Client-Perpetrated Violence Among Frontline Child Welfare Workers

Abstract

Client-perpetrated violence against child protective services (CPS) workers is common and detrimental to worker wellbeing with consequences including physical injury and psychological illnesses. Despite the impacts of violence, few studies capture violence in a representative sample of CPS workers. This study examines prevalence and applies a structural inequality framework to consider who among CPS workers is most susceptible. This article used the Florida Study of Professionals for Safe Families dataset, a longitudinal panel study of newly hired CPS workers (N = 1,501) to examine the prevalence of violence in the first 6 months of employment and consider the influence of individual characteristics. CPS workers experienced high levels of non-physical violence (75%), threats (37%), and physical violence (2.3%). Age and race were significantly related to instances of violence as were college major and position. Relationships were nuanced and did not uniformly follow a structural inequality framework. The high prevalence of violence within the first months on the job illustrates the importance of conflict and violence training before workers acquire independent caseloads. In addition, institutionalized mandatory reporting procedures with definitions of non-physical violence, threats, and physical violence may promote a culture of safety rather than an attitude that violence is part of the job. Finally, supervisors and peers can benefit from opportunities to support one other. The high prevalence of violence within certain subgroups of workers coupled with its wide-ranging consequences call for additional theoretical and empirical research and responsive policy to prioritize worker safety and wellbeing.

Keywords

workplace violence child protection workers client perpetration worker safety

Background

The prevalence of workplace violence among social service workers has increased and is a growing concern (e.g., Jayaratne, Croxton, & Mattison, 2004; Respass & Payne, 2008). According to Bureau of Labor Statistics data, social assistance workers were almost six times more likely than other workers to experience workplace violence requiring at least one day away from work (18.6 vs. 2.7 incidence rate; Respass & Payne, 2008). Moreover, despite the declining trend in workplace injuries across sectors, the incidence rate among social service occupations is rising (Bureau of Labor Statistics, 2013).

The Context of Child Protective Services (CPS) Work and Violence Susceptibility

CPS workers, a subset of social services workers responsible for investigating and providing services to families with alleged or verified reports of child maltreatment, are one vulnerable subgroup of social service workers (Jayaratne et al., 2004). They face 8 of 10 factors identified by the National Institute of Occupational Safety and Health (NIOSH) for workplace violence including delivery of services, home visits, work in high-crime areas, independent work, irregular hours, and hostile clients (NIOSH, 1996). To prepare them for this work, CPS workers possess bachelors’ degrees in a variety of fields predominantly in the arts and sciences; only a minority possess degrees in social work, a field traditionally associated with the work (Barth, Lloyd, Christ, Chapman, & Dickinson, 2008).

Although most CPS worker–client interactions do not result in violence, the context of the job heightens workers’ susceptibility to violence (e.g., McPhaul, Lipsomb, & Johnson, 2010). As part of their positions, CPS workers conduct home visits, interact with clients, and ask personal, intrusive questions regarding the home environment which may anger clients, particularly given that the interactions may result in recommendations to remove children from the home. Common characteristics of non-voluntary clients, the typical status of parents being investigated or monitored, also elevate the risk of violence (McPhaul et al., 2010). Families with alleged or verified maltreatment often present with poor judgment, low impulse control, mental health issues, and are prone to act violently, particularly when feeling threatened with temporary or permanent child removal (Shields & Kiser, 2003). In addition, CPS workers frequently work in dangerous or isolated locations, and in households where threatening animals, firearms, or drug paraphernalia are present (Kendra & George, 2001).

Workplace violence is detrimental to worker wellbeing with consequences including physical injury and psychological illnesses (Flannery, 1999; Robson, Cossar, & Quayle, 2014). Studies indicate that injuries are often substantial and frequently require medical attention (54% of instances; MacDonald & Sirotich, 2001) or time away from work (40% of instances; Snow, 1994), and, in rare instances, worker death (Scalera,1995). In addition, affected workers commonly describe psychological and emotional consequences such as fear and anxiety (Littlechild, 2005a, 2005b). Studies also suggest that violence creates agency costs including medical expenses for victims, lost productivity, lower service quality (Bowie, Fisher, & Cooper, 2012; Flannery, 1999), job turnover (Horwitz, 2006), and a less experienced workforce (Brady & Dickson, 1999; Jayaratne et al., 2004). Moreover, client-perpetrated violence may culminate in unsafe homes as hostile parents create a fearful atmosphere and may affect workers’ ability to conduct comprehensive assessments (Brandon et al., 2008).

Despite CPS workers’ high susceptibility (NIOSH, 1996) and Centers for Disease Control and Prevention’s (CDC) prioritization of identifying vulnerable workers and reducing occupational violence (Greenspan & Noonan, 2012), few studies have examined the prevalence of non-physical and physical workplace violence among CPS workers. Even fewer have considered who among workers are the most susceptible. Evidence from similar helping professions (e.g., nursing) found that exposure to workplace violence increased the likelihood of later exposure indicating the potential importance of individual characteristics (e.g., Grice, Gerberich, Alexander, McGovern, & Ryan, 2012). Measuring CPS violence and who it affects is fundamental to adapting workplace policies and practices to promote safe work environments and worker health.

Prevalence of Client-Perpetrated Violence Against CPS Workers

Notwithstanding the detrimental impacts of violence, few, if any, studies measure violence in representative samples of CPS workers. Limited convenience studies examining violence indicate high rates of verbal assaults (62%-97%) with less frequent threats of physical violence (11%-33%) or physical attacks (13%-34%; Horejsi, Garthwait, & Rolando, 1994; Ringstad, 2009). Assessing violent incidents in the previous year, Horejsi et al. (1994) found that among rural CPS workers in Montana (N = 166), 97% reported being screamed or cursed at, 33% received death threats, and 27% received threats of physical injury. In terms of imminent threat, 8% reported close calls with violence, including a bullet through the windshield, being chased, or struggling over a weapon (Horejsi et al., 1994). In a small convenience sample of CPS workers in one California county (N = 37), 62% of workers reported a psychological assault, 11% reported being threatened, and 19% reported a physical assault in the past year.

Given the limited nature of studies focused specifically on CPS workers, samples of social workers or social service workers that include subsamples of CPS workers offer more representative data. These studies also support the high prevalence of violence among this group (Jayaratne et al., 2004; MacDonald & Sirotich, 2001; Rey, 1996; Ringstad, 2009; Robson et al., 2014). In a subsample of CPS workers (n = 39) within a national sample of National Association of Social Workers members (NASW; N = 507), Jayaratne et al. (2004) measured lifetime prevalence of client-perpetrated violence and found that 68% of CPS workers reported a verbal assault, 39% reported a physical threat, and 2.6% reported a physical assault. Among the NASW sample, CPS workers experienced the highest rates of violence of any social work setting. In a convenience sample of social workers in England (N = 590) who were largely CPS workers (72%), 50% worked with hostile or intimidating parents at least weekly, 61% received a threat within the previous 6 months, and 18% reported a physical assault within the previous 6 months (Hunt, Goddard, Cooper, Littlechild, & Wild, 2016).

Theoretical Link Between the Violence Susceptibility and Individual Characteristics

Structural inequality and the importance of social location may contribute to understanding workers’ experience of violence. A framework of structural inequality and corresponding theories within it (e.g., social dominance theory, Sidanius & Pratto, 2004) recognize that categories of individuals are treated unequally based on status characteristics and differential treatment contributes to an individual’s social location. Determined by one’s race, gender, age, sexual orientation, or other characteristics for which society and its members classify the social world, social location influences one’s sense of self and, subsequently, how they approach relationships including behaviors and micro-level social interactions (Anderson, 2010; Sidanius & Pratto, 2004).

Social location is important to consider in violence victimization as it shapes workers’ attitudes, perceptions, stressors, and resources (Anderson, 2010). Structural inequality recognizes that institutions and processes outside of the individual influence one’s position, relationships, and reactions within a social structure. Women, for example, may be more at risk of violent incidents than men. Social location also may lead workers to classify incidents differently depending on context or culture (Anderson, 2010). One worker may classify a client’s statements as emotional and invested, while another may classify the same expression as verbal violence because of different cultural understandings of verbal expressions. However, CPS workers’ power and authority to make determinations on child removal or return may alter relationship dynamics such that social location operates differently. The underdeveloped theoretical literature examining the role of worker characteristics in the incidence of client-perpetrated violence justifies additional work in this area.

Empirical Links Between Violence Susceptibility and Individual Characteristics

In the only known study examining demographic characteristics and client-perpetrated violence among solely CPS workers, Ringstad (2009) analyzed a convenience sample of 37 workers in one California county and found no differences by age, years of experience, gender, race or ethnicity, or educational level. While these results suggested that demographic characteristics were not related to violence experiences, the author also noted that the small sample was a limitation and may have accounted for the lack of significant findings.

Larger studies of workers in social service professions, often including CPS workers, identify the importance of demographics (Robson et al., 2014). Studies indicate that younger social service workers were more susceptible to violence than their older counterparts (Bride, Choi, Olin, & Roman, 2015; Jayaratne et al., 2004; Jayaratne, Vinokur-Kaplan, Nagda, & Chess, 1996; Ringstad, 2009; Spencer & Munch, 2003). For example, among a national sample of NASW members (N = 1,129), workers younger than 45 years experienced higher incidence of physical threats, verbal abuse, and sexual violence than their older counterparts (Jayaratne et al., 2004). In terms of years of experience, Jayaratne et al. (1996) found that social workers with less experience received more threats in the past year, but not more verbal or physical assaults, compared to workers with more experience. In samples of nursing personnel and substance abuse counselors, however, those with more experience encountered higher rates of verbal assaults and physical threats in the prior year (Bride et al., 2015; Campbell et al., 2011). Others found that years of training were not related to client-perpetrated violence among social service workers (Newhill, 1996; Winstanley & Hales, 2008).

Findings are also mixed with regard to the role of gender in incidents of child-perpetrated violence. Although NIOSH and other researchers suggest women face particular risk of violence (Bent-Goodley, Butler, Pennefather, Perry, Scheyett, Whitaker & Webb, 2017; Knight, 1999), other studies find the opposite (Jayaratne et al., 2004; Jayaratne et al., 1996; Newhill, 1996). Sample composition and complex relationships likely account for much of the differences. For example, men may engage in riskier job tasks (Newhill, 1996) or be particularly vulnerable to certain types of violence (e.g., threats and physical assaults; Jayaratne et al., 2004).

Most available studies of social services workers indicate that client-perpetrated violence does not vary by race (Jayaratne et al., 2004; Newhill, 1996; Ringstad, 2009); however, most studies are relatively racially homogeneous or have small sample sizes. In larger samples of nurses and substance abuse counselors, Campbell et al. (2011) and Bride et al. (2015), respectively, found that White workers were more likely to experience client-perpetrated violence than workers of color. For example, in a sample of substance abuse counselors (N = 1,890), workers of color had 41% and 24% lower odds of psychological assault and physical threat, respectively, compared to White workers (Bride et al., 2015).

Relatively few studies examine how educational background contributes to client-perpetrated violence. Although no identified studies examined the role of education in violence among CPS workers, in a national sample of NASW members (n = 1,029), Ringstad (2005) reported no differences between social workers who had their masters’ degrees and those who did not. However, evidence from home health nursing suggests that nurses with more years of education have lower odds of both non-physical and physical violence compared to their counterparts with lower levels of education (Nachreiner et al., 2007).

Contribution of the Current Study

Studies largely using convenience samples suggest CPS workers face high rates of client-perpetrated workplace violence. Little research, however, has considered who among CPS workers is most susceptible, and available research from other professions is inconclusive. This study uses a statewide sample of newly hired CPS workers and applies a structural inequality framework to examine the following: (a) the prevalence of non-physical violence, threats, and physical violence in the first 6 months of employment and (b) how prevalence varies by race and ethnicity, gender, age, education, and experience. In addition, this study extends available literature by examining the characteristics of nativity and sexual orientation as potential factors in violence exposure. In light of mixed or few empirical findings, the structural inequality framework led to a central hypothesis: advantaged workers (e.g., White, older, male, native-born, more education, specialized training) will be less likely to experience violence than their less-advantaged peers (e.g., Black, Hispanic, younger, female, immigrant, less education, no specialized training). Through identifying workers most at risk of violence, findings can inform interventions designed to reduce violence among CPS workers, as there are currently no empirically tested approaches available. In addition, findings can inform the CDC’s public health model that identifies the importance of individual characteristics in reducing occupational violence (e.g., Greenspan & Noonan, 2012).

Methodology

Sample

The study uses the Florida Study of Professionals for Safe Families (FSPSF) dataset, a 5-year, longitudinal panel study of CPS workers (N = 1,500) designed to provide insight into CPS work experiences and subsequent employment decisions. All workers hired in Florida between September 2015 and December 2016 were eligible to participate. Participants, including those who leave their CPS positions, completed bi-annual online surveys to measure individual (e.g., demographic) and organizational (e.g., supervision, training) influences on worker experiences, wellbeing, and turnover. A full 84% of eligible workers participated in Wave 1, and 80% of Wave 1 respondents completed Wave 2 6 months later (n = 1,206; see Wilke, Radey, & Langenderfer-Magruder, 2017 for complete data description). The first 6 months of employment represented approximately 2 months of training and 4 months of casework with CPS clients. The sample includes individuals who completed both Waves 1 and 2. Due to minimal missing data on analytic variables (<4%, n = 44), listwise deletion resulted in 1,162 respondents in the analytic sample. The Institutional Review Board (IRB) at the authors’ university approved the study protocol.

Measures

Dependent variables

To measure client-perpetrated violence, workers completed the Workplace Violence Scale (McPhaul et al., 2010), a 9-item dichotomous index in which workers responded as to whether they had experienced the following forms of violence from a client or a member of a client’s household in the past 6 months: (a) yelled at, shouted at, or sworn at; (b) threatened by a client without physical contact; (c) threatened with property damage or theft; (d) threatened with a weapon; (e) personal or workplace property damage; (f) physical assault without physical injury; (g) physical assault with mild soreness or minor injury; (h) physical assault with pain, but no emergency room or physician visit; or (i) physical assault requiring emergency room or physician visit. The index did not include frequency of each type of violence. The index also did not include definitions of violence; rather, as is common in classifying actions as threatening, hostile, or offensive in the incidence of sexual abuse (e.g., Barth, Bermetz, Heim, Trelle, & Tonia, 2013), workers used their judgment in classifying incidents as violent or not. Workers’ subjective judgments of their experiences allowed for a range of behaviors and the consideration of context when classifying incidents as violent (MacDonald & Sirotich, 2001). From this categorical index of violence, the analysis includes a dichotomous variable (1 = yes, 0 = no) indicating whether or not workers experienced any type of violence, and three additional dichotomous variables indicating the experience of non-physical violence (i.e., yelled at, property damage), threats, or physical violence. Violence types are not mutually exclusive; workers who experienced more than one type of violence were coded “1” in the variable of experiencing any violence, and “1” in each of the violence types that they experienced.

Independent variables

Workers self-reported their demographic and individual characteristics. Age and work experience were measured in years. In terms of gender, respondents identified as male, female, or transgender. One person identified as transgender and was excluded from analyses. Workers selected a race from the U.S. Census categories and indicated whether or not they were of Hispanic origin. From these variables, the analysis includes one variable with four mutually exclusive categories: Hispanic, non-Hispanic Black, non-Hispanic White, and another non-Hispanic race. Additional demographic variables included nativity (U.S.-born or not) and sexual orientation (heterosexual/straight or not).

Educational background included four dichotomous measures. The analysis included whether or not the workers’ level of highest education was more than the position-required bachelor’s degree; whether or not the worker’s highest level of education was in social work; whether or not the worker had specialized practical training in child welfare (i.e., an internship or other exposure to practical child welfare training); and whether or not the worker had academic training in child welfare (i.e., child welfare elective course, child welfare certificate).

The analysis also controlled for job position. The sample included child protective investigators (CPIs) and case managers. CPIs determine child safety in instances of maltreatment allegations, and case managers provide services to children and families in situations when children are at risk of maltreatment. Evidence suggests that CPIs may be more vulnerable to violence because of responsibilities to assess child safety (NIOSH, 1996).

Analyses

Analyses consisted of three stages. First, the analysis examined demographic characteristics by whether workers experienced any type of client-perpetrated violence and by specific violence type using t-tests and chi-square tests as appropriate. Second, logistic regressions for any client-perpetrated violence and each violence type considered the contribution of each covariate net of the others. Multicollinearity prevented the inclusion of both worker age and number of years of work experience in regression analyses. Because age was more closely related to experiencing violence, models did not include years of work experience. Results provided odds ratios, where values less than one indicated a decrease in the likelihood of experiencing violence and values greater than one indicate an increase in the likelihood. They also provided predicted probabilities of violence or violence type for each variable significantly related to an outcome, holding other covariates at their means. The relatively few instances of physical violence (n = 27; 2%) created unstable estimates (e.g., confidence intervals for the odds ratios often in excess of 10). Therefore, results are not provided. Third, interactions between race and ethnicity, age, and gender tested if characteristics worked in combination to influence the odds of violence. All analyses were conducted in Stata 13.1.

Results

Descriptive Findings

Table 1 provides a description of newly hired workers and their exposure to client-perpetrated violence. Over three-fourths of workers (77%) experienced violence in their first 6 months of employment, most often non-physical violence (75%). Almost 75% of workers reported being yelled at, shouted at, or cursed at by a client or a client’s household member. Relatively few workers (4%) experienced personal or workplace property damage, which most commonly was accompanied with yelling, shouting, or cursing. Approximately 37% of workers experienced a threat, including threats without physical contact (36%), threats with property damage (8%), and threats with a weapon (2.5%, n = 29). Physical violence was relatively uncommon (2.3%, n = 27), however, not trivial considering workers only had approximately 4 months of potential exposure. Most incidents involving physical violence resulted in no injury or minor injuries, but four workers went to the doctor or emergency room due to their injuries.

Table 1.

Descriptive Means (SDs) and Percentages by Type of Violence: Florida Study of Professionals for Safe Families.

	Sample Distribution (n = 1,162)		Any Client-Perpetrated Violence		Verbal		Threat		Physical
Total			77.0		75.3		37.1		2.3
Age (in years)	31.6	(9.5)
Experienced violence (yes)			30.9***	(9.0)	30.9***	(9.1)	31.0^†	(9.2)	30.9	(7.8)
Experience violence type (no)			34.0	(10.6)	33.7	(10.5)	32.0	(9.7)	31.9	(9.6)
Mean years of work experience	8.9	(8.8)
Experienced violence type (yes)			8.4***	(8.4)	8.5**	(8.4)	8.7	(8.5)	7.9	(5.5)
Experience violence type (no)			10.5	(9.8)	10.3	(9.7)	9.0	(9.0)	8.9	(8.9)
Female (yes)	85.0		77.7		75.8		38.0		2.3
Race and Ethnicity
Hispanic	16.1		78.1*^a		75.9*^a		35.3		1.1
Non-Hispanic, Black	32.9		65.7***^b		63.9***^b		30.6***^b		3.4^†b
Non-Hispanic, White	47.0		84.5		82.8		41.8		1.7
Non-Hispanic, other race	4.0		78.7		78.7		42.6		6.4*^c
Born in United States (yes)	89.5		77.3		75.7		36.9		2.3
Heterosexual (yes)	91.9		76.7		75.0		37.0		2.4
Master’s degree or more (yes)	19.2		72.6^†		71.2		32.9		1.4
Social work degree (yes)	20.1		72.1*		70.4^†		36.1		0.4*
CW coursework (yes)	32.3		80.3^†		78.2		43.9**		2.9
CW practical training (yes)	14.7		84.8**		83.6**		40.4		1.2
CPI (yes)	42.1		80.8**		79.1**		38.9		1.4^†

Note. n = 1,052 due to additional missing data in the county variable. CW = child welfare; CPI = child protective investigator.

Hispanics-White difference.

Blacks-White difference.

Of another non-Hispanic race-White difference.

†

p < .10.*p < .05. **p < .01. ***p < .001.

In post hoc analyses examining the overlap among violence incidents, 53% of those who experienced violence only experienced one type of violence, 44% experienced two types, and 2% experienced all three types. Most workers (96%) who experienced threats or physical violence also experienced non-physical violence.

Several demographic characteristics predicted exposure to violence. Older workers and those with more work experience had lower rates of non-physical violence than younger workers and those with less work experience. Race and ethnicity also predicted all three types of violence. Black and Hispanic workers experienced lower rates of non-physical violence compared to White workers, and Black workers experienced lower rates of threats. Although relationships should be interpreted with caution due to the small number of physical violence incidents, workers of another non-Hispanic race were more likely to experience physical violence compared to White workers and the relationship approached significance for Black workers (p = .08). Other demographic characteristics, including gender, nativity, and sexual orientation, did not predict violence exposure.

In terms of education and training, workers’ highest degree level did not predict exposure to violence. However, workers whose highest degree was in social work were less likely to experience violence. Unexpectedly, workers with exposure to child welfare coursework or practical training were more likely to experience violence compared to workers without such training. Although not the focus of this analysis, in terms of the control variable, CPIs were more likely to experience non-physical violence than case managers. The reverse relationship with physical violence approached significance with case managers experiencing greater vulnerability (p = .08).

Logistic Regression Results

We conducted three separate logistic regressions distinguishing workers who experienced (a) any violence, (b) any non-physical violence, and (c) any threat from those who did not. Table 2 displays the results. The comparison of Models A and B indicates identical predictors reached statistical significance and in relatively the same magnitude. The similarities in the two models reflect that the vast majority of violence that workers experienced was non-physical violence and those who experienced threats and physical violence also experienced non-physical violence. No interactions between covariates reached statistical significance indicating that covariates operated independently of one another. Examining these models together, the significant predictors in the descriptive analyses were significant in the regressions. Older workers experienced lower odds of any violence and non-physical violence than younger workers. For each year, the odds of experiencing non-physical violence decreased by 3% (1 – 0.97). Predicted probabilities demonstrate how the age effect culminates in varying risk for older and younger workers. For example, at age 25, a worker’s probability for non-physical violence with all other model covariates at their means is 0.81 compared to a 0.65 probability for a worker at age 50.

Table 2.

Logistic Regression Models of Violence Measures: Florida Study of Professionals for Safe Families (n = 1,162).

	Client-perpetrated violence
	Model A: Experienced Any Type of Violence		Model B: Experienced Non-Physical Violence		Model C: Experienced Threats
	OR	SE	OR	SE	OR	SE
Race and ethnicity
Hispanic	0.63*	0.14	0.65*	0.14	0.72†	0.13
Non-Hispanic Black	0.32***	0.05	0.34***	0.06	0.59***	0.09
Non-Hispanic other race (reference: Non-Hispanic White)	0.64	0.25	0.74	0.29	1.00	0.31
Born in United States (yes)	0.82	0.20	0.92	0.21	0.79	0.17
Female (yes)	1.25	0.25	1.15	0.22	1.30	0.23
Heterosexual (yes)	0.95	0.27	0.98	0.27	1.03	0.23
Age (in years)	0.96***	0.01	0.97***	0.01	0.99	0.01
Have master’s degree or more (yes)	0.93	0.17	0.95	0.17	0.85	0.14
Social work degree (yes)	0.69**	0.11	0.61**	0.12	0.87	0.14
CW course work (yes)	1.35†	0.22	1.28	0.20	1.56**	0.21
CW clinical work (yes)	1.81*	0.45	1.86*	0.45	1.03	0.19
CPI (yes)	1.46*	0.22	1.44*	0.21	1.16	0.15
Constant	16.46***		11.85***		0.82
Pseudo R²	0.08		0.07		0.02

Note. CW = child welfare; CPI = child protective investigator.

†

p < .10.*p < .05. **p < .01. ***p < .001.

Race, educational background, and job position also predicted experience of client-perpetrated violence. Black and Hispanic workers had lower odds of experiencing any violence and non-physical violence compared to their White counterparts. Black and Hispanic workers had 66% and 35% lower odds of non-physical violence, respectively. In terms of predicted probabilities, net of other covariates, White workers had a 0.84 probability of non-physical violence, while Black and Hispanic workers had significantly lower probabilities at 0.64 and 0.77, respectively. Social workers had 39% lower odds of non-physical violence compared to workers whose highest degree was not in social work. Social workers had a 0.70 predicted probability of non-physical violence compared to a 0.79 probability for workers whose highest degree was not in social work. However, prior exposure to the child welfare system through internships or other opportunities increased odds of experiencing non-physical violence by 86% (adjusted predicted probabilities 0.85 vs 0.76).

Examining the incidence of threats in Model C, only race and ethnicity and child welfare course exposure significantly predicted the incidence of violence. Black workers had 41% lower odds of being threatened than White workers. Black workers had a 0.30 predicted probability net of covariates compared to a 0.42 probability for White workers. Although the Hispanic-White difference suggested lower odds of threats for Hispanic workers, the relationship only approached significance. In addition to main effects, race and ethnicity interacted with age such that, net of other covariates, age significantly decreased the probability of being threatened for Black workers, but not for White workers (see Figure 1). Although the predicted probability of being threatened also decreased for Hispanic workers and workers of another non-Hispanic race, these relationships were not statistically significant. The only other variable significantly related to being threatened was child welfare course completion. Similar to the increased vulnerability for non-physical violence among workers with practical experience in child welfare, those who completed child welfare courses had 56% higher odds of being threatened than those without a course. Workers who completed a child welfare course had a 0.44 adjusted probability of being threatened compared to a 0.33 probability for those without a course.

Figure 1.

Adjusted predictions of experiencing threats.

Discussion

This study examined client-perpetrated violence among a statewide sample of recently hired frontline child welfare workers. Consistent with prior studies using non-representative samples (Horejsi et al., 1994; Ringstad, 2009), CPS workers experienced high levels of non-physical violence (75%), threats (37%), and physical violence (2.3%) during their first few months with client caseloads. Because new workers often receive protected caseloads relative to both size and severity (Florida Department of Children and Families, 2016), the high prevalence may understate violent incidents among the larger CPS workforce. Although worker characteristics significantly predicted the incidence of violence, the relationships appear more nuanced than consistently aligning with a structural inequality framework in which disadvantaged groups face a greater risk. In some instances, advantage decreased likelihood of violence (e.g., being older, having social work course work) and in others disadvantage did so (e.g., being Black or Hispanic). In addition, more individual characteristics predicted non-physical violence than threats or physical violence, indicating the importance of differentiating types of violence when accessing risk.

Congruent with structural inequality framework and existing studies in other social service disciplines (Bride et al., 2015; Jayaratne et al., 2004; Jayaratne et al., 1996; Ringstad, 2005), being older and having specialized education (i.e., a social work degree) decreased the odds of experiencing non-physical violence. In addition to an independent effect of age, findings suggest that older age particularly protected Black workers from threats. However, counter to the structural inequality framework, being Black or Hispanic, when significant, decreased the odds, and exposure to child welfare content increased the odds of exposure to violence. Similarly, and in contrast to earlier findings (e.g., Jayaratne et al., 2004; Newhill, 1996; Ringstad, 2005) and structural inequality theory, gender did not alter the odds of exposure to violence. Neither nativity nor sexual orientation was significant.

The descriptive nature of the study along with limited theoretical and empirical work lead only to speculation of why these findings may be true. Both age and a social work degree may provide workers with experience, authority, and strategies to diffuse violence. However, exposure to child welfare courses or fieldwork did not provide the same protection and, in some instances, increased workers’ susceptibility. Although most CPS workers lack social work degrees, it may be that a full social work curriculum provides sustained opportunities for developing rapport-building, empathy, and conflict management skills that limited or no child welfare courses or fieldwork do not.

The impact of race/ethnicity and gender on susceptibility of violence also warrants further investigation. These findings counter the structural inequality framework. However, context and culture are important elements of structural inequality and may contribute to perceptions of violence. A higher likelihood of physical violence approached significance for Black workers and was significant for workers of another non-Hispanic race. Workers of color may have higher thresholds for classifying an incident as violent. For example, when presented with a set of materials depicting potentially violent incidents, White college students scored all episodes as more aggressive than Black students did (Phelps, Meara, Davis, & Patton, 1991). Alternatively, Black and Hispanic workers may experience less violence because their race allows them to develop rapport with clients as Black and Hispanic families are overrepresented among CPS families (Child Welfare Information Gateway, 2016). In terms of gender, the lack of significant findings may reflect counterbalancing forces of women’s greater risk of violence (Bent-Goodley at al., 2017) and men’s disproportionate assignment to potentially dangerous clients (e.g., Lowe, 2011). Future research, such as qualitative interviews with workers regarding the context of client behavior, could explore these possibilities.

Limitations

Findings should be interpreted in the context of the study’s limitations. First, the study measured perceived violence rather than actual violence. Although literature identifies the importance of subjectivity in classifying violence and considering its consequences (e.g., Barth et al., 2013), workers’ reports of violence may not match those of supervisors or peers. As mentioned above, two workers may define the same client situation differently; one perceives violence and the other does not. Second, data did not include client characteristics limiting the ability to consider the context of violent incidents. For example, were violent incidents more or less common when the worker–client duo was of the same race or ethnicity? This information could be useful in interpreting the mechanisms through which individual characteristics (e.g., age, race) influence exposure to violence. In addition, context surrounding violent incidents may provide insight into why a social work degree decreased violent incidents while child welfare course completion or field training increased them.

Implications

Within study limitations, findings advance knowledge about violence toward CPS workers and contribute to three main implications. First, the high prevalence of violence within the first months on the job illustrates the importance of violence training and preparation before workers acquire independent caseloads. Although CPS workers often complete training for 6 weeks or more prior to working with clients, many states have no mandated safety training (Milner & Hornsby, 2004). Trainings focused on judging the potential for client-perpetrated violence and deescalating violent situations could equip workers with skills to avoid or, when unavoidable, handle violent situations. The Department of Labor’s Occupational Safety and Health Administration (2016) establishes guidelines for preventing workplace violence among social service workers. The voluntary nature of the guidelines, however, means that workers only benefit when employers deem this an organizational priority. However, the high prevalence of client-perpetrated violence calls for training workers on such guidelines to decrease incidents and adequately respond when violence occurs.

Second, institutionalized mandatory reporting procedures with definitions of non-physical violence, threats, and physical violence may promote a culture of safety rather than an attitude that violence is part of the job. MacDonald and Sirotich (2001) found that many workers who experienced violence did not report it, considering it as part of the job. In addition, our finding that Black and Hispanic workers experienced a decreased likelihood of exposure to non-physical violence calls into question whether cultural perspectives on behaviors used in this definition of violence may contribute to disparities. Established criteria can guide workers and organizations when considering and responding to violent incidents.

Third, supervisors and peers can benefit from opportunities to support one other. Together with high CPS caseloads, the nature of the work creates a fast-paced, unpredictable, and stressful environment (e.g., DePanfilis & Zlotnik, 2008). The high prevalence of violence further indicates the need for workplace support. Supportive work environments in child protection contribute to increased worker wellbeing and retention (Boyas & Wind, 2010; Radey, Schelbe, & Spinelli, 2018). Workers, for example, could operate in teams when particularly susceptible to violence (e.g., home visits). Supervisor and peer support also provide an opportunity to process violent incidents and connect workers to necessary agency and outside resources to handle both psychological and physical consequences of violence. Observations of workers indicate that those who experience support following a stressful client situation such as violence, experience more positive future client interactions and report lower levels of anxiety than workers without such support (Rose, Jones, & Fletcher, 1998).

Future Directions

The present study, coupled with findings from smaller, less representative studies, indicates CPS workers are at high risk of client-perpetrated violence, particularly non-physical violence, with a substantial proportion experiencing threats. The amount of physical violence (n = 37; 2.3%), although occurring less frequently, is not inconsequential considering workers had, on average, four months of contact with clients, and newly-hired workers often receive smaller, less intense caseloads (Florida Department of Children and Families, 2016). Given workers’ vulnerability, findings related to the role of individual characteristics in the incidence of client-perpetrated violence introduce many questions that warrant future research. First, a structural inequality framework does not seem to explain the nuanced relationships found in this study. Instead, both advantage (e.g., being older; having a social work degree) and disadvantage (e.g., being Black, being Hispanic) decreased exposure to violence. The CPS environment may contribute to unique relationships and intersecting identities. The high prevalence of violence in the midst of power imbalance between workers and clients suggests the importance of a better understanding of the dynamics underlying CPS worker–client relationships.

Complementing theory advancement, future studies can benefit from examining details and consequences of violence. First, what is the nature of violent incidents that CPS workers experience? Qualitative studies, including interviews with workers who have experienced client-perpetrated violence, can provide insight to inform policy and increase worker safety. For example, varying definitions of violence, a range of rapport-building skills, or worker–client race compositions all offer potential explanations for racial violence disparities. Second, how does exposure to violence change over time? Longitudinal data provide an opportunity to assess the impact of workers’ exposure to violence over time. For example, tenure in child welfare may lead to more efficacious workers in preventing or responding to violence or workers may contextualize client behavior and no longer consider particular events violent. In addition, hierarchical modeling provides a mechanism to consider the roles of the county, agency, and supervisor in workers’ susceptibility to violence.

Conclusion

This study provides evidence from a statewide sample that client-perpetrated violence is a significant issue facing CPS workers. Over three-fourths of workers reported violence within the first few months of interacting with clients. The high prevalence of violence within certain subgroups of workers coupled with its wide-ranging consequences call for additional theoretical and empirical research and responsive policy to prioritize worker safety and wellbeing.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was supported in part by a grant from The Florida Institute of Child Welfare.

ORCID iD

Melissa Radey

Author Biographies

Melissa Radey, PhD, is an associate professor in the College of Social Work at Florida State University. Her work examines vulnerability and support among vulnerable populations, including low-income families as well as stakeholders across the child welfare system. She uses both quantitative and qualitative research to promote just policies and practices to enhance public and private safety nets.

Dina J. Wilke, PhD, is a professor in the College of Social Work at Florida State University. Her research examines professional training and workforce development, with a focus on early career patterns of social workers and other human service workers. She is the Principal Investigator for the Florida Study of Professionals for Safe Families, a longitudinal study of newly hired employees into child welfare workforce.

References

Anderson

K. L.

(2010). Conflict, power, and violence in families. Journal of Marriage and Family, 72, 726-742.

Barth

Bermetz

Heim

Trelle

Tonia

(2013). The current prevalence of child sexual abuse worldwide: A systematic review and meta-analysis. International Journal of Public Health, 58, 469-483.

Barth

R. P.

Lloyd

E. C.

Christ

S. L.

Chapman

M. V.

Dickinson

N. S.

(2008). Child welfare worker characteristics and job satisfaction: A national study. Social Work, 53, 199-209.

Bent-Goodley

Butler

Pennefather

Perry

Scheyett

Whitaker

Webb

(2017). Executive summary: Social workers’ perceptions of workforce challenges among women employed in social service settings. Washington, DC: National Association of Social Workers, National Committee on Women’s Issues.

Bowie

Fisher

Cooper

(2012). Workplace violence. New York, NY: Routledge.

Boyas

Wind

L. H.

(2010). Employment-based social capital, job stress, and employee burnout: A public child welfare employee structural model. Children and Youth Services Review, 32, 380-388.

Brady

Dickson

(1999). Violence in health care settings. In Leather

Brady

Lawrence

Beale

Cox

(Eds.), Work-related violence: Assessment and intervention (pp. 166-182). London, England: Routledge.

Brandon

Belderson

Warren

Howe

Gardner

Dodsworth

Black

(2008). Analysing child deaths and serious injury through abuse and neglect: What can we learn? A biennial analysis of serious case reviews 2003–2005 (DCSF-RR023). London, England: Department for Children Schools and Families.

Bride

B. E.

Choi

Y. J.

Olin

I. W.

Roman

P. M.

(2015). Patient violence towards counselors in substance use disorder treatment programs: Prevalence, predictors, and responses. Journal of Substance Abuse Treatment, 57, 9-17.

10.

Bureau of Labor Statistics. (2013). The 2013 survey of occupational injuries and illnesses. Retrieved from https://www.bls.gov/iif/oshwc/osh/os/osch0052.pdf

11.

Campbell

J. C.

Messing

J. T.

Kub

Agnew

Fitzgerald

Fowler

. . . Bolyard

(2011). Workplace violence: Prevalence and risk factors in the safe at work study. Journal of Occupational and Environmental Medicine, 53, 82-89.

12.

Child Welfare Information Gateway. (2016). Racial disproportionality and disparity in child welfare. Washington, DC: U.S. Department of Health and Human Services, Children’s Bureau.

13.

DePanfilis

Zlotnik

J. L.

(2008). Retention of front-line staff in child welfare: A systematic review of research. Children and Youth Services Review, 30, 995-1008.

14.

Department of Labor Occupational Safety and Health Administration. (2016). Guidelines for preventing workplace violence for health care social service workers, OSHA 3148-06R. Washington DC: Author.

15.

Flannery

R. B.

(1999). Violence in the workplace, 1970–1995: A review of the literature. Aggression and Violent Behavior, 1, 57-68. doi:10.1016/1359-1789(95)00005-4

16.

Florida Department of Children and Families. (2016). Child protective investigator and child protective investigator supervisor educational qualifications, turnover, and working conditions status report. Tallahassee: Author.

17.

Greenspan

A. I.

Noonan

R. K.

(2012). Twenty years of scientific progress in injury and violence research and the next public health frontier. Journal of Safety Research, 43, 249-255.

18.

Grice

M. M.

Gerberich

S. G.

Alexander

B. H.

McGovern

P. M.

Ryan

A. D.

(2012). Past violence and future work-related violence: A case-control study of Minnesota nurses. Italian Journal of Public Health, 7(2), 41-47.

19.

Horejsi

Garthwait

Rolando

(1994). A survey of threats and violence directed against child protection workers in a rural state. Child Welfare, 73, 173-179.

20.

Horwitz

M. J.

(2006). Work-related trauma effects in child protection workers. Journal of Social Service Research, 32(3), 1-18. doi:10.1300/J079v32n03_01

21.

Hunt

Goddard

Cooper

Littlechild

Wild

(2016). “If I feel like this, how does the child feel?” Child protection workers, supervision, management and organisational responses to parental violence. Journal of Social Work Practice, 30, 5-24.

22.

Jayaratne

Croxton

Mattison

(2004). A national survey of violence in the practice of social work. Families in Society, 85, 445-453.

23.

Jayaratne

Vinokur-Kaplan

Nagda

Chess

W. A.

(1996). A national study on violence and harassment of social workers by clients. Journal of Applied Social Sciences, 20, 1-14.

24.

Kendra

M. A.

George

V. D.

(2001). Defining risk in home visiting. Public Health Nursing, 18, 128-137.

25.

Knight

(1999). The implications of BSW students’ experiences with danger in the field practicum. The Journal of Baccalaureate Social Work, 4, 133-149.

26.

Littlechild

(2005a). The nature and effects of violence against child-protection social workers: Providing effective support. The British Journal of Social Work, 35, 387-401.

27.

Littlechild

(2005b). The stresses arising from violence, threats and aggression against child protection social workers. Journal of Social Work, 5, 61-82.

28.

Lowe

T. B.

(2011). Practitioners’ risk exposure to client violence: A test of gender-sensitive case assignment among supervisors. The Clinical Supervisor, 30, 19-35.

29.

MacDonald

Sirotich

(2001). Reporting client violence. Social Work, 46, 107-114.

30.

McPhaul

K. M.

Lipsomb

Johnson

(2010). Assessing risk for violence on home health visits. Home Healthcare Nurse, 28, 278-289.

31.

Milner

Hornsby

(2004). Training of child welfare staff and providers: Findings from the Child and Family Service Review. Protecting Children, 19(3), 4-14.

32.

Nachreiner

N. M.

Hansen

H. E.

Okano

Gerberich

S. G.

Ryan

A. D.

McGovern

P. M.

. . . Watt

G. D.

(2007). Difference in work-related violence by nurse license type. Journal of Professional Nursing, 23, 290-300.

33.

National Institute for Occupational Safety and Health. (1996). Violence in the workplace: Risk factors and prevention strategies. Current Intelligence Bulletin, 57, 96-100.

34.

Newhill

C. E.

(1996). Prevalence and risk factors for client violence toward social workers. Families in Society, 77, 488-495.

35.

Phelps

R. E.

Meara

N. M.

Davis

K. L.

Patton

M. J.

(1991). Blacks’ and Whites’ perceptions of verbal aggression. Journal of Counseling & Development, 69, 345-350.

36.

Radey

Schelbe

Spinelli

C. L.

(2018). Learning, negotiating, and surviving in child welfare: Social capitalization among recently hired workers. Journal of Public Child Welfare, 12, 79-98.

37.

Respass

Payne

B. K.

(2008). Social services workers and workplace violence. Journal of Aggression, Maltreatment, & Trauma, 16, 131-143. doi:10.1080/10926770801921287

38.

Rey

(1996). What social workers need to know about client violence. Families in Society, 77, 33-39.

39.

Ringstad

(2005). Conflict in the workplace: Social workers as victims and perpetrators. Social Work, 50, 305-313.

40.

Ringstad

(2009). CPS: Client violence and client victims. Child Welfare, 88, 127-144.

41.

Robson

Cossar

Quayle

(2014). The impact of work-related violence towards social workers in children and family services. British Journal of Social Work, 44, 924-936. doi:10.1092/bjsw/bcu015

42.

Rose

Jones

Fletcher

(1998). Investigating the relationship between stress and worker behaviour. Journal of Intellectual Disability Research, 42, 163-172.

43.

Scalera

N. R.

(1995). The critical need for specialized health and safety measures for child welfare workers. Child Welfare, 74, 337-350.

44.

Shields

Kiser

(2003). Violence and aggression directed toward human service workers: An exploratory study. Families in Society, 84, 13-20.

45.

Sidanius

Pratto

(2004). Social dominance theory: A new synthesis. In Jost

J. T.

Sidanius

(Eds.), Key Readings in Social Psychology. Political psychology: Key readings (pp. 315-332). New York, NY: Psychology Press.

46.

Snow

(1994). “Aggression: Just part of the job?” The psychological impact of aggression on child and youth workers. Journal of Child & Youth Care, 9, 11-29.

47.

Spencer

P. C.

Munch

(2003). Client violence toward social workers: The role of management in community mental health programs. Social Work, 48, 532-544.

48.

Wilke

D. J.

Radey

Langenderfer-Magruder

(2017). Recruitment and retention of child welfare workers in longitudinal research: Successful strategies from the Florida Study of Professionals for Safe Families. Children and Youth Services Review, 78, 122-128.

49.

Winstanley

Hales

(2008). Prevalence of aggression towards residential social workers: Do qualifications and experience make a difference? Child & Youth Care Forum, 37, 103-110.