Work Support, Role Stress, and Life Satisfaction among Chinese Social Workers: The Mediation Role of Work-Family Conflict

The current study examined the relationships among work support, role stress, work-family conflict, and life satisfaction, with a sample of social workers in China’s Pearl River Delta (N = 1414). Using structure equation modelling, the study revealed that social workers’ life satisfaction reduced because of role conflict and work-family conflicts. Work-family conflict partially mediated the negative effects of role ambiguity and conflict on social workers’ life satisfaction. Work support from their director, manager, supervisor, and co-workers protectively reduced role stress and work-family conflict. The findings emphasize the significance of managing the interference between work and family for social workers’ well-being.


Introduction
Social workers' well-being has attracted growing attention due to the contamination of their work experiences [1][2][3]. Providing emotional labor, social workers are one of the at-risk professionals in experiencing negative work-family interference with exposures of various negative work events [4][5][6][7]. Workplace experiences, such as stress and burnout [8][9][10][11][12], are likely to extend into the domain of their private life and to lead to high risks of negative work-to-family spill-over that could impair their life satisfaction [9]. While the social work literature has mainly focused on the negative pathways from social workers' stress and strains to their problematic work outcomes [13], research on the influence of work-family conflict (WFC) on social workers' life satisfaction remains sparse in the social work context (see exceptions of [9,[13][14][15]).
Like the global situation, social workers in China also confront risks to their quality of life, which has been scarcely studied. With the rapid development of social work in Mainland China in the past three decades, social workers face high workplace stress [16], experience high turnover [17], and low job satisfaction [18]. Besides the common risk factors (see a meta-analysis of [19]), the recent transitions occurring in community-level social work organizations in China's Pearl River Delta (PRD) are creating work stress. As government procurement in social services led the nationwide development in Chinese social work starting from 2006 in PRD, Shenzhen, a major PRD city, has initiated a community-level administrative transition in the past four years. Social work services are integrating with community councils (the primary administrative unit of governance). These integrated service centers require community social workers to cooperate with local administrators in managing community affairs. turnover intention [17]. Their results suggested WFE significantly influence job satisfaction but failed to support WFE mediating between job autonomy and turnover intention. Therefore, the mechanism of role stress, WFC, and LS needs further clarification.
Based on previous studies reviewed, we hypothesize WFC mediates between role stress and LS among social workers: Hypothesis 1 (H1). That role stress (i.e., role ambiguity, role conflict, and role overload) positively predicts WFC (i.e., WIF and FIW) and negatively predicts LS.

The Protective Role of Work Support
Work support, is defined as psychological or material resources provided through social relationships that can mitigate work stress or strain at workplace [58]. Based on the main effect hypothesis of social support on stress [59,60], social support serves to manage stress, mitigate strain, and maintain well-being. That is, supportive relationships protect one from stress or strain, and thus enhance one's LS [61]. The literature demonstrated that work support eases work stress [39,62,63] and WFC [64,65]. For example, work support by supervisors and coworkers helps to reduce the role stress caused by overload, and leads to less WFC [66]. Further, meta-analyses have revealed that social support directly and indirectly improves one's LS [58,67].
In the social work domain, there has been a lack of clarity on how the mechanism of social support influence WFC and LS. By mitigating social workers' stress or strain [11,29], social support protects social workers from negative work experiences, such as burnout [5,68] and promotes subjective well-being [9]. Of the few studies that considered WFC, one revealed that social workers took advantage of work support to cope with WFC [15]. Lizano and Mor Barak [69] revealed that work support moderated the influence of role stress and work-family conflict on burnout and job satisfaction among social welfare workers. Kalliath and colleagues found that at a higher level of family support, WFE promoted Indian social workers' work well-being and satisfaction [70]. In the Chinese context, possible cultural differences of the perceived availability of social support [71] or appropriateness of seeking social support [72] might appear. However, how social support protects the impact of WFC and role stress on LS were largely uncharted, especially in the Chinese social work context. Based on previous reviewed studies, we further propose the following hypotheses: Hypothesis 4 (H4). Work support from one's director, direct manager, supervisor, and co-workers, negatively predicts role stress and WFC, and positively predicts LS.

Data and Sample
The current study employed a subsample from the Pearl River Delta from the first-wave data from the Chinese Social Work Longitudinal Survey conducted in 2019 (CSWLS 2019) [73,74]. CSWLS 2019 is the first national-wide survey of social work organizations and social workers. The ethical approval for CSWLS 2019 was achieved by the research ethics committee in a Chinese university [74]. The survey employed paper-pencil questionnaires with a stratified random sampling method, and surveyed social work agencies and social workers from 56 cities of 22 provinces in Mainland China [73,74]. First, a selected pool of social work organizations formed according to the Chinese Social Organization Platform. Then, 33 social work agencies were randomly selected from the capital city or municipality, while 11 social work agencies per city were randomly selected for the rest. At last, 10-30 social workers were randomly selected in every agency according to the agencies' total number of employees [74]. Informed consent was obtained [73,74].
The current study employed its PRD subsample of CSWLS2019, which including nine cities of Dongguan (DG), Foshan (FS), Guangzhou (GZ), Huizhou (HZ), Jiangmen (JM), Shenzhen (SZ), Zhaoqing (ZQ), Zhongshan (ZS), and Zhuhai (ZH) in Guangdong Province. This subsample is a representative microcosm of social work development in Mainland China. PRD is a leading area in Mainland China for social work development, as reflected in its highest number of social work agencies and qualified social workers [75]. Moreover, the organizational transition of combining social work service with community councils was widely adopted by several PRD cities such as Guangzhou and Jiangmen.

Measures
The current study employed social worker dataset in PRD from CSWLS 2019 and includes the following measures:

Outcome Variable
Life satisfaction was measured by the single-item global scale [76] with "how satisfied are you with your life as a whole?" in C16 section of the questionnaire. It was originally measured by a 5-point rating scale from "Very Satisfied" = 1 to "Very Unsatisfied" = 5. We transformed to a reverse-coding according to the global scale [76]. A higher score presents a greater degree of LS.

Independent Variable
Work support was measured by the validated Chinese version of House and Wells's scale (I section) [77,78]. Four types of work support sources were asked about, including agency director, direct manager, social workers' supervisor, and co-workers. It was measured on a 5-point scale from "Very Unsatisfied" = 1 to "Very Satisfied" = 5. A composite of these four sources had higher scores representing greater perceived work support. The reliability (Cronbach's α) ranged from 0.89 to 0.92.

Mediator
Role stress was measured by role ambiguity and conflict, as well as role overload. The validated Chinese version of 6-item role ambiguity and 8-item conflict scale [79,80] were used to measure role stress (H section). The responses ranged on a 5-point Likert scale of agreement from "Strongly Disagree" = 0 to "Strongly Agree" = 4. The role ambiguity and conflict subscales yielded good reliability (Cronbach's α = 0.73 and 0.76, respectively). Followed Rapoport and Rapoport [81], role overload was indicated by time-based work overload with the question "In general, how much overtime work do you do?" (D28D). The response of "None" was coded as "0," whereas all other options of working overtime on workdays, weekends and holidays were coded a "1." Work-family conflict was measured by the validated Chinese version of Work-Family Balance Scale in S section of the survey [82,83]. The responses ranged on a 5-point agreement scale from "Strongly Disagree" = 0 to "Strongly Agree" = 4. Two subscales of 3-item WIF and 4-item FIW yielded good reliability (Cronbach's α = 0.82 and 0.86). The composed scores of WIF and FIW were used, and higher scores represented greater conflicts in work-to-family interference and family-to-work interference, respectively.

Controlled Covariates
Job tenure refers to years of working as a social worker. As a longer tenure presents accumulated resources in the workplace [84], it was controlled. We transformed it from the date format (D3A) into years of tenure.
Social work qualifications (B7) was a control factor because such professional qualifications may lead to differences in work contents or coping strategies [15]. The three levels of social work qualifications in Mainland China were: Assistant, junior, and senior level. As the senior qualification began in 2019, only two levels existed in the dataset. The qualifications of assistant and junior social workers were dummy-coded, respectively.

Demographic Characteristics
As marital status, gender, age, and years of education influence WFC and LS, these characteristics are used for control purposes [3,39]. As divorced or widowed cases were limited, marital status (A9) was dummy-coded into "married" = 1, and "single/divorced/widowed" = 0. Gender (A1) was coded as "female" = 1 and "male" = 0. Age (A2) was transferred from birth date to years of age. Education (B1) was recoded into years of education for 9-19 years.

Data analysis
Regarding missing data, as CSWLS 2019 employed return visits to re-fulfil the data, the proportions of missing data were smaller than 4.60% among all variables and thus the following analysis was performed by default. Zero-order correlations and structure equation modelling (SEM) with maximum likelihood (ML) estimation were conducted with Stata 15 (StataCorp LLC, College Station, TX, USA) for hypotheses testing. A mediation model with multiple mediators was properly tested with SEM [85]. The indirect effect of mediation was tested through the Monte Carlo method, as it was preferred for SEM [86]. As the current study aimed to understand the protective role of general work support than the specification of sources, work support from the director, manager, supervisor, and co-workers was modelled as a latent variable to capture the shared variance of work support [87]. According to the hypotheses (Figure 1), work support was estimated to predict role stress (i.e., role ambiguity, role conflict, and role overload), WFC (i.e., WIF and FIW), and LS. Role stress was estimated to mediate between work support and WFC, as well as between work support and LS. WFC was estimated to mediate between work support and LS, as well as between stress and LS. Controlled variables were estimated as covariates for independent, mediating, and outcome variables. As marital status, gender, age, and years of education influence WFC and LS, these characteristics are used for control purposes [3,39]. As divorced or widowed cases were limited, marital status (A9) was dummy-coded into "married" = 1, and "single/divorced/widowed" = 0. Gender (A1) was coded as "female" = 1 and "male" = 0. Age (A2) was transferred from birth date to years of age. Education (B1) was recoded into years of education for 9-19 years.

Data analysis
Regarding missing data, as CSWLS 2019 employed return visits to re-fulfil the data, the proportions of missing data were smaller than 4.60% among all variables and thus the following analysis was performed by default. Zero-order correlations and structure equation modelling (SEM) with maximum likelihood (ML) estimation were conducted with Stata 15 (StataCorp LLC, College Station, TX, USA) for hypotheses testing. A mediation model with multiple mediators was properly tested with SEM [84]. The indirect effect of mediation was tested through the Monte Carlo method, as it was preferred for SEM [85]. As the current study aimed to understand the protective role of general work support than the specification of sources, work support from the director, manager, supervisor, and co-workers was modelled as a latent variable to capture the shared variance of work support [86]. According to the hypotheses (Figure 1), work support was estimated to predict role stress (i.e., role ambiguity, role conflict, and role overload), WFC (i.e., WIF and FIW), and LS. Role stress was estimated to mediate between work support and WFC, as well as between work support and LS. WFC was estimated to mediate between work support and LS, as well as between stress and LS. Controlled variables were estimated as covariates for independent, mediating, and outcome variables.

Results
The demographic characteristics of the PRD subsample in CSWLS 2019 are presented in Table  1. Although the response rate was not available according to the survey report [73], the average age of the PRD subsample was 29.21 (see Table 1). This statistic was highly close to the average age of 29.37 reported by the Social Work Association of Shenzhen city [73], which partially supported the sample representativeness.

Discussion
The current study examined the pathways among work support, role stress, WFC, and LS, with a representative sample of social workers from nine cities in PRD, China. To our knowledge, this is the first study to preliminarily explore social workers' LS in relation to social support, WFC and role stress with quantitative method, especially in a Chinese context. It revealed that Chinese social workers' LS reduced directly because of role conflict and WFC (i.e., WIF and FIW). Further, role ambiguity and conflict reduced LS indirectly through the mediation of WFC. Work support, as a protective factor, directly reduced role ambiguity, role conflict, and WFC. Moreover, work support indirectly benefited LS through lessening the negative influence of role ambiguity and conflict.
Work-family interference is noteworthy regarding social workers' quality of life, especially among female workers. Coinciding with Yucel and Minnotte [46], we indicated that WFC directly impacted LS and aggravated the negative influence of role stress on LS. It contributes and adds to the literature of negative work-family spill-over perspective [35], by linking social workers' WFC, role stress, and life quality [13,90]. Regarding the individualism-collectivism spectrum, compared with studies conducted in an individualism cultural setting [43], the current study appears not evident that WFC impairs LS to a greater extent [24]. Gender difference in WFC is worth noted that, we indicated that female social workers experience significantly greater WFC than did males (see Table 3), which was also in line with Baum's findings [14].
Second, social workers' role stress of role ambiguity and conflict at workplace were highlighted. The current study elaborates on how work stress worsened WFC and adds to the occupational health literature of social workers, especially in a Chinese cultural context [17]. As role stress is commonly reported among social workers [65,91], we highlighted that, role ambiguity influences WFC more than other variables, whereas role conflict was the only role stress factor that facilitated WFC and spontaneously reduced one's LS, which echoes Michel and colleagues' results of role conflict as the factor most associated with WFC [65]. It might result from the previous mentioned community-level administrative transition at PRD, which needs further exploration. In addition, we indicated as age increased, social workers face significantly less role overload, which possibly because of experience or skills (e.g., resilience) accumulated [92]. Furthermore, we found that married status reduced role ambiguity and conflict, which possibly stems from family support in promoting job involvement [93].
Third, the protective role of work support regarding social workers' LS is evident. Work support from the director, direct manager, supervisor, and co-workers directly protects social workers from role stress, and mildly while indirectly promotes LS. Echoing the main effect hypothesis of social support on stress [59], we indicated that Chinese social workers employ a certain degree of resources derived from workplace social interactions in coping with role stress, which is also in line with findings in other cultural contexts [15]. Additionally, our result suggests that social workers with less education received more work support. A possible reason is that job mentoring is likely to happen among those with less professional education.
In addition, the current study indicates that social work job qualifications matter. Compared with those with no qualifications, assistant social workers faced greater role overload and WFC (Table 3). However, this effect did not present among junior social workers. Possible explanations of work tenure [14,94] and professional skills [95,96] are worth further research.
The limitations of this study are noteworthy. First, we only focused on workplace antecedents, and did not distinguish between the types of WFC [13]. Future study may expand to look at stress and support in the family domain and elaborate WFC. Moreover, as the current study rested on the main hypothesis of social support on stress, future research may elaborate the buffering hypothesis about the effect of social support on stress [57], and the sources or alternative protective mechanisms of work support [58]. Second, the dataset of CSWLS 2019 rests on a cross-sectional design of its first wave. Causal sequences need further exploration. For example, WFC was alternatively defined as a work-related stressor rather than a strain [35]. Longitudinal data, including those from the second or the third wave, are imperative for future work. Third, further cross-cultural comparisons are worthy of research in future.

Conclusions
The current study highlights the negative impact of role conflict and WFC in reducing Chinese social workers' LS. In addition, work support from the agency director, direct manager, supervisor, and co-workers protects social workers' LS by lessening role conflicts. Social workers' length of work, professional qualification, gender, and marital status are significantly associated with role conflict, WFC, and LS. It has several practical implications. First, the administrative practice of work support would be worth pursuing. As work support's protective role in reducing social workers' work stress is evident by the current study and others [59], administrative encouragement of social interactive collaboration among the workplace would optimize social workers' performance. Second, possible administrative support handling work-family interference among female social workers would be recommended. As female social workers face greater WFC than do males in the current sample and general population [97,98], possible policy regarding parenting support or alternative work schedules is worth considering to improve female social workers' work-family balance and well-being. Finally, interventions into social workers' work-related stress or strain are advisable. As social workers experienced negative work-family spill-over, some effective interventions such as resilience [96] and mindfulness [99] programs would enhance social workers' stress management, self-care and self-compassion, and, thus, benefit their well-being.