Abstract
We investigate the relationship between the prevalence of workplace aggression and two key outcomes: employee engagement and organisational effectiveness. Drawing on social capital theory, we propose that the level of employee engagement within the organisation helps explain the association between workplace aggression and organisational effectiveness. We used secondary survey data and an important indicator of organisational effectiveness in the healthcare sector (i.e. rates of hospital-acquired infections (HAIs)) from 101 hospitals in NSW, Australia. We found that hospitals with higher rates of workplace aggression had higher rates of patients with HAIs and that employee engagement was an important mechanism that helped explain this effect. These findings underscore the potential benefits of management policies and practices that are aimed at preventing workplace aggression and support greater employee engagement.
JEL Classification:
1. Introduction
Workplace aggression (i.e. abuse, threats or assault towards staff in circumstances related to their work) is a significant problem for organisations worldwide (Spector et al., 2007; Stutzenberger and Fisher, 2014). Globally, the cost of workplace aggression to organisations is enormous, affecting core business and resulting in losses of millions of dollars each year (Chappell and Di Martino, 2006). Workplace aggression also has devastating effects on individuals, families and society as a whole. In Australia, the prevalence of workplace aggression is not well documented, as only systematic reports of homicide in the workplace are collected (WorkingSafe, 2016). However, a recent survey of healthcare workers in rural and remote Australia found that 57% experienced verbal abuse and 21% physical violence (Rural Doctors Association of Australia, 2012). In the United States, where incidents are recorded, workplace aggression resulted in 521 deaths and 570,000 nonfatal assaults in 1 year alone (Bureau of Justice Statistics, 2011). It is worth noting that these only represent the most serious incidents of workplace violence. More recently, in the United Kingdom, there were 569,000 reported incidents of violence at work (Health and Safety Executive (HSE), 2016). Less severe forms of violence, such as nonphysical aggression (e.g. yelling, swearing) are even more widespread (National Institute for Occupational Safety and Health (NIOSH), 1996).
In this article, we draw on the typology developed by Griffin and Lopez (2005) to define workplace aggression as ‘all forms of behavior by which individuals attempt to harm others at work’ (p. 990). This definition, originally proposed by Baron and Neuman (1998), includes covert aggression, overt aggression, verbal aggression, obstructionism and workplace violence. Workplace aggression can be differentiated from organisational aggression (where the target of the aggression is the organisation (e.g. damage to property or equipment; Hershcovis et al., 2007) or organisation-motivated aggression, which is in response to environmental factors such as a perceived injustice; O’Leary-Kelly et al., 1996).
Given the alarming statistics in Australia and worldwide, managing the issue of workplace aggression is a significant challenge for organisations. In order to foster a work environment free from harm or risk, organisations must implement appropriate policies and practices to deter and manage workplace aggression. Organisations that fail to do this will have employees who feel unsafe and disengaged at work, reducing the organisation’s social capital and engagement and, ultimately, the organisation’s overall effectiveness. For example, there may be higher levels of sick leave and turnover, as employees withdraw from the workplace temporarily or permanently. Levels of customer complaints may increase as levels of service decline. Although the organisational consequences of workplace aggression are often discussed, they are rarely studied empirically. To date, most of the research on workplace aggression has been on how it affects an individual’s health, well-being and job performance. At the extreme, the experience of workplace aggression can be life changing (Jenkins, 1996). But even when workplace aggression is less severe, it can lead to a range of negative outcomes for individuals, such as the fear of future aggression and decreased emotional and physical well-being (Bowling and Beehr, 2006; Cortina and Magley, 2003; Hershcovis and Barling, 2010; LeBlanc and Kelloway, 2002; Mueller and Tschan, 2011). Studies have also found that individuals affected by workplace aggression experience reduced affective commitment, increased intentions to leave (Barling et al., 2001), reduced individual job performance (Barling et al., 2001; Schat and Kelloway, 2000) and individual productivity (Budd et al., 1996). However, while we can assume that these individual consequences will collectively affect an organisation’s effectiveness (e.g. by increasing turnover), there have been almost no studies that have provided evidence that workplace aggression affects organisational-level consequences, such as an organisation’s overall effectiveness, or the process by which workplace aggression might affect organisational outcomes.
The dearth of studies on the organisational consequences of workplace aggression is surprising, given public concerns about this topic, not only for individuals’ safety, productivity and health, but also for society in general. In this study, we contribute to the literature by developing and testing a theoretical model linking the prevalence of workplace aggression, employee engagement (aggregated to the organisational level) and organisational effectiveness (measured as the rate of hospital-acquired infections (HAIs)) in a sample of 101 hospitals in NSW, Australia. Our focus on the organisational-level effects of workplace aggression is an important contribution to the literature because we cannot assume that workplace aggression affects organisations in the same way that it affects individuals (Schneider et al., 2003). In fact, there is growing evidence that organisational-level effects have incremental predictive value beyond individual-level effects (e.g. Kuenzi and Schminke, 2009) as the flow-on effects of individual experiences create a context which shapes behaviours and attitudes. A better understanding of how and why this happens is important for managers in order to prioritise and better manage the implementation of strong workplace policies and practices aimed at preventing and managing workplace aggression.
There is virtually no empirical research investigating mechanisms through which workplace aggression affects organisational effectiveness. This is most likely due to the scarcity of sufficiently large organisational-level datasets available to test mediating effects with sufficient statistical power (Harter et al., 2002). Most research on organisational effectiveness relies on just one or a few key informants at the top of the organisation to capture the potential drivers of effectiveness (e.g. Youndt and Snell, 2004) rather than experience of employees across the organisation.
In this article, we test the mediating role of engagement (defined as the combined cognitive and emotional attachment of employees to the organisation and the energy and behaviours directed towards achieving the organisation’s goals) and argue that the prevalence of workplace aggression in organisations leads to lower levels of employee engagement, which in turns leads to reduced organisational effectiveness. We draw on social capital theory (Bourdieu, 1986; Dess and Shaw, 2001; Leana and Van Buren, 1999; Nahapiet and Ghoshal, 1998) to argue that organisations consist of relationships and that the organisation’s employee engagement is the resource created through effective collective action. The prevalence of workplace aggression can disrupt or deplete the social fabric within an organisation. It can reduce its social capital and, as a consequence, the engagement within the organisation and, ultimately, its overall effectiveness.
Finally, our study addresses another significant challenge in the literature. We demonstrate an empirical link between employee experiences and both ‘soft’ outcomes (such as employee engagement) as well as ‘hard’ outcomes (such as objective organisational indicators of effectiveness). Specifically, we utilise rare organisational-level data from hospitals to link the prevalence of workplace aggression to an objective measure of organisational effectiveness pertinent to the context. The selection of our measure of organisational effectiveness was based on practical, theoretical and contextual grounds. Public hospitals in Australia are funded and regulated by state governments; therefore, typical measures of organisational effectiveness, such as profit, are not relevant or meaningful. In addition, it is not possible to compare patient outcomes such as mortality rates or recovery rates as these differ depending on factors that are unrelated to the internal environment of the hospital, such as whether the hospital is a tertiary referral hospital or a community-based mental health facility.
Therefore, in this study we use the incidence rate of HAIs as our measure of organisational effectiveness. HAIs are infections acquired by patients during a hospital stay that are unrelated to the patients’ medical condition. The purpose of hospitals is to treat patients and discharge them in better health than when they arrived. A hospital’s capacity to do so (i.e. preventing patients acquiring new infections) is an important indicator of the hospital’s effectiveness and a key performance indicator in healthcare. HAI is considered a highly reliable indicator of organisational effectiveness because the formula that is used to calculate the incidence per hospital takes into account the key differences between hospitals such as the number of patient admissions. It is less vulnerable to external factors such as competition and economic trends that might affect indicators such as profit and productivity (Harter et al., 2002) or accountancy practices used to produce profit results (Wall et al., 2004). While there is some evidence that external economic pressures may contribute to adverse medical events through work intensification in Australia (Duffield et al., 2011), these pressures are likely to be experienced equally in our study because all the hospitals in our sample are funded and regulated by the same institution (NSW Health). Our theoretical model is illustrated in Figure 1.
Theoretical model.
1.1 Effects of workplace aggression on organisational effectiveness through employee engagement: a social capital perspective
Organisational social capital is defined as a ‘resource reflecting the character of social relations within the organization’ (Leana and Van Buren, 1999: 538). Social capital can be distinguished from other types of capital such as physical capital (e.g. equipment, materials) and human capital (e.g. knowledge, expertise). But unlike these kinds of capital, it cannot be bought, sold or hired. Rather, it is created through the relationships within organisations. While some scholars narrowly define social capital as structural networks (e.g. Baker, 1990), others argue that resources are created through these relationships (e.g. Putnam, 1995). In this article, we adopt the latter approach and argue that an organisation’s employee engagement (i.e. the commitment, motivation and dedication directed towards the organisation by employees) is a resource created through social capital.
According to Leana and Van Buren (1999), there are two key ways that social capital is created in organisations. The first is through associability (also referred to as collectivity), that is, ‘the willingness and ability of participants in an organization to subordinate individual goals and associated actions to collective goals and actions’ (p. 541). These collective goals and actions become implicit norms that enable employees to coordinate their actions and create a desire or willingness to engage in collective action. The second way is through relational trust, a deep trust that can survive occasional breaches or deferred benefits. Relational trust is based on beliefs about the integrity and intentions of others in the organisation. Putnam (1993) argues that relational trust is embedded in norms and behaviours that are generalisable (i.e. they apply to everyone in the organisation) rather than personal (i.e. they apply to named individuals). In short, trust is transferred to others, regardless of specific interactions, and trust and associability both create and maintain social capital.
There is evidence of the importance of social capital as a predictor of group and organisational effectiveness. For example, Youndt and Snell (2004) found that organisations with bundles of human resources (HR) practices designed to enhance social capital have greater knowledge sharing and collective exchange and that these HR practices are stronger predictors of organisational performance than HR practices designed to enhance either human or physical capital. In addition, there is evidence that when social capital is undermined (e.g. through turnover), organisational performance declines (Shaw et al., 2005). This suggests that social capital creates an environment in which employees are more engaged, are more willing to act towards the ‘greater good’ and have a stronger affective attachment to the organisation. When organisational members are engaged, they are more likely to focus on the parts of their roles that are important and helpful to the organisation, thus achieving valued outcomes and decreasing self-interested behaviour (e.g. rushing a patient examination or not taking time to coordinate action with other healthcare professionals). These norms and positive relationships between employees and the organisation and between employees within the organisation are more powerful in creating engagement than the monitoring or rewarding of compliance with organisational rules that is needed in organisations that have low social capital.
The notion of employee engagement has received ample attention by both academics and practitioners, although the conceptualisation of employee engagement is often ambiguous and unclear (cf., Macey and Schneider, 2008; Purcell, 2014). The academic literature on engagement has seen a proliferation of different conceptualisations and operationalisations of the construct, which has somewhat impeded its theoretical development. While most early research has conceptualised engagement at the individual level, more recent research has increasingly focused on organisational employee engagement that goes beyond a purely employee-focused conception (e.g. Harter et al., 2002; Pugh and Dietz, 2008; Saks, 2006; Salanova et al., 2005). As a result, two dominant frameworks of engagement have emerged. The first, ‘work engagement’, is used to refer to one’s connection with the job, role or work activity (e.g. Schaufeli et al., 2002). The second, ‘employee engagement’ is used more broadly to refer to the connection with the workplace or organisation itself (Leiter and Bakker, 2010). The latter suggests that, at the organisational level, engagement reflects or emerges from an organisation’s social capital in that it captures the norms that have been created and reinforced through collective action and trust, which in turn shape the behaviours and actions that are focused on the common good.
In this article, we adopt this approach given our interest in engagement with an organisation created through the organisation’s social capital rather than the individual’s engagement with their work. In other words, we adopt Kahn’s (1990) original notion of engagement as reflecting employees’ cognitive and emotional attachment to the organisation as well as the energy and behaviours directed towards achieving the organisation’s goals or the common good, rather than self-interest.
Examining employee engagement at the organisational level as the potential mechanism that transmits the effects of workplace aggression is important not just theoretically but also for practical reasons. In many industries, senior management use engagement indices or metrics as organisational barometers of success for external stakeholders (e.g. shareholders), which are presented as precursors to future growth. Many annual reports include employee engagement scores alongside traditional measures of success. 1 In addition, management consultants frequently advocate organisational change to enhance engagement as the way to improve bottom-line outcomes (Macey and Schneider, 2008). However, little empirical evidence exists to support the notion that engagement at the organisation level is associated with an organisation’s effectiveness, or the conditions that undermine engagement. Thus, our study provides an important step in demonstrating much-needed empirical evidence for the link between employee engagement and organisational-level outcomes.
In this study, we suggest that an organisation’s engagement is a proximal outcome that is influenced by the prevalence of workplace aggression through the erosion of the organisation’s social capital. At the individual level, there is some empirical evidence that workplace risks and hazards, such as aggression, are related to lower levels of engagement. For example, meta-analytic results suggest that workplace ‘risks and hazards explained the largest amount of variance in engagement (64.8%) in the health care industry’ (Nahrgang et al., 2011: 13). At the organisational level, we also expect a negative association between the prevalence of workplace aggression and employee engagement. This is based on evidence from social capital theory (Bourdieu, 1986; Nahapiet and Ghoshal, 1998), which suggests that when trust is eroded, the social fabric that holds an organisation together becomes brittle and there is less collective willingness to subordinate individual goals and collective action towards an organisation’s goals.
Workplaces with high levels of aggression are often characterised as hostile or ‘toxic’ (Ghosh et al., 2011). Within these workplaces, employees feel unsafe and unsupported, the antithesis of a non-threatening, predictable and trusting environment that Kahn (1990) argues for as an important pre-condition to engagement. In such an environment, employees are likely to withhold effort and transfer energy and attention away from organisational goals to more immediate personal goals of safety and security (Blau, 1964; Kahn, 1990). In addition, employees who feel unsafe are less likely to try new things, are more likely to remain silent and are less likely to learn (Edmondson, 1999). Poor relationships between knowledge holders and recipients are also one of the biggest barriers to knowledge transfer and the spread of best practice (Szulanski, 2002). Indeed, evidence suggests that threat, fear of or exposure to others who experience violence is sufficient to produce these negative outcomes (Leather et al., 1998; Mueller and Tschan, 2011; Spector et al., 2007). We therefore predict:
Hypothesis 1. The prevalence of workplace aggression is negatively associated with employee engagement.
1.2 Organisational effectiveness and the mediating role of employee engagement
As discussed earlier, the prevalence of workplace aggression depletes an organisation’s social capital and engagement. We propose that employee engagement, manifested through social capital, helps our understanding of how workplace aggression affects an organisation’s effectiveness. When there is a greater prevalence of workplace aggression, employees are more likely to be alert to the possibility that aggression may occur. This awareness and possible fear of aggression has been shown to reduce job satisfaction (Driscoll et al., 1995), decrease well-being (Schat and Kelloway, 2003) and lead to less commitment to the organisation and dedication to the goals of the organisation (Van Emmerik et al., 2007). In addition, there is evidence that aggression affects individuals who are not directly targeted more strongly than those who are directly targeted (Houshmand et al., 2012). In the absence of a network of trust-based, collaborative relationships with colleagues, employees are likely to withdraw or conserve their energy, and their behaviours tend to be more focused on achieving personal rather than the organisation’s goals. Furthermore, employees will be more likely to detach cognitively and emotionally from the organisation. In other words, the depletion of social capital and employee engagement explains why workplace aggression reduces organisational effectiveness.
There is compelling evidence that, at the individual level, engaged employees have higher task performance (Christian et al., 2011; Rich et al., 2010), are more proactive and innovative (Hakanen et al., 2008; Salanova and Schaufeli, 2008) and are less likely to be absent or leave the organisation (Bakker et al., 2003). Although far less frequent, evidence of unit-level (e.g. business unit) engagement suggests that it is positively related to greater return on assets (Schneider et al., 2009), greater earnings-per-share growth, profitability, safety performance and customer loyalty (Harter et al., 2002; Salanova et al., 2005). However, evidence at an even higher level of analysis, that is, the organisational level, is very rare and does not provide a compelling picture about the extent to which engagement affects organisational effectiveness.
We argue that when employees feel safe from aggression, organisational effectiveness is improved, as such an environment promotes greater levels of relational trust, willingness and energy to focus on organisational goals. Where organisations do not have effective workplace policies and practices in place to either prevent or reduce the effects of aggression, the social fabric within the organisation deteriorates, and employees withdraw or withhold discretionary effort as their trust in the relational networks is challenged. The organisation’s engagement that is created through social capital therefore helps explain why the prevalence of aggression in the organisation reduces organisational effectiveness. Thus, we expect workplace aggression to be indirectly related to organisational effectiveness through employee engagement. We do not expect there to be a direct relationship between the prevalence of workplace aggression and organisational effectiveness as there are various mechanisms that transmit the impact that workplace aggression has on organisational effectiveness indicators. Even at the individual level, the experience of workplace aggression on individual performance has been found to be indirect, mediated through fear and negative mood (Barling et al., 2001; Neilsen and Einarsen, 2012). At the organisational level, we test one key mechanism, that is, the extent to which the aggression undermines the social capital and engagement with the organisation and how this in turn reduces the overall effectiveness of the organisation. Therefore, we predict:
Hypothesis 2. There is an indirect relationship between workplace aggression and organisational effectiveness through the organisation’s employee engagement.
2. Method
2.1 Study context
While workplace aggression can occur in any organisation, healthcare is among the most at risk, as aggression within hospitals often stems from patients, visitors, as well as co-workers (Bureau of Justice Statistics, 2011). Patient violence and aggression is often attributed to drugs, alcohol, anxiety or cognitive impairment (Hahn et al., 2010). But patients’ families may also be the source of violent or aggressive reactions triggered by anger, fear, delays or other precipitating events (Crilly et al., 2004; Gacki-Smith et al., 2009).
To test our hypotheses, we used secondary data from 101 Australian hospitals. Studying this phenomenon in hospitals has two advantages. First, healthcare workers are frequently exposed to violence (Fernandes et al., 1999). Almost 50% of Canadian nurses (n = 8000) from 210 hospitals reported experiencing a violent incident in the workplace in the previous week (Duncan et al., 2001). Consequently, dealing with violence and aggression is considered ‘part of the job’ for many healthcare workers (LeBlanc and Kelloway, 2002; O’Leary-Kelly et al., 1996). Second, hospitals are located in communities based primarily on their population density. These communities differ on socioeconomic indicators and community violence, and there is some evidence that violence within organisations is associated with violence within communities due to spillover effects (Dietz et al., 2003). This is particularly relevant for hospitals as community members make up the majority of patients and visitors. In this study, we control for community violence in order to investigate workplace aggression over and above community violence.
Secondary data from three sources within Australia were used to test our hypotheses: New South Wales (NSW) Health, the Australian Institute of Health and Welfare, and the NSW Bureau of Crime Statistics and Research. The use of secondary data has been advocated when the phenomenon under investigation is of a sensitive nature, or where the outcome of interest is only measured at the organisational level (Griffin and Lopez, 2005; Schneider et al., 2003; West et al., 2006).
In early 2011, NSW Health commissioned a workplace survey of all employees from public hospitals and healthcare facilities. Prior to the survey, 20 focus groups were held across the state to identify key workplace culture issues, followed by interviews and an iterative process with a range of NSW Health employees and senior executives to develop the final survey. One of the key concerns raised by employees in the focus groups was exposure to workplace aggression within hospitals. The survey results were publicly released on the NSW Health website, and the aggregated data for hospitals for the measures of workplace aggression and employee engagement were obtained from this site.
At the end of 2011, information for the previous year regarding cases of Staphylococcus aureus bacteraemia – an HAI – was released by the Australian Federal Government. Staphylococcus aureus bacteraemia is a serious bloodstream infection that is associated with hospital care and is often used as a key organisational effectiveness indicator for hospitals in Australia (Australian Institute of Health and Welfare (AIHW), 2011). These data are also publicly available for hospitals in NSW through the Australian Institute of Health and Welfare website.
2.2 Sample and procedure
The survey was sent electronically to all employees, and paper copies were distributed within hospitals for employees with limited computer access. Completion of the survey was voluntary and anonymous. Surveys took approximately 10 minutes to complete.
A total of 31,493 surveys were returned from employees across 192 hospitals and healthcare facilities. Reports, containing summary data, were generated for the 131 hospitals and healthcare facilities where at least 30 surveys were returned. These reports were made publicly available online (NSW Health, 2011) and are the source of our organisation-level data. Of the 131 hospitals and healthcare facilities, 101 could be matched to the 2011 HAI data. Hospitals without HAI data were either healthcare facilities without overnight patients, mental health facilities or support facilities. These organisations are not included in the analysis.
Hospitals included in the final sample had an average of 718 employees (standard deviation (SD) = 983, median = 318). The average response rate within hospitals was 43% (median = 41%). In total, 75% of the sample were female and 53% were full-time permanent employees. And 75% of the sample were clinicians, 68% were over the age of 45 and 78% had tenure of more than 5 years.
2.3 Measures
Workplace aggression was measured using two items based on Spector et al. (2007). The items were embedded in a section of the workplace survey titled ‘inappropriate behaviour’ in a larger section titled ‘the workplace environment’. Specifically, the two items were as follows: ‘In the last twelve months, have you personally experienced any incident where you have been abused or assaulted?’ and ‘In the last twelve months, have you personally experienced repeated behaviour which is offensive, intimidating, intended to humiliate or threaten?’ with the answer ‘Yes/No’. For each organisation, the measure represents the percentage of respondents who answered ‘yes’ to either one or both of these items.
Employee Engagement was conceptualised as the employee’s engagement with the organisation, aggregated to the organisational level. This measure has a particular focus on the behavioural/energetic and emotional/cognitive dimensions of engagement. The five items specifically refer to the workplace. Emotional/cognitive items included: ‘Overall I am proud to be a part of this workplace’; ‘I have a strong sense of belonging to my workplace’ and ‘Overall I am satisfied to be working here at the present time’. Behavioural/energetic items included: ‘I feel motivated to contribute more than what is normally required at work’ and ‘I would recommend my workplace as a good place to work’. Anchors for the items were 1 (strongly disagree) to 5 (strongly agree). Following recommendations by May et al. (2004), we use organisation’s average engagement score in the analysis.
HAI rates were obtained from the Australian Institute of Health and Welfare database (AIHW, 2011) in the year the survey data were collected. HAI data are only recorded and available at the organisational level and were thus matched to each hospital. HAIs are cases of Staphylococcus aureus bacteraemia, a serious bloodstream infection which hospitals aim to minimise to reduce complications, longer hospital stays and increased patient care costs. Serious infections may also result in death (AIHW, 2011). Rates of HAI are recorded as the number of patient infections per 10,000 patient days, with the national benchmark set at 2.0 cases per 10,000 patient days. This formula takes into account the key differences between hospitals, such as the number of patients admitted, and provides a standardised measure of effectiveness. Within our sample, the rate of HAI ranged from 0 to 4.17 cases per 10,000 patient days. Nationwide, there were 1873 cases of HAI reported for approximately 17 million days of patient care.
Control variables. Based on empirical and theoretical relationships among the study variables, we included three control variables in our analysis. In order to isolate the unique effects of hospital levels of aggression, we controlled for community violence and aggression. While our focus in this study is workplace aggression, our data do not identify the aggressor (i.e. insider or outsider). Most areas of hospitals are accessible to the public, and admissions to hospitals are influenced by community aggression and violence (Bellis et al., 2008) and healthcare professionals work with people (patients, relatives, friends) who are experiencing stress and anxiety. Therefore, the levels of violence in the community that the hospital services is a potential factor in levels of aggression experienced by staff in the hospital. We defined community violence as all incidents of violence- and aggression-related crimes (specifically domestic violence, non-domestic violence, sexual offences, harassment and threatening behaviour, offensive conduct, offensive language, alcohol-related offences, breaches of apprehended violence orders and resisting or hindering police officers) per 100,000 population. These data are recorded by NSW Police for each of the 152 Local Government Areas (LGAs; NSW Bureau of Crime Statistics and Research, 2012), the lowest tier of government in Australia.
In addition, we controlled for hospital size (i.e. the number of employees within each hospital) as there is evidence that individuals feel less valued in large organisations, where policies may be more formal and where there may be less individualisation and flexibility (Dekker and Barling, 1995). Following other organisation-level studies (e.g. Weare et al., 2007), we also controlled for the response rate for each hospital as this varies in our sample from 10% to 90%, which may be indicative of the underlying differences in variables such as engagement (Rogelberg and Stanton, 2007). It was not possible to identify whether a non-response bias exists in the data, and
in the absence of good information about presence, magnitude, and direction of non-response bias, ignoring the results of a study with a 10% response rate – particularly if the research question explores a new and previously unaddressed issue – […] is foolish. (Rogelberg and Stanton, 2007: 198)
Thus, controlling for hospital size and response rate within each hospital provides a more accurate test of our proposed model.
3. Results
3.1 Validity and descriptive statistics
Means, SDs and correlation coefficients for all variables are presented in Table 1.
Means, standard deviations and correlation coefficients.
N = 101.
Rate per 100,000 population.
Rate per 10,000 occupied hospital bed days.
p < .05 (2-tailed).
p < .01 (2-tailed).
Response rate and hospital size are negatively related to each other, suggesting that the larger the hospital, the lower the response rate (r = –.40, p < .01). The response rate is also negatively related to HAIs (r = –.44, p < .01), suggesting that hospitals with lower response rates are associated with higher patient infection rates. Hospital size is also negatively associated with employee engagement, with larger hospitals having lower engagement scores (r = –.23, p < .01). Community violence has a moderate negative association with hospital size (r = –.29, p < .01) and with response rate (r = –.37, p < .01), with smaller hospitals and hospitals with higher response rates being more likely to be located in communities with higher incidents of community violence. Community violence also has a small negative association with patient infection rates (r = –.21, p < .05). The associations between the control variables (i.e. hospital size, response rates and community violence) and the focal variables underscore the importance of controlling for these three variables in testing our hypotheses. The relationships among the study variables are all in the expected direction.
3.2 Results of hypothesis testing
We used SPSS 21 and the PROCESS macro (model 4; Hayes, 2013) to estimate the equations and test the indirect effects of violence and aggression on HAIs through employee engagement. The PROCESS macro uses bootstrapping, a nonparametric re-sampling procedure that does not impose the assumption of normality of the sampling distribution (Hayes, 2013), to obtain bias-corrected confidence intervals (using 5000 bootstrap samples).
Table 2 shows that the path coefficient (a) from workplace aggression (X) to employee engagement (M), controlling for community violence, hospital size and response rate (a = –.02, p < .001), is significant and negative, which provides support for Hypothesis 1 in that higher workplace aggression is associated with lower employee engagement. The path coefficient (b) from employee engagement (M) to organisational effectiveness (Y) is significant and negative (b = –.91, p = .014), controlling for violence and aggression (X), community violence, hospital size and response rate. There is also evidence of the indirect effect of workplace aggression on HAIs through engagement (ab; a point estimate, or effect size of .02 with 95% bias-corrected bootstrap confidence interval (.01, .04)), thus providing support for Hypothesis 2. That is, there is an indirect relationship between workplace aggression to organisational effectiveness (HAI rates) that is mediated or explained by employee engagement. This suggests that, for every percentage increase of staff reporting violence and aggression, there is an increased incidence of .02 HAIs per 10,000 occupied hospital bed days. Given that this is based on an average HAI rate of .72 per 10,000 occupied hospital bed days, it suggests an average HAI increase of 2.78% for every percentage increase in employees reporting workplace violence.
Mediation model examining the indirect effect of workplace aggression (X) in hospitals on organisational effectiveness (HAI) (Y) through employee engagement (M).
SE: standard error.
N = 101.
Significance levels were based on bootstrapping (5000) bias-corrected and accelerated confidence intervals.
p = < .001.
p = < .05 (2-tailed).
4. Discussion
Many incidents of workplace aggression are not officially reported (Bureau of Justice Statistics, 2011), and anonymous reports of aggression across many organisations are more likely to capture the phenomenon and enable the examination of the effects on organisational outcomes (Harter et al., 2002). The vast majority of research on workplace aggression has focused on the antecedents or consequences for individuals (Aquino and Thau, 2009; Bowling and Beehr, 2006; Hershcovis and Barling, 2010; Lim et al., 2008). Indeed, there is substantial evidence that an individual’s experience of workplace aggression is associated with negative outcomes (Barling et al., 2001; Budd et al., 1996; LaPierre et al., 2005; Schat and Kelloway, 2000).
In this study, we address a gap in the literature by demonstrating that organisations with a higher incidence of workplace aggression are less effective. The mechanism that helps explain this link is the organisation’s social capital or level of employee engagement. Employees who experience, anticipate or fear violence and aggression lose trust in relational networks and respond by disengaging physically and psychologically. In addition, by controlling for community violence in our analyses, we can rule out one competing explanation, namely, that the relationship between aggression within a hospital and the hospital’s effectiveness is due in part to a spillover from the local community to the hospital. This alternate explanation is based on evidence of increased violent events in hospitals close to nightclubs and alcohol outlets (Schofield and Denson, 2013). However, our research suggests that workplace aggression exists within hospitals irrespective of community violence and aggression, suggesting that it is a phenomenon inherent to the organisation and therefore within an organisation’s span of control.
Employees in organisations with a higher frequency of workplace aggression are more likely to perceive this as a signal that the organisation is not doing enough to prevent or manage violence (Hershcovis and Barling, 2010). The more widespread the experience of violence and aggression, the more likely it is that employees may perceive it to be a systemic failure on the part of the organisation rather than random acts. This general attribution of the cause of the violence to the organisation undermines the social capital which is reflected in lower employee engagement. A complementary perspective comes from the literature on perceived organisational support (POS). This literature provides evidence that organisations with higher levels of POS have higher engagement (Gavino et al., 2012; Rhoades and Eisenberger, 2002). Low POS may further exacerbate the relationship between workplace aggression and effectiveness by depleting the organisation’s social capital. In addition, there is evidence that proactive management and supervisory engagement in conflict resolution are positively associated with a range of organisational outcomes, suggesting an important role for supervisors in managing aggression (Roche and Teague, 2012). Future research could tease out the extent to which POS may act as a resource that moderates the relationship between workplace aggression and employee engagement and as a way of sustaining and enriching the social capital of the organisation.
As discussed earlier in the article, engagement has received a lot of attention by consultants and managers, partly driven by claims of a conclusive, compelling relationship between engagement and profitability (Purcell, 2014). Yet, most of the empirical evidence to date is at the individual rather than the organisational level (for an exception, see Harter et al., 2002). In this study, we contribute to the literature by focusing on engagement at the organisational level, which is of both theoretical and practical importance, given that it is ‘the level at which employee survey data are typically reported’ (Harter et al., 2002: 268). As Wefald and Downey (2009) state, ‘the gap between academia and practice […] with engagement must be resolved now, before the separate ideas become entrenched in the minds of both groups. Clearly, the two worlds need to meet somewhere in the middle’ (p. 145). We believe that our results provide evidence for the importance of clearly distinguishing between different levels of analysis in the study of engagement in order to improve concept clarity.
4.1 Practical implications
This study highlights the detrimental effects of exposure to workplace aggression, not just for an individual’s well-being and engagement with the organisation, but also for organisational effectiveness. Bowling and Beehr (2006) found meta-analytic evidence that work cultures that tolerate or fail to minimise aggression might encourage harassing behaviours and, ultimately, create a climate of harassment. Workplace aggression is associated with diminished social capital, such as engagement. These results suggest that organisations with management policies and practices targeted at preventing aggression, and addressing aggression when it occurs, would be more effective.
There is evidence that organisations that impose strong sanctions reduce levels of violence and aggression (Cortina and Magley, 2003; Dupré and Barling, 2006). A zero-tolerance approach to workplace aggression from both customers/clients/patients and staff may be an important step in reducing its prevalence and demonstrating an organisation’s commitment to supporting employees, thus increasing engagement and effectiveness. There is also evidence that training and development focused on preventing escalation of aggression, working as a team and resolving problem situations safely can reduce rates of moderate and severe aggression in the hospitality industry (Graham et al., 2004).
As mentioned earlier, engagement is a frequently touted panacea (Purcell, 2014) for organisational effectiveness, and yet there are relatively few organisational-level studies demonstrating that organisations with higher employee engagement have better outcomes. By providing evidence for this link, we provide some justification for investing in the development and maintenance of employee engagement.
Finally, within the healthcare sector, the incidence of HAI experienced by patients during a hospital stay is a critically important effectiveness indicator. HAIs are a major cause of illness and death and have serious economic costs for both patients and hospitals (Klein et al., 2007). It has been estimated that 477,927 hospitalisations in the United States in 2005 occurred due to HAI. The number of deaths was estimated to be greater than 90,000 per year, and the estimated direct cost to the healthcare system was US$6 billion (Klein et al., 2007). Much of the prevention focus in the medical literature is on clinicians’ knowledge and skills as well as technical aspects of patient care (Donabedian, 1989). While these are important, it neglects the context in which patient care takes place (Donabedian, 1989; West et al., 2006). Aggression is part of this context of patient care and influences relationships and trust among healthcare professionals and their patients and, ultimately, impacts patient safety. There is evidence that interpersonal factors and context, as reflected by HR systems and poor-quality work environments, affect patient mortality rates (Aiken et al., 2013; West et al., 2006), lead to adverse outcomes during a hospital stay and impact patient satisfaction ratings (Weinberg et al., 2013). So while HAIs are a useful indicator of organisational effectiveness, our research also points to the importance of managing workplace aggression within hospital settings for patient safety.
4.2 Limitations
The results of this research should be viewed in light of several limitations. We only had access to aggregated data at the organisational level rather than raw data, which means that it was not possible to establish the extent of the variability in employee engagement within organisations. In addition, it was not possible to establish the reliability or validity of the measures through traditional methods such as Cronbach’s alphas or confirmatory factor analyses (CFAs). However, our measures do align with measures from published research, and the scarcity of data from large samples of hospitals or other organisations that can be matched to objective outcomes means that our results serve as a valuable starting point and provide an important contribution to the literature. In addition, we based our arguments on social capital theory but do not measure it directly. Future research could explore how to measure social capital in organisations as there are no established organisational measures. One avenue to achieve this might be social network analysis, which would tap into this key mechanism more directly. However, there is a significant challenge of capturing this construct across multiple organisations.
Furthermore, we are not able to differentiate the source of aggression within our data, that is, whether it is initiated by outsiders (e.g. patients) or by insiders (co-workers/supervisors). We also cannot distinguish between more fine-grained types of aggression, such as physical and verbal aggression. The findings of previous research suggest that the effects of aggression from insiders may be more insidious than that from outsiders (Hershcovis and Barling, 2010), although evidence concerning physical and verbal aggression is less clear. Future research should use more specific measures of aggression that delineate different sources and types of violence in order to examine their relative impact.
Finally, HAIs are driven by multiple elements within hospitals, not just engagement. It is difficult to attribute outcomes to specific aspects of employee experiences or to specific elements of the system. In this study, while the outcome measure is weighted for some differences between hospitals, we have been unable to adjust for patient differences, selection effects and random variations (Shojania et al., 2001). In reality, data such as HAIs, even under a mandatory reporting regime, are somewhat dependent on the reporting body. So a hospital with less effective systems may miss cases of HAI that would be picked up within a more vigilant system. This is likely to mask the relationships between employee engagement and HAI rates and suggests that our findings are a conservative estimate of this relationship. In addition, we cannot rule out the possibility of reverse causality, particularly as workplace aggression and employee engagement were measured at the same time. For example, lower employee engagement or less effective organisations may be more vulnerable to violence and aggression. However, we provide strong theoretical arguments for the direction of the proposed relationships in this study.
5. Conclusion
In this study, we demonstrate that the level of workplace aggression within an organisation is associated with organisational effectiveness, and that this link can be explained by the effects of workplace aggression on the organisation’s level of engagement. While the links between violence, engagement and effectiveness have been demonstrated in the literature at the individual level, the same relationship has received far less attention at the organisational level, even though most indicators of organisational effectiveness are only available at this level. The paucity of research is partly due to the rarity of these data. Our study helps bridge this gap by analysing a rare example of data with individual-, organisational- and industry-level applicability to provide insights into how workplace aggression is detrimental to organisational effectiveness. It also suggests that investing in violence minimisation programmes and employee engagement is likely to have significant benefits for organisational effectiveness.
