Although challenging behaviors have been identified as a source of staff stress, few researchers have directly addressed this relationship. In the present study, 55 teachers and support staff in special schools for children with mental retardation completed questionnaires assessing burnout, coping strategies for challenging behavior, and their exposure to challenging behavior. Results showed that (a) use of maladaptive coping strategies for challenging behaviors constitutes a risk for staff burnout, (b) this risk is in addition to that associated with exposure to challenging behavior, and (c) use of maladaptive coping strategies moderated the impact of exposure to challenging behaviors on emotional exhaustion burnout. Implications for future research and for the support of staff working with individuals who have challenging behaviors are discussed.
Editor in charge: Edwin Helmstetter
Challenging behaviors (e.g., self-injury, physical and verbal aggression, property destruction, and sexually inappropriate actions) displayed by people with mental retardation have been linked to staff stress in educational and other service settings. In the existing research literature on staff stress in mental retardation services, challenging behavior is consistently rated by staff members as either the most significant or one of the most significant sources of stress in their work (e.g., Bersani & Heifetz, 1985; Buckhalt, Marchetti, & Bearden, 1990; Corrigan, 1993; Hatton, Brown, Caine, & Emerson, 1995). Furthermore, staff perceptions of the demands associated with challenging behaviors have been shown to be significant predictors of stress and burnout (e.g., Dyer & Quine, 1998). These studies provide only indirect evidence of an association between challenging behavior and staff well-being because they do not involve measurement of challenging behaviors themselves.
Other researchers have tested the relationship between challenging behaviors and staff well-being more directly (i.e., by including some measurement of staff contact with individuals who have challenging behaviors). This has been achieved in two ways. First, a small number of researchers have compared staff members who work in challenging behavior services to those in other services or with normative samples. For example, Jenkins, Rose, and Lovell (1997) compared staff well-being in services where challenging behavior was reported by managers to be present and those in which it was not. Using this “expert” categorization, these authors reported that staff members working with individuals who had challenging behavior were found to have higher levels of anxiety, felt less supported practically, were less clear about identifying risk situations, and had lower job satisfaction. In another study using a group design, Chung, Corbett, and Cumella (1995) administered the Maslach Burnout Inventory (Maslach, Jackson, & Leiter, 1996) to 26 staff members working in four residential units for adults with challenging behavior. Their data showed that these staff members were significantly more emotionally exhausted than were those working in mental health settings described in the Maslach Burnout Inventory normative sample.
The second approach to directly addressing the relationship between challenging behavior and staff well-being has been focused on correlational designs. For example, Murray, Sinclair, Kidd, and Quigley (1999) explored the effects of service user aggressive behavior on 18 staff members in a residential service for adults with challenging behavior. Using records of assaults on staff members and records of their sick leave (workdays missed due to illness), Murray et al. found no significant associations. Exposure to aggressive challenging behavior within this specialist service did not predict staff sick leave.
At best, direct evidence of an association between support staff exposure to challenging behaviors and their psychological well-being is weak. There are at least two main difficulties with the methodology of existing research in which this question is addressed. First, the measurements of exposure to challenging behavior may not be valid in the sense that they do not reflect realistic staff exposure to challenging behaviors. In particular, individuals exhibiting challenging behaviors may not always be identified for receipt of specialized psychological treatment (Lowe et al., 1995). Therefore, challenging behaviors are likely to be present in many service settings, even if settings are not labeled as specializing in the care or education of people with challenging behaviors. Relying on formal records of aggressive assaults is also problematic in that it is unlikely that all such incidents are recorded. Furthermore, the negative impact of challenging behaviors on staff is not confined to aggressive assaults; staff members also report negative emotions associated with self-injury and other challenging behaviors (e.g., Bromley & Emerson, 1995; Dagnan, Trower, & Smith, 1998; Hastings, 1995).
The second methodological problem with existing research is that there has been no exploration of psychological factors that might affect the relationship between challenging behaviors and staff well-being. This issue is not only of theoretical significance but is likely to be crucial when developing interventions for staff under stress. In a recent study, Mitchell and Hastings (2001) suggested that coping, especially those strategies typically labeled as maladaptive (or “emotion focused”) as opposed to adaptive (or “problem focused”), may be an important variable. They asked 83 support staff members to complete a standard coping measure in terms of which strategies they used to cope with challenging behaviors in their work environment. Through factor analysis, three dimensions were identified: Disengagement (e.g., giving up the attempt to cope, substance use, and engaging in displacement activities), Denial (i.e., denying its significance and use of religious coping behaviors), and Adaptive Coping (e.g., planning, using support from others, and taking action to deal with the behaviors). The Maslach Burnout Inventory was also administered to staff members to assess three dimensions of burnout: Emotional Exhaustion (e.g., feeling emotionally drained), Depersonalization (e.g., treating people like impersonal objects), and (lack of) Personal Accomplishment (e.g., feeling that one has not accomplished worthwhile things in one's job). Mitchell and Hastings found that disengagement coping strategies were predictive of emotional exhaustion and personal accomplishment burnout, and adaptive coping strategies were predictive of personal accomplishment.
Although this recent research suggests that the way in which staff members cope with challenging behaviors is related to their reported stress, there was no measure of staff exposure to challenging behavior included in the analysis. Thus, there are still no published data on whether coping may affect the relationship between challenging behavior and staff well-being. Stress and coping theories suggest that coping would act to moderate the impact of a stressor (e.g., challenging behavior) on the well-being of support staff or family members (cf. Lazarus & Folkman, 1984; McCubbin & Patterson, 1983). Moderation in this discussion is being used in a methodological sense to refer to an interaction between two variables. For example, the relationship between challenging behavior and staff stress might be affected by the coping strategies that staff adopt. If they adopt one approach to coping, then they may be less stressed by challenging behaviors than if they adopt an alternative approach.
A recent study of parents of young children with autism lends some support to the theoretical prediction that coping will moderate the relationship between challenging behaviors and staff stress. Hastings and Johnson (2001) found evidence of an interaction between maladaptive family coping strategies (Passive Appraisal—minimization of response to problematic issues, such as “Believing if we wait long enough, the problem will go away”) and children's autistic behaviors in the prediction of parents' stress. Specifically, parents in families who were more likely to use passive appraisal coping reported higher levels of pessimism about their child's future but only when the child also displayed high levels of autistic behaviors. When the child had lower levels of autistic behaviors, passive appraisal coping did not affect parents' reports of pessimism about their child's future.
In the present study we addressed two main issues: (a) direct measurement of staff exposure to challenging behaviors and (b) exploration of the mechanism of action of coping on the relationship between challenging behavior and staff burnout. Based on previous research, we expected staff members' exposure to challenging behaviors and their use of maladaptive coping strategies to be positive predictors of burnout. Furthermore, we expected maladaptive coping strategies to moderate the impact of exposure to challenging behaviors on emotional exhaustion burnout. Specifically, we predicted that staff members with high levels of exposure to challenging behavior would report more feelings of emotional exhaustion if they adopted maladaptive coping strategies to deal with their exposure to challenging behavior. Predictions relating to depersonalization burnout are less clear because coping was not related to this dimension of burnout in the Mitchell and Hastings (2001) study.
Staff from three schools for children with developmental disabilities participated in the research. One school served 40 children with autism who were between 7 and 17 years of age, and two schools provided services to children with mild to moderate mental retardation (one served 50 children ages 4 to 11 years; and the other, 100 children ages 11 to 19 years). The school for children with autism was run by an autism charity. The majority of the children were low functioning, and one half of them boarded on a weekly basis at the school. The remaining children returned to their family home at the end of the school day. The two other schools were state-run day schools. All of the children at the three schools had disabilities, and the school day was designed to reflect as far as possible the experiences of children in English mainstream schools following a national curriculum.
Fifty-five staff members (14 males, 41 females) participated in the research. They had been working with children who had mental retardation and/or autism for a mean of 88.02 months (standard deviation [SD] = 67.96). Twenty-seven participants were special education teachers, and 28 were support staff members (i.e., teacher aides) with no special education qualifications. All of the staff members worked with children during the school day (i.e., they were not involved with residential provision at the school for children with autism). Although most staff members received training for at least 4 or 5 days during each school year, this training was rarely focused on understanding or treating challenging behaviors. Furthermore, there was no dedicated support from behavior consultants available to the schools at the time of the research. Thus, staff members had little support available to them in terms of addressing challenging behaviors.
Data were gathered using a self-report questionnaire containing four sections. In the first section, participants were asked for demographic information, their special education experience, and qualifications (see Participants section). The three remaining sections of the questionnaire contained measures of exposure to challenging behaviors, strategies staff members used to cope with challenging behaviors, and staff burnout.
Exposure to challenging behavior
Staff members were asked to report about how they had experienced aggressive challenging behaviors in their working environment during the preceding month in each of four domains developed for the present study. These domains were chosen to reflect the most common forms of challenging behaviors that may have a compounding effect. First, they were asked about aggression directed at them. This was accomplished by asking them whether or not they had been the target of physical aggression (defined using examples—kicking, biting, scratching, punching) resulting in injury to them (i.e., bruising, bleeding, or other tissue damage), physical aggression that did not result in injury, and verbal aggression (defined using examples—shouting or screaming, verbal abuse, threats). Second, staff members were asked whether or not they had witnessed the same three categories of behavior directed at other staff members or children. Third, they indicated whether or not they had witnessed self-injury (defined using examples—face-slapping, banging head against body or objects, scratching or biting self) resulting in tissue damage and self-injury not resulting in tissue damage. Finally, they were asked similarly about aggression toward objects (defined using examples—banging or kicking furniture or other property, pulling curtains, throwing objects) that either did or did not result in physical damage to property.
Four domain scores were used to produce a total Exposure to Challenging Behavior score. For the Physical Aggression domains, staff members used the following scores: 0 (experienced or witnessed no aggression), 1 (verbal aggression), 2 (physical aggression), and 3 (physical aggression with injury). In each case, the most severe behavior experienced/witnessed was scored. For the Self-Injury and Property Aggression domains, 0 was applied to staff with no experience; 1, when no damage or injury was experienced, and 2, when damage did occur. Thus, two scores ranging from 0 to 3, and two scores ranging from 0 to 2 constituted the exposure scale. Due to the fact that these items used different scales (0 to 2, or 0 to 3), scores on the four items were z-transformed. These standarized scores were then summed to produce a total exposure score for each participant. Using these z-transformed four items to produce a summed score, we obtained a Cronbach's alpha coefficient of .68, indicating an acceptable level of reliability. No other psychometric data were available on this scale.
Challenging behavior coping strategies
Staff strategies for coping with challenging behavior were measured using Carver's (1997) brief situational format of the COPE inventory (Carver, Scheier, & Weintraub, 1989). The COPE was developed by Carver and colleagues as a flexible multi-dimensional coping inventory for a broad range of applications in applied psychology. In the Brief COPE, 28 items are presented in the form of a coping statement, and respondents are asked to rate, using a fully anchored 4-point scale ranging from I haven't been doing this at all to I've been doing this a lot, whether they have or have not been using each way of coping. The Brief COPE has 14 subscales representing a broad range of coping strategies. The Brief COPE was chosen in preference to other coping questionnaires because it (a) encompassed a broad range of coping strategies; (b) can be presented in a situational rather than a trait format, allowing us to explore challenging behavior-focused coping; and (c) is shorter and, therefore, quicker to administer than the full version of the COPE.
Two scores were derived from the Brief COPE: (a) a summed score of ratings for coping strategies typically identified as adaptive (active coping, planning, positive reframing, acceptance, humor, religion, using emotional support, and using instrumental support) (e.g., Carver et al., 1989) and (b) a summed score of ratings of maladaptive coping strategies (self-distraction, denial, venting of emotions, substance use, behavioral disengagement, and self-blame). As the Brief COPE has not previously been used in published research with staff members who work in mental retardation services, we explored the reliability of the two dimensions of staff adaptive and maladaptive coping strategies. These two scale scores had good levels of reliability as assessed via Cronbach's alpha, indicating that they were internally consistent dimensions (adaptive coping, 16 items, α = .83; maladaptive coping, 12 items, α = .75).
The educator version of the Maslach Burnout Inventory was used as our measure of staff distress. It was chosen in preference to other measures of stress mainly because this inventory specifically addresses the feelings of staff working in education settings. Using 7-point fully anchored scales ranging from never to every day, staff rate 22 statements addressing their feelings about their job, which contribute to three subscale scores: Depersonalization (development of negative and cynical attitudes toward service users—e.g., “I treat some students as if they were impersonal objects”), Emotional Exhaustion (staff feeling that they have little left to give, at a psychological level, to their work—e.g., “I feel emotionally drained from my work”), and lack of Personal Accomplishment (staff evaluate themselves and their accomplishments negatively—e.g., not feeling that “I have accomplished many worthwhile things in this job”).
These three burnout dimensions in educators have been confirmed in factor analytic studies (Gold, 1984; Iwanicki & Schwab, 1981), and reliability data also indicate high levels of internal consistency for the three subscales (Maslach et al., 1996). A note on the interpretation of scores on the Maslach Burnout Inventory is worthwhile at this point. Burnout would be defined as the presence of high scores on the Emotional Exhaustion and Depersonalization subscales but as the presence of low scores on the Personal Accomplishment subscale. That is, burnout is related to a lack of feelings of personal accomplishment in one's work.
Special education staff were invited to participate in the research by their head teacher, who distributed questionnaires to all teachers and support staff in their school. Ninety questionnaires were given to the head teachers. The raw response rate of 61% was likely to have been affected by the fact that the questionnaires were completed and returned anonymously. Although head teachers were asked to issue a verbal reminder to their staff members, it was not possible to issue individually addressed reminders. However, the response rate compared favorably with other similar research (e.g., Mitchell & Hastings, 2001). No data are available on nonresponders, although the head teachers confirmed that the gender, age, and qualification (i.e., teacher vs. support staff) distributions represented the staff groups within their schools.
In order to generate some information about the suitability of the data for parametric statistical analysis, we used one-sample Kolmogorov-Smirnov tests to compare the distributions of Exposure, Coping, and Burnout scores to a normal distribution. These tests were also used to explore the distributions of the continuous demographic variables (staff age and length of experience in special education). The results of these tests were nonsignificant, indicating that the data were reasonably normally distributed, for all variables apart from the scores on depersonalization burnout. This variable was transformed by adding the value one to all scores and then using a logarithmic transformation. The value of one was added to each score before transformation due to the number of zero scores on this measure. Both the nontransformed and transformed variables were used in the statistical analyses reported later.
Associations between demographic variables and the three burnout scores were explored using independent samples t tests (for dichotomous variables—gender and special education qualification), and Pearson's correlation coefficients (for continuous demographic variables—age and length of special education experience). None of these demographic variables were associated with burnout scores at the .05 level and, therefore, were not included in subsequent analyses. Furthermore, staff members from each of the three schools did not differ in terms of their recent exposure to challenging behavior. Therefore, the data were analyzed for the whole sample together.
The main statistical analysis was performed using three linear regression models, one for each of the dimensions of burnout. In each case, five predictors were included in the regression model: Exposure to Challenging Behavior (total score), Adaptive Coping, Maladaptive Coping, and two interaction terms. The interaction terms were generated from the product of the z-transformed total scores on the exposure measure and the relevant coping scale (as recommended by Baron & Kenny, 1986). All of the independent variables were entered simultaneously into the regression models so that their independent contribution could be established. If the product interaction terms were found to be significant, then this would constitute evidence of their role as moderators of the impact of challenging behavior on staff well-being (cf. Baron & Kenny, 1986). The results of these analyses are displayed in Table 1.
Two of the regression models were significant, demonstrating that the predictors accounted for a significant proportion of the variance in burnout scores. For emotional exhaustion, R2 = .43, R = .66, F(5, 49) = 7.44, p < .001, exposure to challenging behavior and maladaptive coping strategies both had significant main effects, but the key finding is that the interaction between these two variables was also significant. In order to explore the nature of this interaction effect, we derived a data plot based on the guidelines developed by Aiken and West (1991). Figure 1 shows predicted values (i.e., derived from the regression equation) for emotional exhaustion at low (one SD below the mean) and high (one SD above the mean) exposure scores for each of three levels of maladaptive coping (one SD below the mean, at the mean value, and one SD above the mean). Figure 1 demonstrates that at high levels of exposure to challenging behavior, staff members who make more use of maladaptive coping strategies report higher levels of emotional exhaustion. At lower levels of exposure, the use of maladaptive coping does not appear to affect burnout.
The regression model for depersonalization burnout was also significant, R2 = .32, R = .57, F(5, 49) = 4.58, p < .01. Main effects for both exposure to challenging behavior and maladaptive coping were found to be significant, but there was no main effect of adaptive coping, and neither of the interaction terms were significant in this analysis. Identical results were obtained using the transformed depersonalization scores, although the proportion of explained variance and the standardized regression coefficients for the two significant predictors were slightly larger. The regression model for personal accomplishment burnout was not significant, R2 = .14, R = .39, F(5, 49) = 1.64. Therefore, the results in Table 1 showing a significant main effect of adaptive coping should be interpreted with extreme caution.
The results of the present study confirmed, through staff members' reports of their recent exposure to challenging behaviors, that staff well-being is associated with the challenging behaviors of people with developmental disabilities. Our data also confirm and add to a pattern of findings from previous research: (a) Use of maladaptive coping strategies to deal with challenging behaviors constitutes a risk in terms of staff burnout. (b) This risk is in addition to the risk associated with exposure to challenging behavior itself. (c) Challenging behaviors and associated psychological factors do not account for a very large proportion of the variance in personal accomplishment burnout. It is likely that other occupation or personality variables are more strongly related to feelings of personal accomplishment.
Finally, for the first time in the literature, in the present study we demonstrated a potential psychological mechanism explaining the link between challenging behaviors and staff well-being. More frequent use of maladaptive coping strategies was found to moderate the impact of exposure to challenging behaviors on emotional exhaustion burnout. Staff members with a relatively high exposure to challenging behavior who also used maladaptive coping strategies frequently, reported the highest levels of emotional exhaustion burnout. At low levels of exposure to challenging behavior, the frequency of use of maladaptive challenging behavior coping strategies appeared to carry little risk in terms of feelings of emotional exhaustion.
These results, and those of Mitchell and Hastings (2001), are consistent with research in other fields where the relationship between burnout and coping strategies has been explored. Other research on caregivers has also shown that their adoption of maladaptive coping strategies is associated with reports of burnout, whereas the use of adaptive coping strategies typically is not. For example, in a study of nurses in physical rehabilitation units, Elliott, Shewchuk, Hagglund, Rybarczyk, and Harkins (1996) found that emotion-focused coping was reliably associated with higher levels of burnout. Similar findings have been found in research on secondary school teachers (Chan & Hui, 1995) and professional mental health workers in a state psychiatric facility (Thornton, 1992).
The results of the present study clearly need to be replicated with more representative samples and with staff who work with people who have mental retardation in other contexts. Furthermore, there are three conceptual issues that should be considered when interpreting the results of this study. First, the assessment of coping in the present study was at a different level of measurement (i.e., toward challenging behavior specifically rather than the general stressors for staff in special education contexts) than that of the Maslach Burnout Inventory (i.e., general job-related feelings, not necessarily related to challenging behaviors). Although exposure to challenging behavior and frequency of use of maladaptive challenging behavior coping strategies did predict general feelings of burnout (see Table 1), the difference in level of measurement may explain the lack of association between adaptive coping and burnout. Future researchers could explore this issue by using well-being measures that also focus specifically on challenging behaviors, such as staff day-to-day emotional reactions (e.g., Mitchell & Hastings, 1998).
The second potential conceptual problem is that there is some overlap between items in the maladaptive coping scales (e.g., self-distraction, and especially behavioral disengagement) and items of the Maslach Burnout Inventory. This conceptual relationship seems to be closest for the Depersonalization subscale and its associated items (essentially a detached style of interaction with students). Maladaptive coping was found to be a predictor of depersonalization burnout and so the potential effects of this measurement overlap need to be considered carefully. We ran the regression for Depersonalization again, omitting the scores on self-distraction and behavioral disengagement coping items from the Maladaptive Coping subscale and generating a new interaction term. The pattern of results obtained was identical to that reported in Table 1 (i.e., significant main effects of exposure to challenging behavior and maladaptive coping and no other significant predictors). Therefore, although measurement overlap did not appear to affect the results of the present study, it is an issue of importance for further research.
The third conceptual issue for consideration here relates to the measure of staff exposure to challenging behaviors. Such a measure independent of their own reports would be ideal in future research. In particular, observational data on staff members and their interactions with service users would be useful. The present results may also have been affected by the fact that the severity of staff exposure to challenging behavior was assessed rather than their frequency of exposure. Perhaps continued exposure to low severity challenging behavior has an effect on staff well-being similar to that for infrequent but severe episodes of challenging behavior. Such questions should be explored in future research.
The general approach taken in the present study, of focusing on variables that may identify a psychological mechanism explaining the links between challenging behavior and staff well-being, is important theoretically and practically. At the theoretical level, there are likely to be other variables that moderate the effects of challenging behavior on staff well-being. For example, staff support resources may play a moderating role (i.e., poorly supported staff may be at high risk of stress in contexts where they are exposed to high levels of challenging behavior), and variables such as staff feelings of self-efficacy may play a mediating role. These and other possibilities need to be tested in future research in order to build a more comprehensive model of the interrelationships between staff cognitive and emotional responses and their exposure to challenging behaviors.
At the practical level there are at least two reasons to be interested in staff well-being and how it may be affected by challenging behavior. First, staff stress and burnout may be related to job-seeking behavior and turnover (e.g., Rose, 1995). This is also likely to have an indirect effect on the quality of services provided to people with developmental disabilities because of discontinuities of care or other support. Second, staff members' feelings of burnout may be related to their behavior towards children and adults with developmental disabilities in their care (e.g., Wisniewski & Gargiulo, 1997).
These negative effects of burnout for staff, services, and people with developmental disabilities could be reduced by developing interventions based on the findings of the present study and on future research. We can take the example of the present results in order to illustrate the significance of research addressing the mechanisms underlying the relationship between challenging behavior and staff well-being. First, research data could be used to build models that might help to identify subgroups of staff members who may be at particular risk. For example, the present data suggest that staff members exposed to high levels of challenging behavior and who adopt maladaptive strategies in order to cope with the situation may be at highest risk of burnout. The measures used in the present study show promise for identifying at risk staff as defined by these two variables. Identification of staff at risk is the first step in developing interventions that support staff.
Second, identification of intervening psychological variables, both moderators and mediators (see Baron & Kenny, 1986), often suggest an approach to psychological intervention for staff. For example, our results suggest that use of maladaptive coping strategies by staff was particularly problematic. Staff training or support interventions could be developed that reduce staff reliance on maladaptive strategies and encourage the use of adaptive strategies. Present data suggest that uptake of adaptive challenging behavior-focused coping strategies may not reduce staff feelings of burnout, but it is possible that such strategies are related to staff perceptions of successful coping in their work environment. These perceptions may have other positive benefits.
A final point worthy of note here is that issues relating to the psychological effects of challenging behaviors on staff may well be more dynamic than was possible to address in the present study. For example, it is likely that challenging behaviors not only impact upon staff psychological well-being, but that stress and burnout affect staff perceptions and emotional reactions to challenging behaviors. Furthermore, staff perceptions and reactions to challenging behaviors are likely to be affected by individual incidents or individual children. Researchers should address some of these issues, but the implication at present is that a more flexible intervention for staff may be required. In particular, the role of clinical supervision may be crucial as a mechanism for exploring perceptions and emotional and behavioral reactions to individual children and their challenging behaviors.
Authors: Richard P. Hastings, PhD ( email@example.com), Senior Lecturer, and Tony Brown, DClinPsych, Lecturer, Centre for Behavioural Research Analysis and Intervention in Developmental Disabilities, Department of Psychology, University of Southampton, Highfield, Southampton, SO17 1BJ, United Kingdom