Explaining Self-Reported Resilience in Child-Protection Social Work: The Role of Organisational Factors, Demographic Information and Job Characteristics

Explaining Self-Reported Resilience in Child-Protection Social Work: The Role of Organisational... Abstract Child-protection social work is a stressful occupation that results in workforce concerns about poor levels of staff retention and high levels of inexperience. This paper presents results from a cross-sectional survey and reports findings from a sample of 162 Northern Irish social workers. The sample were measured for ‘resilience’ (acceptance of self and life and individual competence, RS14 Resilience Scale), ‘burnout’ (emotional exhaustion EE, depersonalisation DP, personal accomplishment PA, Maslach Burnout Inventory) and organisational subscales (work-load, community, values, equity and control, Area of Work Life Scale (AWLS)). Pearson zero-order correlations showed that higher resilience was associated with lower EE and higher PA. Hierarchical linear regression analysis was used to identify unique demographic and work-specific predictors of resilience in addition to the AWLS subscales of control and values. The final model significantly accounted for 27 per cent of the variance in resilience scores, providing increased knowledge about resilience enhancing factors. As resilience is not an apolitical concept, the wider debates and politics of resilience are acknowledged. Specifically, contextual concerns are addressed that relate to the organisational factors that impact on social workers. The paper concludes by calling for organisational interventions to support resilience in social workers and maintain expertise in child-protection services. Burnout, child protection, child welfare, organisational factors, resilience Introduction Child-protection social work is acknowledged as a challenging and stressful career that is concerned with protecting children from abuse and neglect. Social workers are required to assess complex risk and make critical decisions that impact on the lives of children and families. This often is in the context of multiple and competing factors, contested information and stressful legal, moral and ethical challenges. Due to the unique nature of the role, retaining experience in child-protection teams has been a major issue due to high levels of staff turnover (Healy et al., 2007). Compassion fatigue and vicarious trauma have been reported in this profession, with concern about this risk increasing levels of burnout and job exit (Bride et al., 2007). In the current economic climate, with permanent job alternatives scarce, increase in use of agency staff and unfilled vacancies, retention of staff may be due to a lack of choice opposed to a desire to stay (Austin et al., 2017). This may impact negatively on not only the well-being of staff, but also the quality of the service they provide. Work-load is acknowledged in the literature as a contributing factor to burnout and staff turnover (Stalker et al., 2007; McFadden et al., 2014). Political and media pressure on the social work profession creates additional unease in front line workers that can result in a change in their practice due to the fear of ‘missing’ signs of dangerousness in families (Smith et al., 2003). The level of administrative tasks and over-bureaucratised systems that take staff away from critical face-to-face social work practice is acknowledged by the government-led Review of Child Protection (Munro, 2011) and in the wider literature (White et al., 2010). Recruiting and retaining experience in teams is critical to quality social work both to support junior staff and add to team stability and camaraderie (Strolin-Goltzman et al., 2008; McFadden et al., 2014). Burns (2011), in an Irish study on child protection, found that turnover was not a consistent issue and that some staff were ‘career converts’, originally entering the profession as a ‘stepping stone’ to something else and then remaining due to having found meaning and satisfaction in the role. This was reportedly related to good supervisory support and relationships that helped sustain individuals. US research on child welfare retention has also found that positive relationships with a supervisor and team can ‘buffer’ the impact of burnout and attrition (Mor Barak et al., 2001; Curry et al., 2005; Smith, 2005; Weaver et al., 2007). Self-reported resilience has also has been found to mediate the relationship between organisational factors and burnout (McFadden et al., 2018). The current paper drills down into this finding by accounting for age and gender as well as job-specific characteristics, which provides further understanding of what factors exist to support social workers to maintain feelings of resilience in this profession. The research question addressed in this paper is as follows: ‘What are the roles of organisational factors, demographic variables and job characteristics in child-protection social worker resilience?’ Theory and definitions Many definitions of resilience exist. From a range of perspectives, resilience can mean different things and the concept is contested and widely debated. Definitional perspectives of resilience have evolved and focus has moved progressively from individual characteristics to systems, context and social justice concerns (Ungar, 2013; Hart, 2016). Resilience is viewed as ‘doing better than expected’ following adversity using a range of protective processes available (Rutter, 1987). When applied to child-protection social work, the relevance of ‘resilience’ has an important application and meaning. However, the organisational environment is the context that must be considered to avoid assumptions about resilient ‘individuals’. Intra-personal references to resilience are criticised for being pathogenic, ignoring contextual, environmental, political and economic factors (Garrett, 2016). Whilst self-reported feelings of ‘resilience’ are factors selected for the current study, this is balanced with the application of organisational factors. These include work-load, social support, equity, values and fairness, to ensure that pathologic explanations do not dominate the findings. Inter-personal dimensions of resilience include references to social resilience, which is multidimensional, and refers to groups and communities coming together to advance their interests in the face of adverse challenges (Hall and Lamont, 2011). When one applies a systems model such as the social ecological perspective (Bronfenbrenner, 1979), it is clear that resilience is not dependent on individual characteristics, but relies on an interactive dynamic between the individual and their environment (Masten and Wright, 2010) and risk and protective factors (Rutter, 1987). Masten and Wright (2010) suggest that systems of meaning making, inter-personal relationships, agency and mastery, emotional and behavioural regulation, problem solving and cultural and religious traditions are related to resilience. Another aspect of resilience research seeks to further understand the developmental biological basis of resilience in interaction with a life-course trajectory that includes environmental factors, showing that resilience is not a ‘fixed’ state, but ‘plastic’ and context, time and experience-specific (Belsky and Pluess, 2009). Wagnild and Young (1993) propose a definition that applies typical individualistic factors to resilient people, such as personal strength, emotional stamina and self-confidence. The tool will be discussed later in the methodology section. In the context of occupational resilience, the concept of ‘job engagement’ is applied in terms of how ‘job-engaged’ (Leiter and Maslach, 2006) or ‘committed’ (Ellett, 2009) workers are despite the pressures and stresses of the job. The current paper does not conceptualise resilience as a personality trait (Werner and Smith, 1982), but as an interactive phenomenon influenced by the person and their environment and an ongoing adaptive capacity and process (Bronfenbrenner, 1979; Rutter, 1987). Therefore, organisational structures and features that provide the context, culture and climate for social work activity are an essential consideration in this debate. Job engagement and burnout run along a line of continuum that is captured by extensive research (Leiter and Maslach, 2005). Burnout is characterised by emotional exhaustion, depersonalisation and low personal accomplishment, whilst job-engaged individuals report opposite characteristics, indicating energy, empathy and efficacy (Maslach and Leiter, 2008). The MBI (Maslach Burnout Inventory) and the AWLS (Area of Work Life Scale) measure this continuum in relation to three interrelated concepts looking at the organisational correlates of work-load, fairness, community, control, values and reward (Maslach and Leiter, 2008). These measures are applied in the current paper. The concept of job burnout became popular in the 1970s and was viewed as ‘A characterization of adverse reactions to work, primarily in the human services settings’ (Freudenberger, 1974, p. 159). Schaufeli et al. (2009, p. 204) argue that burnout is deeply rooted in the broad social, economic and cultural developments that have been part of the occupational landscape for more than thirty-five years. They argue that the shift from being an industrial society to a service-led economy underpins the major occupational transformation. With these many changes, there is a level of social transformation that manifests in psychological pressures on people in transition and change. Not all perspectives on burnout take a ‘political’ stance on the issue but, instead, the risk of burnout can be seen as a ‘cost risk’ with a focus on retention and turnover, absence and sick leave, retraining costs and low experience (Curtis et al., 2010). Curtis et al. (2010) refer to an average eight-year lifespan of social workers compared to the longevity in other professions to demonstrate this point. References to ‘resilience’ in contexts of job-related adversity are criticised for the pathological assumptions made about individuals needing to ‘adapt’ or willingness to ‘change’ having major political underpinnings. Evans and Reid (2013) argue that this imposes a pathological and individualistic responsibility on people that ignores wider contextual and structural concerns, which is politically ‘catastrophic’. When applied to social work, this critique resonates. As social workers self-select into the profession, there is a risk of this perspective being applied to those who struggle with the challenges of the job, with the emphasis on individuals being called on to ‘adapt’ or ‘change’ to manage ‘threats now presupposed as endemic’ (Evans and Reid, 2013, p. 83). This is especially true when a body of research has exposed the specific challenges in relation to contextual concerns of high work-load, bureaucracy, varying levels of supervisory support, burnout and inexperience in teams (Stalker et al., 2007; Healy et al., 2007; Munro, 2011; McFadden et al., 2014). Aims and objectives The overarching aim of this study was to examine the factors that contribute to feelings of resilience and burnout in the child-protection workforce in Northern Ireland. The objectives were to identify the protective factors that sustain resilience in some staff and also to identify and measure the impact of organisational factors and how these relate to experiences of burnout and resilience in this occupational group. Organisational contexts are examined from the perspective of job engagement or burnout to measure the perceived impact of work-load, values, community, equity, control and reward. Demographic variables are also measured. Methodology This study used a cross-sectional survey design using standardised tests (MBI, RS14 and AWLS). Stratified random sampling was applied, as the population of social workers had natural groupings into a number of strata (gateway, family intervention/support, voluntary sector) and further divided into demographic variables such as age, gender, length of experience, job status, employer, caring responsibilities, religion, post qualifications, qualifications and rural or urban caseload. The survey was distributed using Smart Survey (c) software and was delivered by e-mail to social workers who had confidential status to maximise participation. Ethical approval This study was approved by the Office of Research Ethics in Northern Ireland (August 2010). The main ethical issue was the need to consider management of any identification of poor or concerning practice associated with burnout, with an agreed protocol to manage this potential. Although the protocol did not have to be implemented, the process of managing any concerns relating to practice was helpful and also required critical evaluation of a range of considerations such as unprofessional practice or concerning attitudes relating to service recipients. All participants gave informed consent to participate in the study and participant information literature provided all required information, including reference to this concern. Sampling frame The study used stratified random sampling across six participatory organisations: five Health and Social Care Trusts and those employed within a voluntary-sector organisation Child Protection Team. Two statutory structures for front line child-protection staff were the target sample: ‘gateway’ and ‘family intervention’ (intermittently referred to as ‘family support’). Gateway is the structure for all referrals relating to child-protection concerns and the point of entry into social work system. Family support/intervention accept referrals from gateway staff in cases that require long-term family support and child-protection services. This method of sampling increases the ability to claim representativeness; however, the authors use this claim cautiously due to the sample size drawn from each strata (Table 1). Table 1 Variations in response rates across participating organisations HSC Trust  Potential participants  Actual participants  % response rate  A  90  37  41  B  72  33  46  C  76  28  39  D  59  22  37  E  64  24  38  Voluntary sector  19  18  95  Total  380  162  43  HSC Trust  Potential participants  Actual participants  % response rate  A  90  37  41  B  72  33  46  C  76  28  39  D  59  22  37  E  64  24  38  Voluntary sector  19  18  95  Total  380  162  43  Table 1 Variations in response rates across participating organisations HSC Trust  Potential participants  Actual participants  % response rate  A  90  37  41  B  72  33  46  C  76  28  39  D  59  22  37  E  64  24  38  Voluntary sector  19  18  95  Total  380  162  43  HSC Trust  Potential participants  Actual participants  % response rate  A  90  37  41  B  72  33  46  C  76  28  39  D  59  22  37  E  64  24  38  Voluntary sector  19  18  95  Total  380  162  43  The sampling frame for the study included those in their Assessed Year in Employment (AYE—this is the first year following graduation, which ensures the employee has reached an acceptable standard to proceed to fully qualified social worker status). Also included were front line social work posts, those in senior practitioner posts and team leaders. The emphasis was on those who held caseloads and professional accountability for risk and protection of vulnerable children and families. The overall population of social workers who met the inclusion criteria was 380. Sample characteristics The gender of those who took part in the study was 86 per cent female (n = 140) and 14 per cent male (n = 22). This ratio was consistent with the gender profile of social workers in Northern Ireland as a whole and is evidenced in the gender ratio of social work more widely (McFadden, 2013). The contract status levels were reflected across the sample, ranging from those in their AYE through to senior practitioner and those ‘acting’ in team-leader posts. The most frequently participating subjects were those who had permanent contracts at a rate of 67 per cent (n = 109). Senior practitioners represent the second most frequently occurring participation rate at 9 per cent (n = 14). Those in their AYE (8 per cent, n = 13) and agency staff (6 per cent, n = 10) were the least frequently participating groups in the study. Participants’ ages ranged from twenty-two to fifty-eight years old. The mean age of participants was thirty-six years old (SD = 9.31). Research tools All research tools in the study are validated measures. The Resilience Scale RS14 (Wagnild and Young, 1993) has fourteen items measuring self-reported feelings of resilience across five broad factor areas of self-reliance, meaning, equanimity, perseverance and existential aloneness. This was the outcome variable measured in the context of predictor variables that will be outlined below. Factor analysis indicated that the Resilience Scale has two major factors, namely ‘acceptance of self and life’ and ‘individual competence’ (Wagnild and Young, 1993). Perseverance is the act of persistence despite adversity or lack of encouragement. It also means the there is a determination to remain actively involved in the management of adverse life events, being able to redefine oneself and being self-disciplined. Equanimity is the presence of a balanced view of life and experiences, and the ability to form an overview of a broader range of experiences and feel ready for unforeseen challenges, which moderates the potential for an extreme response to adverse events. Meaningfulness is the ability to understand that life has a purpose that one has a role in with ‘meaning’, providing the individual with something to live for. Self-reliance is the ability to have self-belief about one’s role in life and the abilities and capacities present or absent in oneself. Existential aloneness was a particularly useful concept in the measurement of resilience in child-protection social workers. This, generally applied, is the understanding of our aloneness in the world and the uniqueness of one’s life and experiences and individual responsibilities in one’s lifetime (Wagnild, 2009, p. 23). Resilience items measure statements such as ‘I usually manage one way or another’, ‘I usually take things in my stride’, ‘I feel I have accomplished things in life’, ‘I feel I can handle many things at a time’ and ‘In an emergency I am someone people can rely upon’. Ideally, the internal consistency of a scale should be above 0.7 (DeVellis, 2003). Internal consistency reliability (Cronbach’s a coefficient) was 0.90 in the current sample. The MBI (Maslach and Jackson, 1986) measures self-perception of experienced emotional exhaustion, depersonalisation and personal accomplishment. A series of twenty-two questions measures these three dimensions and subjects are asked to report the frequency of experienced feelings and thoughts using a Likert scale ranging from (0) ‘never’ to (6) ‘every day’. This scale measures the individual’s experience on the continuum between burnout and job engagement. Three interrelated dimensions are exhaustion energy, cynicism involvement, and inefficacy and efficacy (Maslach and Leiter, 2008). The exhaustion component is linked to the psychological strain associated with individual experience of burnout. Cynicism is the inter-personal context that refers to an emotional detachment or negative feelings towards service users and other aspects of the job. The efficacy dimension is related to the self-evaluation dimension of burnout and is experienced when an individual has feelings of incompetence or a lack of achievement in relation to their job. Internal consistencies (Cronbach’s a) for the MBI subscales were acceptable—emotional exhaustion (0.90), depersonalisation (0.75) and personal accomplishment (0.78). The AWLS has twenty-nine items and measured organisational factors such as work-load, community, control, fairness, equity and values. The items are worded in terms of perceived congruence or incongruence between self and workplace. Items reflect positively framed statements of congruence, such as ‘I have enough time to do what’s important in my job’, as well as negatively framed items of incongruence, such as ‘Working here forces me to compromise my values’. Subjects indicate their agreement or disagreement with these statements using a five-point Likert scale, from (1) ‘strongly disagree’ to (5) ‘strongly agree’ (Maslach and Leiter, 2008). Scoring for the negatively framed items is reversed. This tool defines congruence as a high score (greater than 3.00), meaning a high degree of ‘fit’ between the person and workplace. Incongruence between the person and the workplace is the reverse. As the score is low (less than 3.00), this suggests a ‘misfit’ between the person and workplace. Maslach and Leiter (2008) report that this misfit is related to burnout. Maslach and Leiter (2008) developed a model of burnout that focused on the perceived congruency between the worker and important aspects of the organisational environment. According to this model, a higher level of incongruence is directly related to an increased risk in burnout. Conversely, the higher degree of job–person congruity is correlated with a higher chance of job engagement. Internal consistency coefficients for the AWLS subscales were above 0.7—work-load (0.84), reward (0.91), community (0.87), fairness (0.77) and values (0.74)—with the exception of ‘control’ (0.68). The demographics measured include: team and sector, status of the post, length of experience, employer, gender, age, marital status and caring responsibilities, religious background, post qualifications, qualifications and rural–urban caseload. Data analysis The data were analysed using SPSS 23. Ordinary least-squares (OLS) hierarchical regression analysis was used to examine the ability of the variables to predict resilience scores (RS14). Response rate The number of child-protection social workers who participated in the survey was 162 out of a possible target sample of 380 social workers. The collective response rate across all six participating organisations was 43 per cent (n = 162). This was a regional study in Northern Ireland that had one voluntary-sector child-protection organisation and all five Health and Social Care Trusts included in the sample. Correlates of resilience Table 2 illustrates the inter-correlations among resilience, MBI and AWLS subscales. This table demonstrates the low to moderate correlations between each of the AWLS and the Resilience RS14 scale. This meets the OLS regression assumption of predictor variables not being too highly correlated for regression analysis. Pallant (2007, p. 132) refers researchers to Cohen’s (1988, p. 79) suggested ranges for interpreting Pearson correlation coefficients as follows: small = 0.10–0.29, medium = 0.30–0.49 and large = 0.50–1.00. Table 2 Pearson correlation matrix of resilience, MBI and AWLS subscales   2  3  4  5  6  7  8  9  10    Exhaustion  Depersonal  Personal accomp.  Workload  Control  Reward  Community  Fairness  Values  1 Resilience  –0.37**  –0.10  0.52**  0.21**  0.22**  0.13  0.12  0.07  0.22**  2. Exhaustion    0.38**  –0.33**  –0.62**  –0.39**  –0.35**  –0.07*  –0.24**  –0.37**  3. Depersonalisation      –0.25**  –0.26**  –0.19*  –0.12  –0.08  –0.09  –0.21**  4. Personal accomplishment        0.35**  0.32**  0.30**  0.09  0.20*  0.25**  5. Workload          0.43**  0.39**  0.17*  0.22**  0.27**  6. Control            0.44**  0.33**  0.32**  0.28**  7. Reward              0.40**  0.51**  0.44**  8. Community                0.44**  0.25**  9. Fairness                  0.40**    2  3  4  5  6  7  8  9  10    Exhaustion  Depersonal  Personal accomp.  Workload  Control  Reward  Community  Fairness  Values  1 Resilience  –0.37**  –0.10  0.52**  0.21**  0.22**  0.13  0.12  0.07  0.22**  2. Exhaustion    0.38**  –0.33**  –0.62**  –0.39**  –0.35**  –0.07*  –0.24**  –0.37**  3. Depersonalisation      –0.25**  –0.26**  –0.19*  –0.12  –0.08  –0.09  –0.21**  4. Personal accomplishment        0.35**  0.32**  0.30**  0.09  0.20*  0.25**  5. Workload          0.43**  0.39**  0.17*  0.22**  0.27**  6. Control            0.44**  0.33**  0.32**  0.28**  7. Reward              0.40**  0.51**  0.44**  8. Community                0.44**  0.25**  9. Fairness                  0.40**  N = 162; * p < 0.05; ** p < 0.01 (two-tailed). Table 2 Pearson correlation matrix of resilience, MBI and AWLS subscales   2  3  4  5  6  7  8  9  10    Exhaustion  Depersonal  Personal accomp.  Workload  Control  Reward  Community  Fairness  Values  1 Resilience  –0.37**  –0.10  0.52**  0.21**  0.22**  0.13  0.12  0.07  0.22**  2. Exhaustion    0.38**  –0.33**  –0.62**  –0.39**  –0.35**  –0.07*  –0.24**  –0.37**  3. Depersonalisation      –0.25**  –0.26**  –0.19*  –0.12  –0.08  –0.09  –0.21**  4. Personal accomplishment        0.35**  0.32**  0.30**  0.09  0.20*  0.25**  5. Workload          0.43**  0.39**  0.17*  0.22**  0.27**  6. Control            0.44**  0.33**  0.32**  0.28**  7. Reward              0.40**  0.51**  0.44**  8. Community                0.44**  0.25**  9. Fairness                  0.40**    2  3  4  5  6  7  8  9  10    Exhaustion  Depersonal  Personal accomp.  Workload  Control  Reward  Community  Fairness  Values  1 Resilience  –0.37**  –0.10  0.52**  0.21**  0.22**  0.13  0.12  0.07  0.22**  2. Exhaustion    0.38**  –0.33**  –0.62**  –0.39**  –0.35**  –0.07*  –0.24**  –0.37**  3. Depersonalisation      –0.25**  –0.26**  –0.19*  –0.12  –0.08  –0.09  –0.21**  4. Personal accomplishment        0.35**  0.32**  0.30**  0.09  0.20*  0.25**  5. Workload          0.43**  0.39**  0.17*  0.22**  0.27**  6. Control            0.44**  0.33**  0.32**  0.28**  7. Reward              0.40**  0.51**  0.44**  8. Community                0.44**  0.25**  9. Fairness                  0.40**  N = 162; * p < 0.05; ** p < 0.01 (two-tailed). MBI subscales Results show that higher resilience scores are strongly associated with higher personal accomplishment (r(160) = 0.52, p < 0.01) and moderately associated with lower emotional exhaustion scores (r(160) = –0.37, p < 0.01). Resilience was not significantly correlated with depersonalisation (p > 0.05). AWLS subscales The majority of study participants (83 per cent) reported feelings of incongruence with between and the workplace in relation to their work-load. Less than a fifth experienced congruence with their work-load. Just under half (45 per cent) experienced misfit in relation to reward and over half (55 per cent) experienced congruence is related to job engagement. With regard to values, around one-third (35 per cent) felt a sense of misfit and about two-thirds (65 per cent) reflected congruence, suggesting alignment between professional and personal values in relation to the job. Fairness scores showed over two-fifths (43 per cent) reporting misfit and over half (57 per cent) a sense of congruence between equity expectations and reality. More than half (54 per cent) scored a misfit on the control items and just under half (46 per cent) scored congruence on the control items. Finally, community showed that a majority (87 per cent) felt congruence with expectations around relationships with co-workers and managers and a minority (13 per cent) reported a misfit in this area. In terms of the relationships between resilience and AWLS subscales, higher resilience was weak to moderately associated with higher reported work-load congruence (r(160) = 0.21, p < 0.01), sense of control (r(160) = 0.22, p < 0.01) and values scores (r(160) = 0.22, p < .01). Resilience scores were not significantly associated with the other AWLS subscales of perceived rewards, sense of community or fairness (p > 0.05). Predicting resilience using multiple regression analysis A series of hierarchical linear regression models were specified in order to further explain variation in resilience scores using blocks of predictor variables (demographic, work-specific and perceptions of AWLS subscales). The intention was to improve upon the known explanatory power of AWLS in the prediction of resilience by also including demographic (age, gender) and other work-specific variables in the model (level of qualifications, extent of experience, status of post, work sector, caring responsibilities, working in urban/rural areas). However, given the large number of work-specific variables, a preliminary model was specified that included both demographic and all work-specific variables in order to identify the variables making significant unique contributions to the prediction of resilience scores. Only the experience and qualifications variables proved to be significant in the initial model alongside age and gender (R2 = 0.36, Adjusted R2 = 0.14). For the sake of model parsimony and statistical power, only experience and qualifications were subsequently included in the hierarchical analyses along with demographic and AWLS subscales. Hierarchical regression models Model 1 included demographic variables (age, gender), with subsequent models using these as control variables. This model accounted for a small but significant amount of variance (R2 = 0.06, Adjusted R2 = 0.05), with females (β= 0.21, p < 0.01) and older workers (β= 0.18, p < 0.05) reporting higher resilience scores. Model 2 included experience and qualifications variables and results indicated that those with four to five years’ experience (β = 0.35, p < 0.01) and those with more than five years’ experience both reported higher mean resilience scores (β = 0.48, p < 0.01) than those in their AYE. Likewise, those with higher qualifications such as the diploma in social work (β = 0.18, p < 0.01) and a degree in social work (β = 0.41, p < 0.01) recorded higher resilience scores than those with the CQSW (Certificate of Qualification in Social Work) only. The difference was greater for those holding degrees in social work. Although those with master’s-level social work qualifications reported higher resilience scores than CQSW, this difference was not statistically significant (β = 0.15, p > 0.05). Model 2 was a significant improvement on the first model (F(10, 151) = 3.61, p = 0.001) and accounted for an additional 15 per cent of the variance in resilience scores (R2 = 0.21, Adjusted R2 = 0.16). Model 3 further included the AWLS subscales of control and values, and this model explained a further 6 per cent of the variance in resilience in addition to the experience and qualification variables (R2 = 0.27, Adjusted R2 = 0.22). Both control (β = 0.17, p < 0.05) and values (β = 0.16, p < 0.05) made small but significant unique contributions to the model. A fourth model was also tested by adding the remaining AWLS subscales (reward, community, fairness and work-load) but this resulted in no significant improvement in the model (R2 increase = 0.02, p > 0.05) and none of these variables made a significant unique contribution to Model 3, with standardised regression slopes in the range of 0.04–0.15 (p > 0.05). For this reason, Model 4 results have not been included in Table 3. Table 3 Hierarchical multiple regression of resilience RS14 scores on demographic, work-specific variables and selected Area of Work Life Scale (AWLS) subscales Variable  B  Standardised  R2  Adjusted R2  R2 Change  F value (change)  p-value (change)  Model 1  –  –  0.06  0.05  0.06  5.02  0.008   Age  0.22  0.21**             Gender  4.98  0.18*            Model 2  –  –  0.21  0.16  0.15  3.61  0.001   Age  0.18  0.17             Gender  3.36  0.12             Experience AYE (reference)  –  –              1–2 years  3.23  0.13              2–3 years  2.01  0.07              3–4 years  2.93  0.11              4–5 years  12.07  0.35**              5+ years  9.92  0.48**             Qualifications CQSW/Diploma (reference)  –  –             DipHE social work  8.79  0.18*             Degree social work  8.70  0.41**             Master’s social work  4.77  0.15            Model 3  –  –  0.27  0.22  0.06  6.55  0.002   Age  0.16  0.15             Gender  2.50  0.09             Experience AYE (reference)  –  –              1–2 years  3.07  0.12              2–3 years  2.05  0.07              3–4 years  .53  0.02              4–5 years  11.02  0.32**              5+ years  9.41  0.46**             Qualifications CQSW/Diploma (reference)  –  –              DipHE social work  9.29  0.19*              Degree social work  8.49  0.40**              Master’s social work  4.68  0.15             AWLS subscales                  Control  2.12  0.17*              Values  2.47  0.16*            Variable  B  Standardised  R2  Adjusted R2  R2 Change  F value (change)  p-value (change)  Model 1  –  –  0.06  0.05  0.06  5.02  0.008   Age  0.22  0.21**             Gender  4.98  0.18*            Model 2  –  –  0.21  0.16  0.15  3.61  0.001   Age  0.18  0.17             Gender  3.36  0.12             Experience AYE (reference)  –  –              1–2 years  3.23  0.13              2–3 years  2.01  0.07              3–4 years  2.93  0.11              4–5 years  12.07  0.35**              5+ years  9.92  0.48**             Qualifications CQSW/Diploma (reference)  –  –             DipHE social work  8.79  0.18*             Degree social work  8.70  0.41**             Master’s social work  4.77  0.15            Model 3  –  –  0.27  0.22  0.06  6.55  0.002   Age  0.16  0.15             Gender  2.50  0.09             Experience AYE (reference)  –  –              1–2 years  3.07  0.12              2–3 years  2.05  0.07              3–4 years  .53  0.02              4–5 years  11.02  0.32**              5+ years  9.41  0.46**             Qualifications CQSW/Diploma (reference)  –  –              DipHE social work  9.29  0.19*              Degree social work  8.49  0.40**              Master’s social work  4.68  0.15             AWLS subscales                  Control  2.12  0.17*              Values  2.47  0.16*            * p < 0.05; ** p < 0.01 (two-tailed); AYE, Assessed Year in Employment. Table 3 Hierarchical multiple regression of resilience RS14 scores on demographic, work-specific variables and selected Area of Work Life Scale (AWLS) subscales Variable  B  Standardised  R2  Adjusted R2  R2 Change  F value (change)  p-value (change)  Model 1  –  –  0.06  0.05  0.06  5.02  0.008   Age  0.22  0.21**             Gender  4.98  0.18*            Model 2  –  –  0.21  0.16  0.15  3.61  0.001   Age  0.18  0.17             Gender  3.36  0.12             Experience AYE (reference)  –  –              1–2 years  3.23  0.13              2–3 years  2.01  0.07              3–4 years  2.93  0.11              4–5 years  12.07  0.35**              5+ years  9.92  0.48**             Qualifications CQSW/Diploma (reference)  –  –             DipHE social work  8.79  0.18*             Degree social work  8.70  0.41**             Master’s social work  4.77  0.15            Model 3  –  –  0.27  0.22  0.06  6.55  0.002   Age  0.16  0.15             Gender  2.50  0.09             Experience AYE (reference)  –  –              1–2 years  3.07  0.12              2–3 years  2.05  0.07              3–4 years  .53  0.02              4–5 years  11.02  0.32**              5+ years  9.41  0.46**             Qualifications CQSW/Diploma (reference)  –  –              DipHE social work  9.29  0.19*              Degree social work  8.49  0.40**              Master’s social work  4.68  0.15             AWLS subscales                  Control  2.12  0.17*              Values  2.47  0.16*            Variable  B  Standardised  R2  Adjusted R2  R2 Change  F value (change)  p-value (change)  Model 1  –  –  0.06  0.05  0.06  5.02  0.008   Age  0.22  0.21**             Gender  4.98  0.18*            Model 2  –  –  0.21  0.16  0.15  3.61  0.001   Age  0.18  0.17             Gender  3.36  0.12             Experience AYE (reference)  –  –              1–2 years  3.23  0.13              2–3 years  2.01  0.07              3–4 years  2.93  0.11              4–5 years  12.07  0.35**              5+ years  9.92  0.48**             Qualifications CQSW/Diploma (reference)  –  –             DipHE social work  8.79  0.18*             Degree social work  8.70  0.41**             Master’s social work  4.77  0.15            Model 3  –  –  0.27  0.22  0.06  6.55  0.002   Age  0.16  0.15             Gender  2.50  0.09             Experience AYE (reference)  –  –              1–2 years  3.07  0.12              2–3 years  2.05  0.07              3–4 years  .53  0.02              4–5 years  11.02  0.32**              5+ years  9.41  0.46**             Qualifications CQSW/Diploma (reference)  –  –              DipHE social work  9.29  0.19*              Degree social work  8.49  0.40**              Master’s social work  4.68  0.15             AWLS subscales                  Control  2.12  0.17*              Values  2.47  0.16*            * p < 0.05; ** p < 0.01 (two-tailed); AYE, Assessed Year in Employment. Limitations of the study This study is limited by the self-reported nature of the survey and the potential motivation to participate being disproportionally drawn from ‘unhappy’ employees. This has the potential of skewing the data; however, the balance of findings reflects both positive and negative experiences. Therefore, it does not appear that the issue of negative skewedness has been evidenced. The number of variables in the regression model could also be considered a limitation; however, it was important to analyse the potential predictive value of factors contained in the original survey questionnaire. The final multiple regression model reports only on significant results, as shown in Table 3. Age is the only exception to this and was left in the final model as a control variable. The response rate of 43 per cent could also be considered a limitation; however, due to the level of work pressure for social workers, including time capacity regarding work-load, this response rate was considered an achievement. Discussion The aim of the paper was to establish the factors that support child-protection social worker resilience within challenging economic, political and social contexts. It is important to acknowledge that, although the paper is reporting on resilience factors, this is balanced with inclusion of organisational factors and the acknowledgement of the wider political issues that creates the context for social work. The results of the AWLS evidence concerning levels of pressures, particularly about work-load, which have been repeatedly evidenced in research literature. The context of a demanding work-load and public, media and political scrutiny contributes to a context of fear about professional error or omission. Social work is often slated as a failing profession when there is a child death, despite the multidisciplinary nature of child protection. If there is a dominant perspective to ‘adapt’ as Evans and Reid (2013) caution, concerns are justified about how this is likely to impact on staff and ultimately on the service to vulnerable children and families. Previous analysis of the data studied in the current paper (McFadden et al., 2018), found resilience to be a mediator variable between organisational factors and burnout; however, only 7 per cent of the variance was explained by an extension of the two-process model (McFadden et al., 2018). The results presented in the current paper help to explain the relationship self-reported between resilience and burnout more fully by including demographic information and job characteristics as predictor variables of resilience. An important finding in this study is the relationship between emotional exhaustion and resilience scores. Emotional exhaustion is the feeling ‘of being over-extended from one’s job’ (Maslach and Jackson, 1986, p. 13) and has been found to be correlated with work-load in a number of studies (Maslach and Leiter, 2008; McFadden et al., 2018). Emotional exhaustion is accepted as a work-related issue of increasing concern in social work (DePanfilis and Zlotnik, 2008; Aarons, 2009). Levels of work-load were not found to be related to higher levels of emotional exhaustion in a UK-wide study (McFadden, 2015) whereby even those with smaller workloads reported high levels of emotional exhaustion. This suggests that further research on the nature of social work workloads is required to further understand this phenomenon. The RS14 Resilience Scale (Wagnild and Young, 1993) measured resilience in relation to two broad factors of ‘acceptance of life and self’ and ‘individual competence’, and these are operationalised across the five categories of meaningfulness, equanimity, existential aloneness, perseverance and self-reliance. A significant negative relationship exists between emotional exhaustion and resilience. So, when an individual is expressing high emotional exhaustion, resilience is low. Examples of emotional exhaustion items are ‘I feel emotionally drained from my work’, ‘I feel used up at the end of the workday’, ‘I feel burned out from my work’ and ‘I feel at the end of my rope’. Conversely, resilience items measured items such as ‘I usually manage one way or another’, ‘I usually take things in my stride’, ‘I feel I can handle many things at a time’ and ‘In an emergency I am someone people can rely upon’. It is therefore theoretically sound that, as one feels more emotionally exhausted, feelings of resilience will be reduced, as evidenced in the inverted correlation presented in the correlation matrix (Table 2). Another important finding is the relationship between resilience and personal accomplishment. As resilience scores increase, so too personal accomplishment scores increase. Personal accomplishment is measured in items such as ‘I feel I am positively influencing peoples lives’, ‘I can easily create a relaxed atmosphere with my recipients’, ‘I have accomplished many worth-while things in this job’ and ‘I feel exhilarated when working closely with my recipients’. One can understand the congruence in resilience and personal accomplishment scores and how these factors are related, and this is evidenced in the correlation matrix (Table 2). Stalker et al. (2007), in a systematic review, found similar results, whereby social workers had reported high levels of job satisfaction in the context of high emotional exhaustion. Analysis of personal accomplishment scores may help to theoretically explain this phenomenon through a resilience theory lens. The items that measure personal accomplishment in the MBI refer to feelings of accomplishment and meaning applied to the job and the impact on service users’ lives, which is congruent with the broad theoretical resilience domains advocated by Wagnild and Young (1993) with regard to ‘meaning’. Meaning making is confirmed as an important aspect of resilience across cultures and has been applied to many studies relating to populations achieving positive adjustment despite adversity and trauma experiences (Theron and Theron, 2014). In Model 1, females were reported as having higher resilience scores than males; however, when experience was introduced to Model 2, gender was no longer significant. Similarly, age was significant in Model 1 but was no longer significant when experience was introduced. As Model 2 included experience and qualifications, those with four to five years’ experience and those with more than five years’ experience had higher mean resilience scores than those in their AYE. So experience tended to significantly impact on higher levels of resilience, which is evidence to persuade employers and influence policy to maintain workforce retention and avoid turnover. Organisational-level interventions to improve job engagement could, for example, include supervisory support (McFadden, 2015), mentoring schemes, counselling and organisational consultancy (Schaufelli, 2009, p. 205). Qualifications were also important in the findings, as those who reported having a degree or a master’s-level qualification (versus CQSW only) had higher resilience scores. Other research has reported important findings about age, experience and qualifications. Goldberg (2004) found age to be a significant factor in retention and showed that workers under thirty-five years old, those who held a graduate degree and those employed for less than two years were more likely to intend to leave than their counterparts. Post-qualifying training at master’s level (Post Qualifying Award 2–6 and Specialist Award are levels of previous post-qualifying framework in Northern Ireland that are based on up to 120 academic credit points) was found to be a significant predictor of resilience. Other studies on training impact on staff (Curry et al., 2005: Turcotte et al., 2009) confirm this finding. The non-significant levels of post-qualifying training in the findings require further exploration, as there is no obvious explanation for this result. Conclusion Child-protection workers are particularly vulnerable to burnout (Anderson, 2000; Tham and Meagher, 2009); nevertheless, there is evidence of job satisfaction that has been linked to consumer satisfaction and client outcomes (Bednar, 2003), suggesting levels of resilience in this population. The link to ‘client outcomes’ is an important aspect of child protection, as the role means that, when a child fatality happens, social workers are held publicly accountable for not ‘preventing’ the incident. Conversely, positive outcomes for children are rarely reported due to the confidential nature of cases. The resultant political and media pressure on the social work profession creates fear in front line workers that can result in a change in their practice due to the fear of ‘missing’ signs of risk in families (Smith et al., 2003; Stanley, 2013). This paper importantly presents new evidence relating to the significance of holding experience in teams, retaining expertise in child protection and preventing unwanted staff turnover. We have found this may be achieved by making organisational-level adjustments to ensure social workers are supported to manage the challenges of the job. We know that having a realistic work-load and organisational supports such as a positive manager and co-worker relationships and a sense of control, fairness, values and reward are important for retaining staff and reducing turnover intent. However, the modernisation agenda, as discussed by White et al. (2010), in the context of New Labour and continuing to progress under the current Tory government, provides a challenging socio-economic and political landscape for social work, with excessive caseloads and high levels of risk to children being balanced in this context. The results in this paper support an argument for employers and policy makers to revisit the evidence regarding the risk of burnout in child protection and take positive action to support a more resilient workforce. This could be achieved by implementing protective processes in organisational policy to ensure this outcome is achieved in this critical area of social work practice. Funding Department of Education and Learning, Northern Ireland, 2009. Conflict of interest statement. None declared. References Aarons G. A., Fettes D. L., Flores L. E., Sommerfeld D. H. ( 2009) ‘ Evidence-based practice implementation and staff emotional exhaustion in children's services’, Behaviour Research and Therapy , 47( 11), pp. 954– 60. Google Scholar CrossRef Search ADS PubMed  Anderson D. G. ( 2000) ‘ Coping strategies and burnout among veteran child protection workers’, Child Abuse and Neglect , 24( 6), pp. 839– 48. Google Scholar CrossRef Search ADS PubMed  Austin G., Royse D., Kulver K., Plescher K., Zhang Y. ( 2017) ‘ Who stays, who goes, who knows?: A state-wide survey of child welfare workers’, Children and Youth Services Review , 77, pp. 110– 17. Google Scholar CrossRef Search ADS   Bednar S. ( 2003) ‘ Elements of satisfying organizational climates in child welfare agencies’, Families in Society , 84( 1), pp. 7– 12. Google Scholar CrossRef Search ADS   Belsky J., Pluess M. ( 2009) ‘ The nature (and nurture?) of plasticity in early human development’, Perspectives on Psychological Science , 4( 4), pp. 345– 51. Google Scholar CrossRef Search ADS PubMed  Bride B. E., Radey M., Figley C. R. ( 2007) ‘ Measuring compassion fatigue’, Clinical Social Work Journal , 35( 3), pp. 155– 63. Google Scholar CrossRef Search ADS   Bronfenbrenner U. ( 1979) The Ecology of Human Development: Experiments by Nature and Design , Cambridge, MA, Harvard University Press. Burns K. ( 2011) ‘ Career preference, transients and converts: A study of social workers’ retention in child protection and welfare’, British Journal of Social Work , 41( 3), pp. 520– 38. Google Scholar CrossRef Search ADS   Cohen J. W. ( 1988) Statistical Power Analysis for the Behavioral Science , 2nd edn, Hillsdale, NJ, Lawrence Erlbaum Associates. Curry D., McCarragher T., Dellmann-Jenkins M. ( 2005) ‘ Training, transfer, and turnover: Exploring the relationship among transfer of learning factors and staff retention in child welfare’, Children and Youth Services Review , 27( 8), pp. 931– 48. Google Scholar CrossRef Search ADS   Curtis L., Moriarity J., Netton A. ( 2010) ‘ The expected working life of a social worker’, British Journal of Social Work , 40( 5), pp. 1628– 43. Google Scholar CrossRef Search ADS   DePanfilis D., Zlotnik J. L. ( 2008) ‘ Retention of front-line staff in child welfare: A systematic review of research’, Children and Youth Services Review , 30( 9), pp. 995– 1008. Google Scholar CrossRef Search ADS   DeVellis R. F. ( 2003) Scale Development: Theory and Applications , 2nd edn, Thousand Oaks, CA, Sage, 171 pages. Ellett A. J. ( 2009) ‘ Intentions to remain employed in child welfare: The role of human caring, self-efficacy beliefs, and professional organizational culture’, Children and Youth Services Review , 31( 1), pp. 78– 88. Google Scholar CrossRef Search ADS   Evans B., Reid J. ( 2013) ‘ Dangerously exposed: The life and death of the resilient subject’, Resilience: International Policies, Practices and Discourses , 1( 2), pp. 83– 98. Faul F., Erdfelder E., Lang A.-G., Buchner A. ( 2007) ‘ A flexibile statistical power analysis program for the social and behavioral sciences’, Behavior Research Methods , 39( 2), pp. 175– 91. Google Scholar CrossRef Search ADS PubMed  Freudenberger H. J. ( 1974) ‘ Staff burnout’, Journal of Social Issues , 30( 1), pp. 159– 65. Google Scholar CrossRef Search ADS   Garrett P. M. ( 2016) ‘ Questioning tales of “‘ordinary magic”: “Resilience” and neo-liberal reasoning’, British Journal of Social Work , 46( 7), pp. 1909– 25. Google Scholar CrossRef Search ADS   Goldberg Levin A. ( 2004) ‘ Professions at risk: Why are so many workers leaving the field of child welfare? Exploring the relationship between diversity, inclusion, supervisory support, stress, job satisfaction and intention to leave among public child welfare workers’, Dissertation Abstracts International, A: The Humanities and Social Sciences , 64( 12), p. 4624. Hall P. A., Lamont M. (eds) ( 2013) Social Resilience in the Neoliberal Era , Cambridge, Cambridge University Press. Google Scholar CrossRef Search ADS   Hart A., Gagnon E., Eryigit-Madzwamuse S., Cameron J., Aranda K., Rathbone A., Heaver B. ( 2016) ‘ Uniting resilience research and practice with an inequalities approach’, SAGE Open , 6( 4), pp. 1– 15. Google Scholar CrossRef Search ADS   Healy K., Meagher G., Cullin J. ( 2007) ‘ Retaining novices to become expert child protection practitioners: Creating career pathways in direct practice’, British Journal of Social Work , 39( 2), pp. 299– 317. Google Scholar CrossRef Search ADS   Leiter M., Maslach C. ( 2006) The Area of Work Life Survey Manual , 4th edn, Nova Scotia, Canada, Centre for Organizational Research and Development, Acadia University. Maslach C., Jackson S. ( 1986) Maslach Burnout Inventory Manual , 2nd edn, California, CPP, Inc. Maslach C., Leiter M. ( 2008) ‘ Early predictors of job burnout and engagement’, Journal of Applied Psychology , 93( 3), pp. 498– 512. Google Scholar CrossRef Search ADS PubMed  Maslach C., Leiter M. P. ( 2005) ‘Stress and burnout: The critical research’, in Cooper C. L. (ed.), Handbook of Stress Medicine and Health , 2nd edn, London, CRC Press, pp. 153– 70. Maslach C., Jackson S., Leiter M. ( 1996) Maslach Burnout Inventory Manual , 3rd edn, California, CPP, Inc. Masten A. S., Wright M. O. ( 2010) ‘Resilience over the lifespan: Developmental perspectives on resistance, recovery and transformation’, in Reich J. W., Zautra A. J., Hall J. S. (eds), Handbook of Adult Resilience , California, New York, The Guilford Press, pp. 213– 37. McEwen B. S., Gianaros P. J. ( 2010) ‘ Central role of the brain in stress and adaptation: Links to socioeconomic status, health, and disease’, Annals of the New York Academy of Sciences , 1186( 1), pp. 190– 222. Google Scholar CrossRef Search ADS PubMed  McFadden P. ( 2013) ‘Resilience and burnout in child protection social work’, PhD thesis, University of Ulster, Northern Ireland. McFadden P. ( 2015) ‘Measuring burnout among UK social workers: A Community Care study’, available online at: https://s3-eu-west-1.amazonaws.com/rbi-communities/wp-content/uploads/sites/7/2015/07/Burnout-among-UK-social-workers.pdf (accessed on 1 July 2015). McFadden P., Campbell A., Taylor B. ( 2014) ‘ Resilience and burnout in child protection social work: Individual and organisational themes from a systematic literature review’, British Journal of Social Work , 45( 5), pp. 1546– 63. Google Scholar CrossRef Search ADS   McFadden P., Mallett J., Leiter M. ( 2018) ‘ An extension of the two process model of burnout: The role of resilience, reward and community relationships in child protection social work’, Stress & Health: Journal of the International Society for the Investigation of Stress , 34( 1), pp. 72– 83. Google Scholar CrossRef Search ADS   Mor Barak M. E., Nissly J. A., Levin A. ( 2001) ‘ Antecedents to retention and turnover among child welfare, social work, and other human service employees: What can we learn from past research: A review and meta-analysis?’, Social Service Review , 75( 4), pp. 625– 61. Google Scholar CrossRef Search ADS   Munro E. ( 2011) Review of Child Protection: A Child Centered System , London, Department for Education. Nordick W. G. ( 2002) ‘Striking balance, enjoying challenge: How social workers in child protection stay on the high wire’, UBC Retrospective theses Digitization Project, available online at: http://www.Library.Ubc.ca/archives/retro_theses/ (accessed on 22 March 2018). Pallant J. ( 2007) SPSS Survival Manual , 3rd edn, Buckingham, Open University Press. Rutter M. ( 1987) ‘Psychosocial resilience and protective mechanisms’, in Rolf J., Masten A. S., Cicchetti D., Nuechterlein K. H., Weintraub S. (eds), Risk and Protective Factors in the Development of Psychopathology , NY, New York, Cambridge Press, pp. 181– 214. Schaufeli W. B., Leiter M. P., Maslach C. ( 2009) ‘ Burnout: 35 years of research and practice’, Career Development International , 14( 3), pp. 204– 20. Google Scholar CrossRef Search ADS   Smith B. D. ( 2005) ‘ Job retention in child welfare: Effects of perceived organizational support, supervisor support, and intrinsic job value’, Children and Youth Services Review , 27( 2), pp. 153– 69. Google Scholar CrossRef Search ADS   Smith M., McMahon N., Nurston J. ( 2003) ‘ Social workers’ experiences of fear’, British Journal of Social Work , 33( 5), pp. 659–7. Stalker C. A., Mandell D., Frensch K. A., Harvey C., Wright M. ( 2007) ‘ Children welfare workers who are exhausted yet satisfied with their jobs: How do they do it?’, Child and Family Social Work , 12( 2), pp. 182– 91. Google Scholar CrossRef Search ADS   Stanley T. ( 2013) ‘“ Our tariff will rise”: Risk, probabilities and child protection’, Health, Risk and Society , 15( 1), pp. 67– 83. Google Scholar CrossRef Search ADS   Strolin-Goltzman J., McCarthy M., Smith B., Caringi J., Bronstein L., Lawson H. ( 2008) ‘ Should I stay or should I go? A comparison study of intention to leave among public child welfare systems with high and low turnover rates’, Child Welfare , 87( 4), pp. 125– 43. Google Scholar PubMed  Tham P., Meagher G. ( 2009) ‘ Working in human services: How do experiences and working conditions in child welfare social work compare?’, British Journal of Social Work , 39( 5), pp. 807– 27. Google Scholar CrossRef Search ADS   Theron L., Theron A. ( 2014) ‘ Meaning-making and resilience: Case studies of a multifaceted process’, Journal of Psychology in Africa , 24( 1), pp. 24– 32. Turcotte D., Lamonde G., Beaudoin A. ( 2009) ‘ Evaluation of an in-service training program for child welfare practitioners’, Research on Social Work Practice , 19( 1), pp. 31– 41. Google Scholar CrossRef Search ADS   Ungar M. ( 2008) ‘ Resilience across cultures’, British Journal of Social Work , 38( 2), pp. 218– 35. Google Scholar CrossRef Search ADS   Ungar M. ( 2013) ‘ Resilience, trauma, context, and culture’, Journal of Trauma, Violence, and Abuse , 14( 3), pp. 255– 66. Google Scholar CrossRef Search ADS   Wagnild G. ( 2009) The Resilience Scale User’s Guide for the US English Version of the Resilience Scale and the 14 item Resilience Scale (RS-14). USA, The Resilience Centre. Wagnild G., Young H. ( 1993) ‘ Development and psychometric evaluation of the Resilience Scale’, Journal of Nursing Measurement , 1( 2), pp. 165– 7. Google Scholar PubMed  Weaver D., Chang J., Clark S., Rhee S. ( 2007) ‘ Keeping public child welfare workers on the job’, Administration in Social Work , 31( 2), pp. 5– 25. Google Scholar CrossRef Search ADS   Werner E., Smith R. ( 1982) Vulnerable but Not Invincible: A Longitudinal Study of Resilient Children and Youth , New York, McGraw Hill. White S., Wastell D., Broadhurst K., Hall C. ( 2010) ‘ When policy o’erleaps itself: The tragic tale of the integrated children’s system’, Critical Social Policy , 30( 3), p. 405– 29. Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press on behalf of The British Association of Social Workers. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The British Journal of Social Work Oxford University Press

Explaining Self-Reported Resilience in Child-Protection Social Work: The Role of Organisational Factors, Demographic Information and Job Characteristics

Loading next page...
 
/lp/ou_press/explaining-self-reported-resilience-in-child-protection-social-work-MDyasNeJpC
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of The British Association of Social Workers. All rights reserved.
ISSN
0045-3102
eISSN
1468-263X
D.O.I.
10.1093/bjsw/bcy015
Publisher site
See Article on Publisher Site

Abstract

Abstract Child-protection social work is a stressful occupation that results in workforce concerns about poor levels of staff retention and high levels of inexperience. This paper presents results from a cross-sectional survey and reports findings from a sample of 162 Northern Irish social workers. The sample were measured for ‘resilience’ (acceptance of self and life and individual competence, RS14 Resilience Scale), ‘burnout’ (emotional exhaustion EE, depersonalisation DP, personal accomplishment PA, Maslach Burnout Inventory) and organisational subscales (work-load, community, values, equity and control, Area of Work Life Scale (AWLS)). Pearson zero-order correlations showed that higher resilience was associated with lower EE and higher PA. Hierarchical linear regression analysis was used to identify unique demographic and work-specific predictors of resilience in addition to the AWLS subscales of control and values. The final model significantly accounted for 27 per cent of the variance in resilience scores, providing increased knowledge about resilience enhancing factors. As resilience is not an apolitical concept, the wider debates and politics of resilience are acknowledged. Specifically, contextual concerns are addressed that relate to the organisational factors that impact on social workers. The paper concludes by calling for organisational interventions to support resilience in social workers and maintain expertise in child-protection services. Burnout, child protection, child welfare, organisational factors, resilience Introduction Child-protection social work is acknowledged as a challenging and stressful career that is concerned with protecting children from abuse and neglect. Social workers are required to assess complex risk and make critical decisions that impact on the lives of children and families. This often is in the context of multiple and competing factors, contested information and stressful legal, moral and ethical challenges. Due to the unique nature of the role, retaining experience in child-protection teams has been a major issue due to high levels of staff turnover (Healy et al., 2007). Compassion fatigue and vicarious trauma have been reported in this profession, with concern about this risk increasing levels of burnout and job exit (Bride et al., 2007). In the current economic climate, with permanent job alternatives scarce, increase in use of agency staff and unfilled vacancies, retention of staff may be due to a lack of choice opposed to a desire to stay (Austin et al., 2017). This may impact negatively on not only the well-being of staff, but also the quality of the service they provide. Work-load is acknowledged in the literature as a contributing factor to burnout and staff turnover (Stalker et al., 2007; McFadden et al., 2014). Political and media pressure on the social work profession creates additional unease in front line workers that can result in a change in their practice due to the fear of ‘missing’ signs of dangerousness in families (Smith et al., 2003). The level of administrative tasks and over-bureaucratised systems that take staff away from critical face-to-face social work practice is acknowledged by the government-led Review of Child Protection (Munro, 2011) and in the wider literature (White et al., 2010). Recruiting and retaining experience in teams is critical to quality social work both to support junior staff and add to team stability and camaraderie (Strolin-Goltzman et al., 2008; McFadden et al., 2014). Burns (2011), in an Irish study on child protection, found that turnover was not a consistent issue and that some staff were ‘career converts’, originally entering the profession as a ‘stepping stone’ to something else and then remaining due to having found meaning and satisfaction in the role. This was reportedly related to good supervisory support and relationships that helped sustain individuals. US research on child welfare retention has also found that positive relationships with a supervisor and team can ‘buffer’ the impact of burnout and attrition (Mor Barak et al., 2001; Curry et al., 2005; Smith, 2005; Weaver et al., 2007). Self-reported resilience has also has been found to mediate the relationship between organisational factors and burnout (McFadden et al., 2018). The current paper drills down into this finding by accounting for age and gender as well as job-specific characteristics, which provides further understanding of what factors exist to support social workers to maintain feelings of resilience in this profession. The research question addressed in this paper is as follows: ‘What are the roles of organisational factors, demographic variables and job characteristics in child-protection social worker resilience?’ Theory and definitions Many definitions of resilience exist. From a range of perspectives, resilience can mean different things and the concept is contested and widely debated. Definitional perspectives of resilience have evolved and focus has moved progressively from individual characteristics to systems, context and social justice concerns (Ungar, 2013; Hart, 2016). Resilience is viewed as ‘doing better than expected’ following adversity using a range of protective processes available (Rutter, 1987). When applied to child-protection social work, the relevance of ‘resilience’ has an important application and meaning. However, the organisational environment is the context that must be considered to avoid assumptions about resilient ‘individuals’. Intra-personal references to resilience are criticised for being pathogenic, ignoring contextual, environmental, political and economic factors (Garrett, 2016). Whilst self-reported feelings of ‘resilience’ are factors selected for the current study, this is balanced with the application of organisational factors. These include work-load, social support, equity, values and fairness, to ensure that pathologic explanations do not dominate the findings. Inter-personal dimensions of resilience include references to social resilience, which is multidimensional, and refers to groups and communities coming together to advance their interests in the face of adverse challenges (Hall and Lamont, 2011). When one applies a systems model such as the social ecological perspective (Bronfenbrenner, 1979), it is clear that resilience is not dependent on individual characteristics, but relies on an interactive dynamic between the individual and their environment (Masten and Wright, 2010) and risk and protective factors (Rutter, 1987). Masten and Wright (2010) suggest that systems of meaning making, inter-personal relationships, agency and mastery, emotional and behavioural regulation, problem solving and cultural and religious traditions are related to resilience. Another aspect of resilience research seeks to further understand the developmental biological basis of resilience in interaction with a life-course trajectory that includes environmental factors, showing that resilience is not a ‘fixed’ state, but ‘plastic’ and context, time and experience-specific (Belsky and Pluess, 2009). Wagnild and Young (1993) propose a definition that applies typical individualistic factors to resilient people, such as personal strength, emotional stamina and self-confidence. The tool will be discussed later in the methodology section. In the context of occupational resilience, the concept of ‘job engagement’ is applied in terms of how ‘job-engaged’ (Leiter and Maslach, 2006) or ‘committed’ (Ellett, 2009) workers are despite the pressures and stresses of the job. The current paper does not conceptualise resilience as a personality trait (Werner and Smith, 1982), but as an interactive phenomenon influenced by the person and their environment and an ongoing adaptive capacity and process (Bronfenbrenner, 1979; Rutter, 1987). Therefore, organisational structures and features that provide the context, culture and climate for social work activity are an essential consideration in this debate. Job engagement and burnout run along a line of continuum that is captured by extensive research (Leiter and Maslach, 2005). Burnout is characterised by emotional exhaustion, depersonalisation and low personal accomplishment, whilst job-engaged individuals report opposite characteristics, indicating energy, empathy and efficacy (Maslach and Leiter, 2008). The MBI (Maslach Burnout Inventory) and the AWLS (Area of Work Life Scale) measure this continuum in relation to three interrelated concepts looking at the organisational correlates of work-load, fairness, community, control, values and reward (Maslach and Leiter, 2008). These measures are applied in the current paper. The concept of job burnout became popular in the 1970s and was viewed as ‘A characterization of adverse reactions to work, primarily in the human services settings’ (Freudenberger, 1974, p. 159). Schaufeli et al. (2009, p. 204) argue that burnout is deeply rooted in the broad social, economic and cultural developments that have been part of the occupational landscape for more than thirty-five years. They argue that the shift from being an industrial society to a service-led economy underpins the major occupational transformation. With these many changes, there is a level of social transformation that manifests in psychological pressures on people in transition and change. Not all perspectives on burnout take a ‘political’ stance on the issue but, instead, the risk of burnout can be seen as a ‘cost risk’ with a focus on retention and turnover, absence and sick leave, retraining costs and low experience (Curtis et al., 2010). Curtis et al. (2010) refer to an average eight-year lifespan of social workers compared to the longevity in other professions to demonstrate this point. References to ‘resilience’ in contexts of job-related adversity are criticised for the pathological assumptions made about individuals needing to ‘adapt’ or willingness to ‘change’ having major political underpinnings. Evans and Reid (2013) argue that this imposes a pathological and individualistic responsibility on people that ignores wider contextual and structural concerns, which is politically ‘catastrophic’. When applied to social work, this critique resonates. As social workers self-select into the profession, there is a risk of this perspective being applied to those who struggle with the challenges of the job, with the emphasis on individuals being called on to ‘adapt’ or ‘change’ to manage ‘threats now presupposed as endemic’ (Evans and Reid, 2013, p. 83). This is especially true when a body of research has exposed the specific challenges in relation to contextual concerns of high work-load, bureaucracy, varying levels of supervisory support, burnout and inexperience in teams (Stalker et al., 2007; Healy et al., 2007; Munro, 2011; McFadden et al., 2014). Aims and objectives The overarching aim of this study was to examine the factors that contribute to feelings of resilience and burnout in the child-protection workforce in Northern Ireland. The objectives were to identify the protective factors that sustain resilience in some staff and also to identify and measure the impact of organisational factors and how these relate to experiences of burnout and resilience in this occupational group. Organisational contexts are examined from the perspective of job engagement or burnout to measure the perceived impact of work-load, values, community, equity, control and reward. Demographic variables are also measured. Methodology This study used a cross-sectional survey design using standardised tests (MBI, RS14 and AWLS). Stratified random sampling was applied, as the population of social workers had natural groupings into a number of strata (gateway, family intervention/support, voluntary sector) and further divided into demographic variables such as age, gender, length of experience, job status, employer, caring responsibilities, religion, post qualifications, qualifications and rural or urban caseload. The survey was distributed using Smart Survey (c) software and was delivered by e-mail to social workers who had confidential status to maximise participation. Ethical approval This study was approved by the Office of Research Ethics in Northern Ireland (August 2010). The main ethical issue was the need to consider management of any identification of poor or concerning practice associated with burnout, with an agreed protocol to manage this potential. Although the protocol did not have to be implemented, the process of managing any concerns relating to practice was helpful and also required critical evaluation of a range of considerations such as unprofessional practice or concerning attitudes relating to service recipients. All participants gave informed consent to participate in the study and participant information literature provided all required information, including reference to this concern. Sampling frame The study used stratified random sampling across six participatory organisations: five Health and Social Care Trusts and those employed within a voluntary-sector organisation Child Protection Team. Two statutory structures for front line child-protection staff were the target sample: ‘gateway’ and ‘family intervention’ (intermittently referred to as ‘family support’). Gateway is the structure for all referrals relating to child-protection concerns and the point of entry into social work system. Family support/intervention accept referrals from gateway staff in cases that require long-term family support and child-protection services. This method of sampling increases the ability to claim representativeness; however, the authors use this claim cautiously due to the sample size drawn from each strata (Table 1). Table 1 Variations in response rates across participating organisations HSC Trust  Potential participants  Actual participants  % response rate  A  90  37  41  B  72  33  46  C  76  28  39  D  59  22  37  E  64  24  38  Voluntary sector  19  18  95  Total  380  162  43  HSC Trust  Potential participants  Actual participants  % response rate  A  90  37  41  B  72  33  46  C  76  28  39  D  59  22  37  E  64  24  38  Voluntary sector  19  18  95  Total  380  162  43  Table 1 Variations in response rates across participating organisations HSC Trust  Potential participants  Actual participants  % response rate  A  90  37  41  B  72  33  46  C  76  28  39  D  59  22  37  E  64  24  38  Voluntary sector  19  18  95  Total  380  162  43  HSC Trust  Potential participants  Actual participants  % response rate  A  90  37  41  B  72  33  46  C  76  28  39  D  59  22  37  E  64  24  38  Voluntary sector  19  18  95  Total  380  162  43  The sampling frame for the study included those in their Assessed Year in Employment (AYE—this is the first year following graduation, which ensures the employee has reached an acceptable standard to proceed to fully qualified social worker status). Also included were front line social work posts, those in senior practitioner posts and team leaders. The emphasis was on those who held caseloads and professional accountability for risk and protection of vulnerable children and families. The overall population of social workers who met the inclusion criteria was 380. Sample characteristics The gender of those who took part in the study was 86 per cent female (n = 140) and 14 per cent male (n = 22). This ratio was consistent with the gender profile of social workers in Northern Ireland as a whole and is evidenced in the gender ratio of social work more widely (McFadden, 2013). The contract status levels were reflected across the sample, ranging from those in their AYE through to senior practitioner and those ‘acting’ in team-leader posts. The most frequently participating subjects were those who had permanent contracts at a rate of 67 per cent (n = 109). Senior practitioners represent the second most frequently occurring participation rate at 9 per cent (n = 14). Those in their AYE (8 per cent, n = 13) and agency staff (6 per cent, n = 10) were the least frequently participating groups in the study. Participants’ ages ranged from twenty-two to fifty-eight years old. The mean age of participants was thirty-six years old (SD = 9.31). Research tools All research tools in the study are validated measures. The Resilience Scale RS14 (Wagnild and Young, 1993) has fourteen items measuring self-reported feelings of resilience across five broad factor areas of self-reliance, meaning, equanimity, perseverance and existential aloneness. This was the outcome variable measured in the context of predictor variables that will be outlined below. Factor analysis indicated that the Resilience Scale has two major factors, namely ‘acceptance of self and life’ and ‘individual competence’ (Wagnild and Young, 1993). Perseverance is the act of persistence despite adversity or lack of encouragement. It also means the there is a determination to remain actively involved in the management of adverse life events, being able to redefine oneself and being self-disciplined. Equanimity is the presence of a balanced view of life and experiences, and the ability to form an overview of a broader range of experiences and feel ready for unforeseen challenges, which moderates the potential for an extreme response to adverse events. Meaningfulness is the ability to understand that life has a purpose that one has a role in with ‘meaning’, providing the individual with something to live for. Self-reliance is the ability to have self-belief about one’s role in life and the abilities and capacities present or absent in oneself. Existential aloneness was a particularly useful concept in the measurement of resilience in child-protection social workers. This, generally applied, is the understanding of our aloneness in the world and the uniqueness of one’s life and experiences and individual responsibilities in one’s lifetime (Wagnild, 2009, p. 23). Resilience items measure statements such as ‘I usually manage one way or another’, ‘I usually take things in my stride’, ‘I feel I have accomplished things in life’, ‘I feel I can handle many things at a time’ and ‘In an emergency I am someone people can rely upon’. Ideally, the internal consistency of a scale should be above 0.7 (DeVellis, 2003). Internal consistency reliability (Cronbach’s a coefficient) was 0.90 in the current sample. The MBI (Maslach and Jackson, 1986) measures self-perception of experienced emotional exhaustion, depersonalisation and personal accomplishment. A series of twenty-two questions measures these three dimensions and subjects are asked to report the frequency of experienced feelings and thoughts using a Likert scale ranging from (0) ‘never’ to (6) ‘every day’. This scale measures the individual’s experience on the continuum between burnout and job engagement. Three interrelated dimensions are exhaustion energy, cynicism involvement, and inefficacy and efficacy (Maslach and Leiter, 2008). The exhaustion component is linked to the psychological strain associated with individual experience of burnout. Cynicism is the inter-personal context that refers to an emotional detachment or negative feelings towards service users and other aspects of the job. The efficacy dimension is related to the self-evaluation dimension of burnout and is experienced when an individual has feelings of incompetence or a lack of achievement in relation to their job. Internal consistencies (Cronbach’s a) for the MBI subscales were acceptable—emotional exhaustion (0.90), depersonalisation (0.75) and personal accomplishment (0.78). The AWLS has twenty-nine items and measured organisational factors such as work-load, community, control, fairness, equity and values. The items are worded in terms of perceived congruence or incongruence between self and workplace. Items reflect positively framed statements of congruence, such as ‘I have enough time to do what’s important in my job’, as well as negatively framed items of incongruence, such as ‘Working here forces me to compromise my values’. Subjects indicate their agreement or disagreement with these statements using a five-point Likert scale, from (1) ‘strongly disagree’ to (5) ‘strongly agree’ (Maslach and Leiter, 2008). Scoring for the negatively framed items is reversed. This tool defines congruence as a high score (greater than 3.00), meaning a high degree of ‘fit’ between the person and workplace. Incongruence between the person and the workplace is the reverse. As the score is low (less than 3.00), this suggests a ‘misfit’ between the person and workplace. Maslach and Leiter (2008) report that this misfit is related to burnout. Maslach and Leiter (2008) developed a model of burnout that focused on the perceived congruency between the worker and important aspects of the organisational environment. According to this model, a higher level of incongruence is directly related to an increased risk in burnout. Conversely, the higher degree of job–person congruity is correlated with a higher chance of job engagement. Internal consistency coefficients for the AWLS subscales were above 0.7—work-load (0.84), reward (0.91), community (0.87), fairness (0.77) and values (0.74)—with the exception of ‘control’ (0.68). The demographics measured include: team and sector, status of the post, length of experience, employer, gender, age, marital status and caring responsibilities, religious background, post qualifications, qualifications and rural–urban caseload. Data analysis The data were analysed using SPSS 23. Ordinary least-squares (OLS) hierarchical regression analysis was used to examine the ability of the variables to predict resilience scores (RS14). Response rate The number of child-protection social workers who participated in the survey was 162 out of a possible target sample of 380 social workers. The collective response rate across all six participating organisations was 43 per cent (n = 162). This was a regional study in Northern Ireland that had one voluntary-sector child-protection organisation and all five Health and Social Care Trusts included in the sample. Correlates of resilience Table 2 illustrates the inter-correlations among resilience, MBI and AWLS subscales. This table demonstrates the low to moderate correlations between each of the AWLS and the Resilience RS14 scale. This meets the OLS regression assumption of predictor variables not being too highly correlated for regression analysis. Pallant (2007, p. 132) refers researchers to Cohen’s (1988, p. 79) suggested ranges for interpreting Pearson correlation coefficients as follows: small = 0.10–0.29, medium = 0.30–0.49 and large = 0.50–1.00. Table 2 Pearson correlation matrix of resilience, MBI and AWLS subscales   2  3  4  5  6  7  8  9  10    Exhaustion  Depersonal  Personal accomp.  Workload  Control  Reward  Community  Fairness  Values  1 Resilience  –0.37**  –0.10  0.52**  0.21**  0.22**  0.13  0.12  0.07  0.22**  2. Exhaustion    0.38**  –0.33**  –0.62**  –0.39**  –0.35**  –0.07*  –0.24**  –0.37**  3. Depersonalisation      –0.25**  –0.26**  –0.19*  –0.12  –0.08  –0.09  –0.21**  4. Personal accomplishment        0.35**  0.32**  0.30**  0.09  0.20*  0.25**  5. Workload          0.43**  0.39**  0.17*  0.22**  0.27**  6. Control            0.44**  0.33**  0.32**  0.28**  7. Reward              0.40**  0.51**  0.44**  8. Community                0.44**  0.25**  9. Fairness                  0.40**    2  3  4  5  6  7  8  9  10    Exhaustion  Depersonal  Personal accomp.  Workload  Control  Reward  Community  Fairness  Values  1 Resilience  –0.37**  –0.10  0.52**  0.21**  0.22**  0.13  0.12  0.07  0.22**  2. Exhaustion    0.38**  –0.33**  –0.62**  –0.39**  –0.35**  –0.07*  –0.24**  –0.37**  3. Depersonalisation      –0.25**  –0.26**  –0.19*  –0.12  –0.08  –0.09  –0.21**  4. Personal accomplishment        0.35**  0.32**  0.30**  0.09  0.20*  0.25**  5. Workload          0.43**  0.39**  0.17*  0.22**  0.27**  6. Control            0.44**  0.33**  0.32**  0.28**  7. Reward              0.40**  0.51**  0.44**  8. Community                0.44**  0.25**  9. Fairness                  0.40**  N = 162; * p < 0.05; ** p < 0.01 (two-tailed). Table 2 Pearson correlation matrix of resilience, MBI and AWLS subscales   2  3  4  5  6  7  8  9  10    Exhaustion  Depersonal  Personal accomp.  Workload  Control  Reward  Community  Fairness  Values  1 Resilience  –0.37**  –0.10  0.52**  0.21**  0.22**  0.13  0.12  0.07  0.22**  2. Exhaustion    0.38**  –0.33**  –0.62**  –0.39**  –0.35**  –0.07*  –0.24**  –0.37**  3. Depersonalisation      –0.25**  –0.26**  –0.19*  –0.12  –0.08  –0.09  –0.21**  4. Personal accomplishment        0.35**  0.32**  0.30**  0.09  0.20*  0.25**  5. Workload          0.43**  0.39**  0.17*  0.22**  0.27**  6. Control            0.44**  0.33**  0.32**  0.28**  7. Reward              0.40**  0.51**  0.44**  8. Community                0.44**  0.25**  9. Fairness                  0.40**    2  3  4  5  6  7  8  9  10    Exhaustion  Depersonal  Personal accomp.  Workload  Control  Reward  Community  Fairness  Values  1 Resilience  –0.37**  –0.10  0.52**  0.21**  0.22**  0.13  0.12  0.07  0.22**  2. Exhaustion    0.38**  –0.33**  –0.62**  –0.39**  –0.35**  –0.07*  –0.24**  –0.37**  3. Depersonalisation      –0.25**  –0.26**  –0.19*  –0.12  –0.08  –0.09  –0.21**  4. Personal accomplishment        0.35**  0.32**  0.30**  0.09  0.20*  0.25**  5. Workload          0.43**  0.39**  0.17*  0.22**  0.27**  6. Control            0.44**  0.33**  0.32**  0.28**  7. Reward              0.40**  0.51**  0.44**  8. Community                0.44**  0.25**  9. Fairness                  0.40**  N = 162; * p < 0.05; ** p < 0.01 (two-tailed). MBI subscales Results show that higher resilience scores are strongly associated with higher personal accomplishment (r(160) = 0.52, p < 0.01) and moderately associated with lower emotional exhaustion scores (r(160) = –0.37, p < 0.01). Resilience was not significantly correlated with depersonalisation (p > 0.05). AWLS subscales The majority of study participants (83 per cent) reported feelings of incongruence with between and the workplace in relation to their work-load. Less than a fifth experienced congruence with their work-load. Just under half (45 per cent) experienced misfit in relation to reward and over half (55 per cent) experienced congruence is related to job engagement. With regard to values, around one-third (35 per cent) felt a sense of misfit and about two-thirds (65 per cent) reflected congruence, suggesting alignment between professional and personal values in relation to the job. Fairness scores showed over two-fifths (43 per cent) reporting misfit and over half (57 per cent) a sense of congruence between equity expectations and reality. More than half (54 per cent) scored a misfit on the control items and just under half (46 per cent) scored congruence on the control items. Finally, community showed that a majority (87 per cent) felt congruence with expectations around relationships with co-workers and managers and a minority (13 per cent) reported a misfit in this area. In terms of the relationships between resilience and AWLS subscales, higher resilience was weak to moderately associated with higher reported work-load congruence (r(160) = 0.21, p < 0.01), sense of control (r(160) = 0.22, p < 0.01) and values scores (r(160) = 0.22, p < .01). Resilience scores were not significantly associated with the other AWLS subscales of perceived rewards, sense of community or fairness (p > 0.05). Predicting resilience using multiple regression analysis A series of hierarchical linear regression models were specified in order to further explain variation in resilience scores using blocks of predictor variables (demographic, work-specific and perceptions of AWLS subscales). The intention was to improve upon the known explanatory power of AWLS in the prediction of resilience by also including demographic (age, gender) and other work-specific variables in the model (level of qualifications, extent of experience, status of post, work sector, caring responsibilities, working in urban/rural areas). However, given the large number of work-specific variables, a preliminary model was specified that included both demographic and all work-specific variables in order to identify the variables making significant unique contributions to the prediction of resilience scores. Only the experience and qualifications variables proved to be significant in the initial model alongside age and gender (R2 = 0.36, Adjusted R2 = 0.14). For the sake of model parsimony and statistical power, only experience and qualifications were subsequently included in the hierarchical analyses along with demographic and AWLS subscales. Hierarchical regression models Model 1 included demographic variables (age, gender), with subsequent models using these as control variables. This model accounted for a small but significant amount of variance (R2 = 0.06, Adjusted R2 = 0.05), with females (β= 0.21, p < 0.01) and older workers (β= 0.18, p < 0.05) reporting higher resilience scores. Model 2 included experience and qualifications variables and results indicated that those with four to five years’ experience (β = 0.35, p < 0.01) and those with more than five years’ experience both reported higher mean resilience scores (β = 0.48, p < 0.01) than those in their AYE. Likewise, those with higher qualifications such as the diploma in social work (β = 0.18, p < 0.01) and a degree in social work (β = 0.41, p < 0.01) recorded higher resilience scores than those with the CQSW (Certificate of Qualification in Social Work) only. The difference was greater for those holding degrees in social work. Although those with master’s-level social work qualifications reported higher resilience scores than CQSW, this difference was not statistically significant (β = 0.15, p > 0.05). Model 2 was a significant improvement on the first model (F(10, 151) = 3.61, p = 0.001) and accounted for an additional 15 per cent of the variance in resilience scores (R2 = 0.21, Adjusted R2 = 0.16). Model 3 further included the AWLS subscales of control and values, and this model explained a further 6 per cent of the variance in resilience in addition to the experience and qualification variables (R2 = 0.27, Adjusted R2 = 0.22). Both control (β = 0.17, p < 0.05) and values (β = 0.16, p < 0.05) made small but significant unique contributions to the model. A fourth model was also tested by adding the remaining AWLS subscales (reward, community, fairness and work-load) but this resulted in no significant improvement in the model (R2 increase = 0.02, p > 0.05) and none of these variables made a significant unique contribution to Model 3, with standardised regression slopes in the range of 0.04–0.15 (p > 0.05). For this reason, Model 4 results have not been included in Table 3. Table 3 Hierarchical multiple regression of resilience RS14 scores on demographic, work-specific variables and selected Area of Work Life Scale (AWLS) subscales Variable  B  Standardised  R2  Adjusted R2  R2 Change  F value (change)  p-value (change)  Model 1  –  –  0.06  0.05  0.06  5.02  0.008   Age  0.22  0.21**             Gender  4.98  0.18*            Model 2  –  –  0.21  0.16  0.15  3.61  0.001   Age  0.18  0.17             Gender  3.36  0.12             Experience AYE (reference)  –  –              1–2 years  3.23  0.13              2–3 years  2.01  0.07              3–4 years  2.93  0.11              4–5 years  12.07  0.35**              5+ years  9.92  0.48**             Qualifications CQSW/Diploma (reference)  –  –             DipHE social work  8.79  0.18*             Degree social work  8.70  0.41**             Master’s social work  4.77  0.15            Model 3  –  –  0.27  0.22  0.06  6.55  0.002   Age  0.16  0.15             Gender  2.50  0.09             Experience AYE (reference)  –  –              1–2 years  3.07  0.12              2–3 years  2.05  0.07              3–4 years  .53  0.02              4–5 years  11.02  0.32**              5+ years  9.41  0.46**             Qualifications CQSW/Diploma (reference)  –  –              DipHE social work  9.29  0.19*              Degree social work  8.49  0.40**              Master’s social work  4.68  0.15             AWLS subscales                  Control  2.12  0.17*              Values  2.47  0.16*            Variable  B  Standardised  R2  Adjusted R2  R2 Change  F value (change)  p-value (change)  Model 1  –  –  0.06  0.05  0.06  5.02  0.008   Age  0.22  0.21**             Gender  4.98  0.18*            Model 2  –  –  0.21  0.16  0.15  3.61  0.001   Age  0.18  0.17             Gender  3.36  0.12             Experience AYE (reference)  –  –              1–2 years  3.23  0.13              2–3 years  2.01  0.07              3–4 years  2.93  0.11              4–5 years  12.07  0.35**              5+ years  9.92  0.48**             Qualifications CQSW/Diploma (reference)  –  –             DipHE social work  8.79  0.18*             Degree social work  8.70  0.41**             Master’s social work  4.77  0.15            Model 3  –  –  0.27  0.22  0.06  6.55  0.002   Age  0.16  0.15             Gender  2.50  0.09             Experience AYE (reference)  –  –              1–2 years  3.07  0.12              2–3 years  2.05  0.07              3–4 years  .53  0.02              4–5 years  11.02  0.32**              5+ years  9.41  0.46**             Qualifications CQSW/Diploma (reference)  –  –              DipHE social work  9.29  0.19*              Degree social work  8.49  0.40**              Master’s social work  4.68  0.15             AWLS subscales                  Control  2.12  0.17*              Values  2.47  0.16*            * p < 0.05; ** p < 0.01 (two-tailed); AYE, Assessed Year in Employment. Table 3 Hierarchical multiple regression of resilience RS14 scores on demographic, work-specific variables and selected Area of Work Life Scale (AWLS) subscales Variable  B  Standardised  R2  Adjusted R2  R2 Change  F value (change)  p-value (change)  Model 1  –  –  0.06  0.05  0.06  5.02  0.008   Age  0.22  0.21**             Gender  4.98  0.18*            Model 2  –  –  0.21  0.16  0.15  3.61  0.001   Age  0.18  0.17             Gender  3.36  0.12             Experience AYE (reference)  –  –              1–2 years  3.23  0.13              2–3 years  2.01  0.07              3–4 years  2.93  0.11              4–5 years  12.07  0.35**              5+ years  9.92  0.48**             Qualifications CQSW/Diploma (reference)  –  –             DipHE social work  8.79  0.18*             Degree social work  8.70  0.41**             Master’s social work  4.77  0.15            Model 3  –  –  0.27  0.22  0.06  6.55  0.002   Age  0.16  0.15             Gender  2.50  0.09             Experience AYE (reference)  –  –              1–2 years  3.07  0.12              2–3 years  2.05  0.07              3–4 years  .53  0.02              4–5 years  11.02  0.32**              5+ years  9.41  0.46**             Qualifications CQSW/Diploma (reference)  –  –              DipHE social work  9.29  0.19*              Degree social work  8.49  0.40**              Master’s social work  4.68  0.15             AWLS subscales                  Control  2.12  0.17*              Values  2.47  0.16*            Variable  B  Standardised  R2  Adjusted R2  R2 Change  F value (change)  p-value (change)  Model 1  –  –  0.06  0.05  0.06  5.02  0.008   Age  0.22  0.21**             Gender  4.98  0.18*            Model 2  –  –  0.21  0.16  0.15  3.61  0.001   Age  0.18  0.17             Gender  3.36  0.12             Experience AYE (reference)  –  –              1–2 years  3.23  0.13              2–3 years  2.01  0.07              3–4 years  2.93  0.11              4–5 years  12.07  0.35**              5+ years  9.92  0.48**             Qualifications CQSW/Diploma (reference)  –  –             DipHE social work  8.79  0.18*             Degree social work  8.70  0.41**             Master’s social work  4.77  0.15            Model 3  –  –  0.27  0.22  0.06  6.55  0.002   Age  0.16  0.15             Gender  2.50  0.09             Experience AYE (reference)  –  –              1–2 years  3.07  0.12              2–3 years  2.05  0.07              3–4 years  .53  0.02              4–5 years  11.02  0.32**              5+ years  9.41  0.46**             Qualifications CQSW/Diploma (reference)  –  –              DipHE social work  9.29  0.19*              Degree social work  8.49  0.40**              Master’s social work  4.68  0.15             AWLS subscales                  Control  2.12  0.17*              Values  2.47  0.16*            * p < 0.05; ** p < 0.01 (two-tailed); AYE, Assessed Year in Employment. Limitations of the study This study is limited by the self-reported nature of the survey and the potential motivation to participate being disproportionally drawn from ‘unhappy’ employees. This has the potential of skewing the data; however, the balance of findings reflects both positive and negative experiences. Therefore, it does not appear that the issue of negative skewedness has been evidenced. The number of variables in the regression model could also be considered a limitation; however, it was important to analyse the potential predictive value of factors contained in the original survey questionnaire. The final multiple regression model reports only on significant results, as shown in Table 3. Age is the only exception to this and was left in the final model as a control variable. The response rate of 43 per cent could also be considered a limitation; however, due to the level of work pressure for social workers, including time capacity regarding work-load, this response rate was considered an achievement. Discussion The aim of the paper was to establish the factors that support child-protection social worker resilience within challenging economic, political and social contexts. It is important to acknowledge that, although the paper is reporting on resilience factors, this is balanced with inclusion of organisational factors and the acknowledgement of the wider political issues that creates the context for social work. The results of the AWLS evidence concerning levels of pressures, particularly about work-load, which have been repeatedly evidenced in research literature. The context of a demanding work-load and public, media and political scrutiny contributes to a context of fear about professional error or omission. Social work is often slated as a failing profession when there is a child death, despite the multidisciplinary nature of child protection. If there is a dominant perspective to ‘adapt’ as Evans and Reid (2013) caution, concerns are justified about how this is likely to impact on staff and ultimately on the service to vulnerable children and families. Previous analysis of the data studied in the current paper (McFadden et al., 2018), found resilience to be a mediator variable between organisational factors and burnout; however, only 7 per cent of the variance was explained by an extension of the two-process model (McFadden et al., 2018). The results presented in the current paper help to explain the relationship self-reported between resilience and burnout more fully by including demographic information and job characteristics as predictor variables of resilience. An important finding in this study is the relationship between emotional exhaustion and resilience scores. Emotional exhaustion is the feeling ‘of being over-extended from one’s job’ (Maslach and Jackson, 1986, p. 13) and has been found to be correlated with work-load in a number of studies (Maslach and Leiter, 2008; McFadden et al., 2018). Emotional exhaustion is accepted as a work-related issue of increasing concern in social work (DePanfilis and Zlotnik, 2008; Aarons, 2009). Levels of work-load were not found to be related to higher levels of emotional exhaustion in a UK-wide study (McFadden, 2015) whereby even those with smaller workloads reported high levels of emotional exhaustion. This suggests that further research on the nature of social work workloads is required to further understand this phenomenon. The RS14 Resilience Scale (Wagnild and Young, 1993) measured resilience in relation to two broad factors of ‘acceptance of life and self’ and ‘individual competence’, and these are operationalised across the five categories of meaningfulness, equanimity, existential aloneness, perseverance and self-reliance. A significant negative relationship exists between emotional exhaustion and resilience. So, when an individual is expressing high emotional exhaustion, resilience is low. Examples of emotional exhaustion items are ‘I feel emotionally drained from my work’, ‘I feel used up at the end of the workday’, ‘I feel burned out from my work’ and ‘I feel at the end of my rope’. Conversely, resilience items measured items such as ‘I usually manage one way or another’, ‘I usually take things in my stride’, ‘I feel I can handle many things at a time’ and ‘In an emergency I am someone people can rely upon’. It is therefore theoretically sound that, as one feels more emotionally exhausted, feelings of resilience will be reduced, as evidenced in the inverted correlation presented in the correlation matrix (Table 2). Another important finding is the relationship between resilience and personal accomplishment. As resilience scores increase, so too personal accomplishment scores increase. Personal accomplishment is measured in items such as ‘I feel I am positively influencing peoples lives’, ‘I can easily create a relaxed atmosphere with my recipients’, ‘I have accomplished many worth-while things in this job’ and ‘I feel exhilarated when working closely with my recipients’. One can understand the congruence in resilience and personal accomplishment scores and how these factors are related, and this is evidenced in the correlation matrix (Table 2). Stalker et al. (2007), in a systematic review, found similar results, whereby social workers had reported high levels of job satisfaction in the context of high emotional exhaustion. Analysis of personal accomplishment scores may help to theoretically explain this phenomenon through a resilience theory lens. The items that measure personal accomplishment in the MBI refer to feelings of accomplishment and meaning applied to the job and the impact on service users’ lives, which is congruent with the broad theoretical resilience domains advocated by Wagnild and Young (1993) with regard to ‘meaning’. Meaning making is confirmed as an important aspect of resilience across cultures and has been applied to many studies relating to populations achieving positive adjustment despite adversity and trauma experiences (Theron and Theron, 2014). In Model 1, females were reported as having higher resilience scores than males; however, when experience was introduced to Model 2, gender was no longer significant. Similarly, age was significant in Model 1 but was no longer significant when experience was introduced. As Model 2 included experience and qualifications, those with four to five years’ experience and those with more than five years’ experience had higher mean resilience scores than those in their AYE. So experience tended to significantly impact on higher levels of resilience, which is evidence to persuade employers and influence policy to maintain workforce retention and avoid turnover. Organisational-level interventions to improve job engagement could, for example, include supervisory support (McFadden, 2015), mentoring schemes, counselling and organisational consultancy (Schaufelli, 2009, p. 205). Qualifications were also important in the findings, as those who reported having a degree or a master’s-level qualification (versus CQSW only) had higher resilience scores. Other research has reported important findings about age, experience and qualifications. Goldberg (2004) found age to be a significant factor in retention and showed that workers under thirty-five years old, those who held a graduate degree and those employed for less than two years were more likely to intend to leave than their counterparts. Post-qualifying training at master’s level (Post Qualifying Award 2–6 and Specialist Award are levels of previous post-qualifying framework in Northern Ireland that are based on up to 120 academic credit points) was found to be a significant predictor of resilience. Other studies on training impact on staff (Curry et al., 2005: Turcotte et al., 2009) confirm this finding. The non-significant levels of post-qualifying training in the findings require further exploration, as there is no obvious explanation for this result. Conclusion Child-protection workers are particularly vulnerable to burnout (Anderson, 2000; Tham and Meagher, 2009); nevertheless, there is evidence of job satisfaction that has been linked to consumer satisfaction and client outcomes (Bednar, 2003), suggesting levels of resilience in this population. The link to ‘client outcomes’ is an important aspect of child protection, as the role means that, when a child fatality happens, social workers are held publicly accountable for not ‘preventing’ the incident. Conversely, positive outcomes for children are rarely reported due to the confidential nature of cases. The resultant political and media pressure on the social work profession creates fear in front line workers that can result in a change in their practice due to the fear of ‘missing’ signs of risk in families (Smith et al., 2003; Stanley, 2013). This paper importantly presents new evidence relating to the significance of holding experience in teams, retaining expertise in child protection and preventing unwanted staff turnover. We have found this may be achieved by making organisational-level adjustments to ensure social workers are supported to manage the challenges of the job. We know that having a realistic work-load and organisational supports such as a positive manager and co-worker relationships and a sense of control, fairness, values and reward are important for retaining staff and reducing turnover intent. However, the modernisation agenda, as discussed by White et al. (2010), in the context of New Labour and continuing to progress under the current Tory government, provides a challenging socio-economic and political landscape for social work, with excessive caseloads and high levels of risk to children being balanced in this context. The results in this paper support an argument for employers and policy makers to revisit the evidence regarding the risk of burnout in child protection and take positive action to support a more resilient workforce. This could be achieved by implementing protective processes in organisational policy to ensure this outcome is achieved in this critical area of social work practice. Funding Department of Education and Learning, Northern Ireland, 2009. Conflict of interest statement. None declared. References Aarons G. A., Fettes D. L., Flores L. E., Sommerfeld D. H. ( 2009) ‘ Evidence-based practice implementation and staff emotional exhaustion in children's services’, Behaviour Research and Therapy , 47( 11), pp. 954– 60. Google Scholar CrossRef Search ADS PubMed  Anderson D. G. ( 2000) ‘ Coping strategies and burnout among veteran child protection workers’, Child Abuse and Neglect , 24( 6), pp. 839– 48. Google Scholar CrossRef Search ADS PubMed  Austin G., Royse D., Kulver K., Plescher K., Zhang Y. ( 2017) ‘ Who stays, who goes, who knows?: A state-wide survey of child welfare workers’, Children and Youth Services Review , 77, pp. 110– 17. Google Scholar CrossRef Search ADS   Bednar S. ( 2003) ‘ Elements of satisfying organizational climates in child welfare agencies’, Families in Society , 84( 1), pp. 7– 12. Google Scholar CrossRef Search ADS   Belsky J., Pluess M. ( 2009) ‘ The nature (and nurture?) of plasticity in early human development’, Perspectives on Psychological Science , 4( 4), pp. 345– 51. Google Scholar CrossRef Search ADS PubMed  Bride B. E., Radey M., Figley C. R. ( 2007) ‘ Measuring compassion fatigue’, Clinical Social Work Journal , 35( 3), pp. 155– 63. Google Scholar CrossRef Search ADS   Bronfenbrenner U. ( 1979) The Ecology of Human Development: Experiments by Nature and Design , Cambridge, MA, Harvard University Press. Burns K. ( 2011) ‘ Career preference, transients and converts: A study of social workers’ retention in child protection and welfare’, British Journal of Social Work , 41( 3), pp. 520– 38. Google Scholar CrossRef Search ADS   Cohen J. W. ( 1988) Statistical Power Analysis for the Behavioral Science , 2nd edn, Hillsdale, NJ, Lawrence Erlbaum Associates. Curry D., McCarragher T., Dellmann-Jenkins M. ( 2005) ‘ Training, transfer, and turnover: Exploring the relationship among transfer of learning factors and staff retention in child welfare’, Children and Youth Services Review , 27( 8), pp. 931– 48. Google Scholar CrossRef Search ADS   Curtis L., Moriarity J., Netton A. ( 2010) ‘ The expected working life of a social worker’, British Journal of Social Work , 40( 5), pp. 1628– 43. Google Scholar CrossRef Search ADS   DePanfilis D., Zlotnik J. L. ( 2008) ‘ Retention of front-line staff in child welfare: A systematic review of research’, Children and Youth Services Review , 30( 9), pp. 995– 1008. Google Scholar CrossRef Search ADS   DeVellis R. F. ( 2003) Scale Development: Theory and Applications , 2nd edn, Thousand Oaks, CA, Sage, 171 pages. Ellett A. J. ( 2009) ‘ Intentions to remain employed in child welfare: The role of human caring, self-efficacy beliefs, and professional organizational culture’, Children and Youth Services Review , 31( 1), pp. 78– 88. Google Scholar CrossRef Search ADS   Evans B., Reid J. ( 2013) ‘ Dangerously exposed: The life and death of the resilient subject’, Resilience: International Policies, Practices and Discourses , 1( 2), pp. 83– 98. Faul F., Erdfelder E., Lang A.-G., Buchner A. ( 2007) ‘ A flexibile statistical power analysis program for the social and behavioral sciences’, Behavior Research Methods , 39( 2), pp. 175– 91. Google Scholar CrossRef Search ADS PubMed  Freudenberger H. J. ( 1974) ‘ Staff burnout’, Journal of Social Issues , 30( 1), pp. 159– 65. Google Scholar CrossRef Search ADS   Garrett P. M. ( 2016) ‘ Questioning tales of “‘ordinary magic”: “Resilience” and neo-liberal reasoning’, British Journal of Social Work , 46( 7), pp. 1909– 25. Google Scholar CrossRef Search ADS   Goldberg Levin A. ( 2004) ‘ Professions at risk: Why are so many workers leaving the field of child welfare? Exploring the relationship between diversity, inclusion, supervisory support, stress, job satisfaction and intention to leave among public child welfare workers’, Dissertation Abstracts International, A: The Humanities and Social Sciences , 64( 12), p. 4624. Hall P. A., Lamont M. (eds) ( 2013) Social Resilience in the Neoliberal Era , Cambridge, Cambridge University Press. Google Scholar CrossRef Search ADS   Hart A., Gagnon E., Eryigit-Madzwamuse S., Cameron J., Aranda K., Rathbone A., Heaver B. ( 2016) ‘ Uniting resilience research and practice with an inequalities approach’, SAGE Open , 6( 4), pp. 1– 15. Google Scholar CrossRef Search ADS   Healy K., Meagher G., Cullin J. ( 2007) ‘ Retaining novices to become expert child protection practitioners: Creating career pathways in direct practice’, British Journal of Social Work , 39( 2), pp. 299– 317. Google Scholar CrossRef Search ADS   Leiter M., Maslach C. ( 2006) The Area of Work Life Survey Manual , 4th edn, Nova Scotia, Canada, Centre for Organizational Research and Development, Acadia University. Maslach C., Jackson S. ( 1986) Maslach Burnout Inventory Manual , 2nd edn, California, CPP, Inc. Maslach C., Leiter M. ( 2008) ‘ Early predictors of job burnout and engagement’, Journal of Applied Psychology , 93( 3), pp. 498– 512. Google Scholar CrossRef Search ADS PubMed  Maslach C., Leiter M. P. ( 2005) ‘Stress and burnout: The critical research’, in Cooper C. L. (ed.), Handbook of Stress Medicine and Health , 2nd edn, London, CRC Press, pp. 153– 70. Maslach C., Jackson S., Leiter M. ( 1996) Maslach Burnout Inventory Manual , 3rd edn, California, CPP, Inc. Masten A. S., Wright M. O. ( 2010) ‘Resilience over the lifespan: Developmental perspectives on resistance, recovery and transformation’, in Reich J. W., Zautra A. J., Hall J. S. (eds), Handbook of Adult Resilience , California, New York, The Guilford Press, pp. 213– 37. McEwen B. S., Gianaros P. J. ( 2010) ‘ Central role of the brain in stress and adaptation: Links to socioeconomic status, health, and disease’, Annals of the New York Academy of Sciences , 1186( 1), pp. 190– 222. Google Scholar CrossRef Search ADS PubMed  McFadden P. ( 2013) ‘Resilience and burnout in child protection social work’, PhD thesis, University of Ulster, Northern Ireland. McFadden P. ( 2015) ‘Measuring burnout among UK social workers: A Community Care study’, available online at: https://s3-eu-west-1.amazonaws.com/rbi-communities/wp-content/uploads/sites/7/2015/07/Burnout-among-UK-social-workers.pdf (accessed on 1 July 2015). McFadden P., Campbell A., Taylor B. ( 2014) ‘ Resilience and burnout in child protection social work: Individual and organisational themes from a systematic literature review’, British Journal of Social Work , 45( 5), pp. 1546– 63. Google Scholar CrossRef Search ADS   McFadden P., Mallett J., Leiter M. ( 2018) ‘ An extension of the two process model of burnout: The role of resilience, reward and community relationships in child protection social work’, Stress & Health: Journal of the International Society for the Investigation of Stress , 34( 1), pp. 72– 83. Google Scholar CrossRef Search ADS   Mor Barak M. E., Nissly J. A., Levin A. ( 2001) ‘ Antecedents to retention and turnover among child welfare, social work, and other human service employees: What can we learn from past research: A review and meta-analysis?’, Social Service Review , 75( 4), pp. 625– 61. Google Scholar CrossRef Search ADS   Munro E. ( 2011) Review of Child Protection: A Child Centered System , London, Department for Education. Nordick W. G. ( 2002) ‘Striking balance, enjoying challenge: How social workers in child protection stay on the high wire’, UBC Retrospective theses Digitization Project, available online at: http://www.Library.Ubc.ca/archives/retro_theses/ (accessed on 22 March 2018). Pallant J. ( 2007) SPSS Survival Manual , 3rd edn, Buckingham, Open University Press. Rutter M. ( 1987) ‘Psychosocial resilience and protective mechanisms’, in Rolf J., Masten A. S., Cicchetti D., Nuechterlein K. H., Weintraub S. (eds), Risk and Protective Factors in the Development of Psychopathology , NY, New York, Cambridge Press, pp. 181– 214. Schaufeli W. B., Leiter M. P., Maslach C. ( 2009) ‘ Burnout: 35 years of research and practice’, Career Development International , 14( 3), pp. 204– 20. Google Scholar CrossRef Search ADS   Smith B. D. ( 2005) ‘ Job retention in child welfare: Effects of perceived organizational support, supervisor support, and intrinsic job value’, Children and Youth Services Review , 27( 2), pp. 153– 69. Google Scholar CrossRef Search ADS   Smith M., McMahon N., Nurston J. ( 2003) ‘ Social workers’ experiences of fear’, British Journal of Social Work , 33( 5), pp. 659–7. Stalker C. A., Mandell D., Frensch K. A., Harvey C., Wright M. ( 2007) ‘ Children welfare workers who are exhausted yet satisfied with their jobs: How do they do it?’, Child and Family Social Work , 12( 2), pp. 182– 91. Google Scholar CrossRef Search ADS   Stanley T. ( 2013) ‘“ Our tariff will rise”: Risk, probabilities and child protection’, Health, Risk and Society , 15( 1), pp. 67– 83. Google Scholar CrossRef Search ADS   Strolin-Goltzman J., McCarthy M., Smith B., Caringi J., Bronstein L., Lawson H. ( 2008) ‘ Should I stay or should I go? A comparison study of intention to leave among public child welfare systems with high and low turnover rates’, Child Welfare , 87( 4), pp. 125– 43. Google Scholar PubMed  Tham P., Meagher G. ( 2009) ‘ Working in human services: How do experiences and working conditions in child welfare social work compare?’, British Journal of Social Work , 39( 5), pp. 807– 27. Google Scholar CrossRef Search ADS   Theron L., Theron A. ( 2014) ‘ Meaning-making and resilience: Case studies of a multifaceted process’, Journal of Psychology in Africa , 24( 1), pp. 24– 32. Turcotte D., Lamonde G., Beaudoin A. ( 2009) ‘ Evaluation of an in-service training program for child welfare practitioners’, Research on Social Work Practice , 19( 1), pp. 31– 41. Google Scholar CrossRef Search ADS   Ungar M. ( 2008) ‘ Resilience across cultures’, British Journal of Social Work , 38( 2), pp. 218– 35. Google Scholar CrossRef Search ADS   Ungar M. ( 2013) ‘ Resilience, trauma, context, and culture’, Journal of Trauma, Violence, and Abuse , 14( 3), pp. 255– 66. Google Scholar CrossRef Search ADS   Wagnild G. ( 2009) The Resilience Scale User’s Guide for the US English Version of the Resilience Scale and the 14 item Resilience Scale (RS-14). USA, The Resilience Centre. Wagnild G., Young H. ( 1993) ‘ Development and psychometric evaluation of the Resilience Scale’, Journal of Nursing Measurement , 1( 2), pp. 165– 7. Google Scholar PubMed  Weaver D., Chang J., Clark S., Rhee S. ( 2007) ‘ Keeping public child welfare workers on the job’, Administration in Social Work , 31( 2), pp. 5– 25. Google Scholar CrossRef Search ADS   Werner E., Smith R. ( 1982) Vulnerable but Not Invincible: A Longitudinal Study of Resilient Children and Youth , New York, McGraw Hill. White S., Wastell D., Broadhurst K., Hall C. ( 2010) ‘ When policy o’erleaps itself: The tragic tale of the integrated children’s system’, Critical Social Policy , 30( 3), p. 405– 29. Google Scholar CrossRef Search ADS   © The Author(s) 2018. Published by Oxford University Press on behalf of The British Association of Social Workers. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

The British Journal of Social WorkOxford University Press

Published: Apr 9, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off