aggressors supports this idea.
In conflict with the received wisdom, studies on younger populations and many other adult studies,10,15,26 we found no statistically significant difference in use of RA between male and female nurses. It would be of interest to compare employees in a profession where there is an equal distribution of men and women.
Other scholars have found that the experience of incivility is related to age, with younger nurses experiencing increased mistreatment.12 In particular, Leiter and colleagues27 found that “Generation X” nurses experienced greater coworker and supervisor incivility than “Baby Boomers”. Our results do not support these findings; we identified no relationship between age and use of RA behaviors.
Our study had several limitations, including the use of a new instrument not previously used with nurses. In addition, we evaluated nurses only at 1 institution. The setting of a teaching hospital may also contribute to bias, and the study should be replicated in a variety of care settings and institutions. Institutional and leadership culture also are possible limitations. Organizational policies such as zero tolerance guidelines may influence responses to research such as this.
Despite these limitations, we believe that we have established benchmark data for levels of relationally aggressive, victim, and bystander behaviors in nurses, which should be helpful to future researchers. In addition, despite the low overall level of aggressive, bystander, and victim-like behaviors, we did find a significant minority of nurses who frequently rely on these dysfunctional relational behaviors.
Implications for Nurse Executives
Creating and sustaining a healthy professional practice environment, free of RA and other negative behaviors, are foundational to the work of the nurse executive and nursing leadership. Achieving better healthcare and better health outcomes at a lower cost means remaining vigilant for cost reductions, productivity improvements, and quality improvements in the practice environment while working toward the vision suggested by the Institute of Medicine’s28 Future of Nursing report.
In conclusion, creating and maintaining a healthy WE are a monumentally difficult but significant imperative for leaders. Although we found low overall use of RA behaviors by nurses, our findings do point toward areas for potential improvement. In particular, 13% (n = 108) of nurses indicated that being the best was all important to them, and 9% (n = 75) of nurses said they would watch another coworker intimidate someone and not intervene.
Medicine is a team profession; every healthcare provider is a member of at least 1 team. Isolating, self-serving, and other negative behaviors by 1 or more individuals can diminish the function of the entire group.29 Scott Hutton30 argues that administrators need to act early to stem the tide of a bad WE: “Like a disease, early diagnosis of a toxic WE and intervention are key to mitigating organizational costs.” The pilot program described in the Document, Supplemental Digital Content 1 (http://links.lww.com/JONA/A300), offers promise as a step that institution leaders should consider to begin this process.
1. Greene MA. Paying for nursing orientation: a huge cost to hospitals. J Nurses Staff Dev. 2010; 26 (6): E3–E7.
2. Aiken LH, Cimiotti JP, Sloane DM, Smith HL, Flynn L, Neff DF. Effects of nurse staffing and nurse education on patient deaths in hospitals with different nurse work environments. J Nurs Adm. 2012; 42 (Suppl 10): S10–S16.
3. Kutney-Lee A, McHugh MD, Sloane DM, et al. Nursing: a key to patient satisfaction. Health Aff. 2009; 28 (4): w669–w677.
4. Dellasega C, Volpe R. Toxic Nursing: Managing Bullying, Bad Attitudes and Total Turmoil. Indianapolis, IN: Sigma Theta Tau International; 2013.
5. Dellasega C, Nixon C. Girl Wars: 12 Strategies That Will End Female Bullying. New York: Touchstone; 2013.
6. Johnson SL, Rea RE. Workplace bullying: concerns for nurse leaders. J Nurs Adm. 2009; 39 (2): 84–90.
8. Schafer M, Korn S, Brodbeck F, Wolke D, Schulz H. Bullying roles in changing contexts: The stability of victim and bully roles from primary to secondary school. Int J Behav Dev. 2005; 29 (4): 323–335.
9. Salin D. Ways of explaining workplace bullying: a review of enabling, motivating and precipitating structures and processes in the work environment. Hum Relat. 2003; 56 (10): 1213–1232.
10. Sellers KF, Millenbach L, Ward K, Scribani M. The degree of horizontal violence in RNs practicing in New York State. J Nurs Adm. 2012; 42 (10): 483–487.
11. Wilson BL, Diedrich A, Phelps CL, Choi M. Bullies at work: the impact of horizontal hostility in the hospital setting and intent to leave. J Nurs Adm. 2011; 41 (11): 453–458.
12. Read E, Laschinger HK. Correlates of new graduate nurses’ experiences of workplace mistreatment. J Nurs Adm. 2013; 43 (4): 221–228.
13. Spence Laschinger HK, Leiter M, Day A, Gilin D. Workplace empowerment, incivility, and burnout: Impact on staff nurse recruitment and retention outcomes. J Nurs Manag. 2009; 17 (3): 302–311.
14. Lewis PS, Malecha A. The impact of workplace incivility on the work environment, manager skill, and productivity. J Nurs Adm. 2011; 41 (1): 41–47.
15. Werner NE, Crick NR. Relational aggression and social-psychological adjustment in a college sample. J Abnorm Psychol. 1999; 108 (4): 615–623.
16. JM, Hart EJ, Kamper KE, Godleski SA. Relational aggression in women during emerging adulthood: a social process model. Behav Sci Law. 2011; 29 (5): 695–710.
17. Linder J, Crick N, Collins W. Relational aggression and victimization in young adults’ romantic relationships: associations with perceptions of parent, peer, and romantic relationship quality. Soc Dev. 2002; 11: 69–86.
18. Harnish L, Bradshaw C, Espelage D, Rodkin P, Swearer S, Horne A. Looking toward the future of bullying research: recommendations for research and funding priorities. J Sch Violence. 2013; 12 (3): 283–295.
19. Sheridan-Leos N. Understanding lateral violence in nursing. Clin J Oncol Nurs. 2008; 12 (3): 399–403.
20. McKenna BG, Smith NA, Poole SJ, Coverdale JH. Horizontal violence: experiences of registered nurses in their first year of practice. J Adv Nurs. 2003; 42 (1): 90–96.
21. Woelfle CY, McCaffrey R. Nurse on nurse. Nurs Forum. 2007; 42 (3): 123–131.
22. Sofield L, Salmond SW. Workplace violence. A focus on verbal abuse and intent to leave the organization. Orthop Nurs. 2003; 22 (4): 274–283.
23. Montgomery N, Dellasega C, Neilson N. Personality Profiles Underlying Relational Aggression and Relational Victimization in Women. Chicago, IL: Association for Psychological Science; 2012.
24. Ulrich C, O’Donnell P, Taylor C, Farrar A, Danis M, Grady C. Ethical climate, ethics stress, and the job satisfaction of nurses and social workers in the United States. Soc Sci Med. 2007; 65 (8): 1708–1719.
25. Lewis MA. Nurse bullying: organizational considerations in the maintenance and perpetration of health care bullying cultures. J Nurs Manag. 2006; 14 (1): 52–58.
26. Weber D, Robinson Kurpius S. The importance of self-beliefs on relational aggression of college students. J Interpers Violence. 2011; 26 (13): 2735–2743.
27. Leiter MP, Price SL, Spence Laschinger HK. Generational differences in distress, attitudes and incivility among nurses. J Nurs Manag. 2010; 18 (8): 970–980.
28. Institute of Medicine (US). Committee on the Robert Wood Johnson Foundation Initiative on the Future of Nursing. The Future of Nursing: Leading Change, Advancing Health. Washington, DC: National Academies Press; 2011.
29. Ramsay M. Conflict in the health care workplace. Proc Baylor Univ Med Center. 2001; 14 (2): 138–139.
30. Hutton SA. Workplace incivility: state of the science. J Nurs Adm. 2006; 36 (1): 22–27; discussion 27-28.