Professional identity and turnover intentions of social workers in Beijing,China: The roles of job satisfaction and agency type

Abstract

In order to address the ascending turnover rate among social workers in China, this article presents the idea, based on social identity theory, that professional identity is an important factor influencing turnover rate. Employing structural equation modelling (SEM) techniques, the study found that professional identity was negatively associated with social workers’ turnover intentions, and that social workers’ job satisfaction mediated the relationship between professional identity and their turnover intentions. The findings suggest that more work needs to be done on social work education and public image construction in China.

Keywords

Job satisfaction professional identity social identity theory social worker turnover intention

Introduction

Social work in China is an emerging profession but turnover rates for social workers are increasing rapidly. The rate of social workers leaving their jobs in Shenzhen, the leading city in social work development in China, has climbed from 8.2 percent in 2008 to 22.2 percent in 2014 (Du, 2015). Statistics regarding other cities are unavailable, but local government officers suggested that about 25 percent of Beijing social workers had left their positions in 2014 (Zhang, 2015), and over 50 percent of frontline social workers in Dong Guan had been considering quitting in the next 6 months (Tian and Jing, 2014). On the other hand, there are approximately 30,000 college social work graduates each year; however, most do not pursue careers in social work. These graduates leave the profession before trying it, and so many non-government organizations (NGOs) providing social work services face the double challenge of recruitment and retention. Social work in China is encountering a crisis regarding workforce shortages threatening development of the profession, and there is a call for more research into the antecedents and interventions related to recruitment and retention.

Worldwide, high workforce turnover in the social work profession is a long-standing concern in many developed countries. Social worker turnover rates perplex social service organizations. Research on social workers’ turnover rates and turnover intentions is thriving, and antecedents and predictors have been identified (Mor Barak et al., 2001; Shim, 2014; Tham, 2007). A meta-analysis of articles on turnover rates within the social work profession showed that the strongest predictors of intention to leave are burnout, lack of job satisfaction, availability of employment alternatives, low organization and professional commitment, stress and lack of social support, which implies that organizational and job-based factors have more influence on turnover than individual factors (Mor Barak et al., 2001).

In the Chinese context these factors are important, but two factors are especially critical: professional identity and job satisfaction. The social work profession is new in China so there is a lack of indigenous theory and skills and a lack of public awareness and acceptance of the field (Gao and Yan, 2015). According to social identity theory, when a public image of social work has not yet been established, social workers’ professional identity is unstable and likely to influence turnover intentions (Ellemers et al., 1999). One of the key facets of the retention crisis among social workers in China is the underdevelopment of a professional identity. Additionally, due to the disadvantaged status and early development stage of the social work profession in China, NGOs providing social work services do not have the ability to pay well or the management capability to create a positive organizational climate, which is believed to be an important factor in regard to turnover (Shim, 2014). The Shenzhen Social Work Association indicated that there were two reasons why social workers left their jobs in Shenzhen: one is lack of a sense of identity and the other is poor remuneration (Shen and Jiang, 2014). These two problems are repeatedly emphasized when discussing the development of social work in China. Therefore, to address high turnover rates, the authors focus on the effects of identity and job characteristics on retention, including the interacting effect of these two aspects in China.

Previous research in social psychology has indicated identification and job satisfaction as important factors in turnover intention (De Moura et al., 2009; Van Dick et al., 2004), but few studies have considered the two factors together as independent variables and examined their compound effects on turnover intention. The purpose of this study is to examine both the main and interaction effects of professional identity and job satisfaction on turnover intention.

Literature review

Social identity theory, professional identity and turnover intention

There can be both intrinsic and extrinsic motivations for work, the former based on the meaning of the work itself and the latter concerned with the benefits derived from it. Intrinsic motivations of social workers can be derived from their identification with the profession; extrinsic motivations relate to their conditions of work. If a social worker cannot derive either or both of these sources of job satisfaction, there could be problems regarding retention.

Deaux (2000: 746–747) argued that identity can be a source of motivation because individuals may choose to identify with a particular group to satisfy a particular set of individual needs. According to Brown (2000), social identity theory ‘proposes that people strive to achieve or maintain a positive social identity (thus boosting their self-esteem) and that this positive identity derives largely from favourable comparisons that can be made between the ingroup and relevant outgroups’.

Given that professional identity is one aspect of a person’s social identity (Miller and Garran, 2008), identification with a profession can enhance a person’s motivation to stay, especially if that person can achieve a positive professional identity in comparison to other professions. This indicates that professional identity, as an intrinsic motivation, can decrease turnover intention.

Image problems and turnover intention

Once a positive professional identity has been formed, social workers become more committed to their profession and present a lower turnover intention (Haslam, 2004). However, if comparison with other professions is unfavourable, the social worker’s professional identity will be damaged. Unfortunately, it is a persistent challenge for the social work profession to maintain a positive public image. Therefore, professional identity is a more salient factor for social workers than for other practitioners.

Worldwide, the status and public image of social work are still in flux. Since its inception as a profession, social workers have been belittled as ‘semi-professionals’ (Etzioni, 1969), as social work cannot adequately meet the traits or criteria used to define ‘full’ or ‘prestigious’ professionals, such as lawyers and doctors. Even in developed economies, social workers are still responding to ‘image problems’ (Singer, 2013).

Leung and Wong’s (2005) review of studies from Western countries (Berry, 2004; LeCroy and Stinson, 2004) shows that if the public has a negative view of the profession, it is difficult for social work to gain support for its service programme and to recruit and retain staff.

The Chinese context

Since the opening up policy in China in the 1980s, there have been escalating social problems characteristic of an industrializing society. The emergence of problems such as family breakdown, mental health issues and criminal offences has urged the Chinese government to promote social welfare (Gao and Yan, 2015; Leung, 2011). Recognizing its limitations, the Chinese government launched the policy direction of socialization of social welfare in the mid-1990s to promote NGOs to share the burden of welfare provision. It also provided impetus for the proliferation of formal social work training in universities. In the first decade of the 21st century, there are already some 320 Bachelor of Social Work (BSW) and 104 Master of Social Work (MSW) programmes, producing approximately 30,000 BSW and 2000 MSW graduates yearly (Ministry of Civil Affairs, 2014).

Given the late development of social work, the general public in China is not well informed on the nature of social work practice. The wide misconception is that social workers are volunteers. Although there have been efforts by social work academics and practitioners to set up professional organizations and registration systems, the status of the profession and social workers as professionals has not yet been fully recognized. This lack of recognition is reflected in the relatively low pay received by social workers. According to our data, the average monthly salary of social workers in Beijing is RMB2933,¹ with the average salary and the minimum salary in 2014 being RMB6463 and RMB1560 respectively (Beijing Statistical Information Net, 2015). Even in Shenzhen, a social worker’s monthly salary is between RMB4000 and RMB5000, much lower than other professions.

According to our survey, a majority of social work practitioners in Beijing do not have a formal education qualification (only 25% of respondents received formal social work education). And this is probably the case in many other cities in China. But even for those holding social work educational qualifications, there may be some who lack professional identification, which is related to the university entrance examination system. Social work programmes fare less well compared to other programmes such as business, medicine and engineering in terms of enrolment numbers and applicants’ public examination results. As a result, there has been a ‘sifting’ process, through which those with lower scores are compelled to take social work programmes that do not require high entrance scores. Consequently, students enroling in social work programmes may not be motivated to study or pursue a career in social work.

According to Tajfel and Turner’s (1979) work on social identity theory and the three cognitive processes that are involved in social identity formation, social work practitioners may need to categorize and identify themselves as belonging to the social work community based on recognition and acceptance by others. As the social work profession and the quality of social work education are still underdeveloped in China, it is worrisome whether practitioners successfully and solidly categorize themselves as belonging to the group. As for the more challenging process of social comparison, social workers may compare themselves adversely to their counterparts in other occupations, in terms of social status, prestige and pay. The disadvantaged position of social workers denies their possible acquisition of self-esteem, a crucial element in the social comparison process of establishing their professional identity. Furthermore, with reference to Burke and Reitzes’ (1991) assertion on commitment, if social workers cannot find a strong anchorage for their professional identity, they might be predisposed to have less commitment.

Job satisfaction and turnover intention

Although professional identity has to be examined to predict turnover intention, especially among social workers in China, the satisfaction derived from job conditions is another factor that cannot be ignored in an explanation of turnover intention. Job- and organization-related conditions are well-known factors when discussing staff retention in regard to social workers (Kim and Stoner, 2008; McFadden et al., 2015; Mor Barak et al., 2001; Shim, 2014; Tham, 2007).

Fuller et al. (2003) suggested that members of an organization have ‘perceived organizational support’, which refers to global beliefs about the extent to which the organization cares about their well-being and values their contributions (Eisenberger et al., 1986: 501, cited in Fuller et al., 2003: 789). Such a relationship between members and their affiliated organization is based on social exchange theory and the norm of reciprocity (Goulder, 1960, cited in Fuller et al., 2003: 789). Perceived organizational support is positively related to organizational commitment. Once again making reference to social identity theory, it has been suggested that people ‘remain loyal when they feel that their organizations … value and appreciate them’ (Tyler, 1999: 235, cited in Fuller et al., 2003: 790).

Given the possible challenges facing the social work profession, it is imperative to investigate how social workers’ professional identity and job satisfaction influence their turnover intentions.

Types of social service organizations in China

Social workers in China are employed in four types of organizations: government departments, social work organizations (SWOs), community organizations (COs) and public institutions (PIs). Due to the difficulty in distinguishing social workers from civil servants, and the small number of frontline social workers, the government sector was excluded from this survey.

As the three types of organization (SWOs, COs and PIs) with social work positions vary in terms of the welfare programmes, service targets, professional identities and job satisfaction (Xu et al., 2017), the type of organization was also used as a variable associated with turnover intention.

SWOs are the newest NGOs, formed in response to the policy of government purchase of social work services. Most employees of SWOs are young graduates with little work experience (Xu et al., 2017). According to Young’s definition of the three government–nonprofit relations – supplementary, complementary and adversarial (Young, 2000) – SWOs in Beijing are complementary to the government, which means SWOs get financed through the government’s contracting out policies and are seen as partners of the government. However, in China the situation is more complicated for SWOs because there is little trust between the government and NGOs due to rising social protests, and ‘the public also have little trust in NGOs given the legal dilemmas, political constraints, financial crises and internal problems faced by the latter’ (Xu, 2013: 316). Government interventions are massive, and crucial to the extent that Gao and Yan (2015) argue that the government has ‘made’ social work in China. Facing the constraints of government (Qi, 2004), some SWOs have had to decouple the form and real operation within the organization (Tian, 2004). Therefore, the image and working conditions of SWOs are disadvantaged compared to government-organized or quasi-governmental organizations (Xu and Ngai, 2011). NGOs and social workers in these organizations need more recognition, otherwise their professional identity will be damaged and create a high turnover.

In contrast, PIs and COs are NGOs, but strictly speaking they are both quasi-governmental organizations as they are funded by the government and function as government offices. They also have different moral resources and political capital with grassroots NGOs (Xu and Ngai, 2011). Their image and reputation are much better than those of SWOs, resulting in better recruitment and retention. Although the three types of organization are mostly funded by government, SWOs have to compete with each other to gain government contracts, whereas COs and PIs receive funding directly through government allocation.

The distinction between funding mechanisms, social status and public image among these organizations indicates two different types of organization – NGOs and quasi-government organizations – and the nature of each will influence professional identity, job satisfaction and turnover intentions of social workers in the different organizations. Therefore, organization type is examined in this article as a moderating variable.

The relationships among professional identity, job satisfaction and turnover intention

As Van Dick et al. (2004) pointed out, ‘previous research has shown that both identification and satisfaction are predictors of turnover intentions. However, their interplay in explaining and predicting such intention is less clear’ (p. 353). Their research showed that organizational identity fed into job satisfaction, which in turn predicted turnover intention. Based on social identity theory, they argued that the relationship between identification and turnover intention was mediated by job satisfaction. However, De Moura et al. (2009), also using social identity theory, found the opposite. Their data supported the notion that job satisfaction and organizational identity are both related to turnover intention, but identity mediates the relationship between job satisfaction and turnover intention.

The aforementioned research used the concept of organization identity which, distinguished from professional identity, has a more complex relationship with job satisfaction, since they both relate to organization, including job and organization conditions. Professional identity, however, which this article addresses, is formed prior to organization identity as professional training comes before organizational affiliation. Social workers’ satisfaction with their job does influence their professional identity; however, the effect that professional identity exerts on job satisfaction is stronger than the opposite direction, not only due to the sequence of formation mentioned above, but also because of the roles that intrinsic motivation and professional identity play in the relationship network, according to social identity theory. Therefore, although the interplay effect of organization identity and job satisfaction on turnover intention is difficult to confirm, it is more certain that the relationship between professional identity and turnover intention is mediated by job satisfaction. With the assumption that professional identity is an important antecedent of the turnover intentions of social workers, this study explores the interrelationship among professional identity, job satisfaction and turnover intentions of social workers in China, with the following hypotheses.

Hypotheses

H1. Social workers’ professional identity and job satisfaction are negatively associated with turnover intentions.

H2. Social workers’ job satisfaction mediates the relationship between professional identity and turnover intentions.

H3. The type of organization moderates the relationship among professional identity, job satisfaction and turnover intention among social workers.

Methodology

Sample and procedures

Beijing was chosen as the study site based on the following considerations. Beijing presents a typical case in which there are various types of organizations that provide a wide range of sampling units for recruiting suitable respondents to the study. Three types of organization (SWOs, COs and PIs) were covered in this research, and it is through these organizations that the researchers located social workers and sent out the self-administered questionnaires to them. With the list provided by the Beijing Civil Affairs Bureau in July 2014, the sampling frames of SWOs, COs and PIs were formed with 105, 6591 and 47 organizations respectively. In order to achieve a representative sample, 101 COs were randomly selected from 6591 communities compared to the number of SWOs. Investigators were sent to all the SWOs, all the PIs and selected COs with the questionnaire. The screening criteria were such that only social work practitioners were invited to participate in the survey. The return rates from the organizations were 75 percent for SWOs, 66 percent for COs and 51 percent for PIs. By November 2014, 1345 valid questionnaires had been collected, with 500 (37.2%) coming from SWOs, 620 (46.1%) from COs and 225 (16.7%) from PIs.

Measures

Dependent variables: Turnover intention

Turnover intention is the best predictor of the real turnover rate in organizations (Mor Barak et al., 2001). Auerbach et al. (2014) suggested that in order to obtain a strong prediction of the actual turnover, the turnover intention scale should be made up of three subscales: thinking, looking and acting. Three items in this study (‘thoughts of leaving’, ‘feeling of having no future’ and ‘looking for information’) were adapted from Auerbach et al.’s (2014) scale, to examine the factors regarding thinking and looking, while the other two items (‘leaving this organization’ and ‘leaving this profession’) were newly developed, in order to reveal whether or not social workers planned to leave the social work profession, as this could be associated with the status of the profession in China.

Independent variable: Professional identity

Professional identity and job satisfaction scales were both adapted from ‘The 2013 Social Work Day (Hong Kong) Organizing Committee’s (SWDOC) Study’, which was an online survey involving 588 registered social workers in Hong Kong. The SWDOC study identified 10 statements with which to measure the professional identity of social workers through a literature review, which can be divided into three facets: (a) professionalism, (b) professional status and (c) identification with professional values (The 2013 Social Work Day (Hong Kong) Organizing Committee, 2013). The 10 items were adopted in this survey, but only the first two factors were significant in our data analysis, with names changed to (a) professionalism and (b) public image in order to illustrate the themes more clearly. The third factor, ‘identification with professional values’, was integrated into the ‘professionalism’ factor according to our data.

Mediating variable: Job satisfaction

Using the Job Descriptive Index (JDI) and Job in General (JIG) as the instruments of job satisfaction, the SWDOC study showed six indicators in the job satisfaction scale, comprising (a) work on present job, (b) pay, (c) opportunities for promotion, (d) supervision, (e) people in your present job and (f) work in general. Adapted from these six indicators, this study used six descriptions of job characteristics as the tools for job satisfaction: salary, welfare, career prospects, working content, working environment and human relationship, with the first three being basic and important factors and the last three being soft and potential factors due to the poor remuneration and image of social workers in China. A 5-point Likert scale was employed both in the professional identity scale and job satisfaction scale.

Moderating variable: Organization type

The type of organization, as a moderating variable, was examined in our study as different organizations face different challenges, indicative of the political context and policy resources of each Chinese organization. Social workers in different types of organization – NGOs and quasi-government organizations – probably have a different professional identity, level of job satisfaction and turnover intention. We used SWOs and other organizations (OOs) as the indicators’ names to examine the moderating effect of organization type on professional identity, job satisfaction and turnover intention.

Validity and reliability of the measurements

The content and face validity for the questionnaire were established by three experts from Hong Kong and Mainland China with extensive research and field experience in social work. In order to determine the reliability, the questionnaire was piloted with 10 social workers comprising five members from SWOs, two members from the community and three members from the university. The reliability of the modified instruments, assessed with Cronbach’s alpha, was 0.837 for the professional identity scale, 0.832 for the job satisfaction scale and 0.87 for the turnover intention scale.

Data analysis method

Before analysing associations between variables, it was necessary to identify the factor structures of professional identity and job satisfaction by conducting an exploratory factor analysis. After that, structural equation modelling (SEM) was used to test the hypotheses. A two-step method was utilized with both a measurement model and a structural model (Kline, 2011). Confirmatory factor analysis was first used to test the measurement model. Maximum likelihood estimation was the default estimator. Several goodness-of-fit indexes were used, including the chi-square test, comparative fit index (CFI), Tucker–Lewis index (TLI), root mean square error of approximation (RMSEA) and standardized root mean square residual (SRMR; Kline, 2011). In the second stage, a structural model was then built and tested to examine the direct and indirect relationships between latent variables and turnover intention. Amos 21 was used in the data analysis process.

Results

Sample characteristics

The characteristics of the respondents are presented in Table 1. The mean age of the respondents was 36 years. Approximately 63.6 percent of the respondents were between 20 and 39 years old. The number of females was more than double the number of males (70% and 30%, respectively). Of the respondents, 55 percent held a Bachelor’s degree or above and 65.8 percent were without a social work licence. The mean salary of the participants was RMB2933 per month. In addition, compared with those in OOs, social workers in SWOs were younger and had a better education background, less social work licences and higher salaries.

Table 1.

Sample characteristics (N = 1345).

Number	Full sample, n (%)	SWOs, n (%)	OOs, n (%)
	1345	500 (37.2)	845 (62.8)
Age (years)
20–29	491 (36.5)	342 (69.0)	149 (17.8)
30–39	364 (27.1)	104 (21.0)	260 (31.0)
40–49	282 (21.0)	37 (7.5)	245 (29.2)
50–59	173 (12.9)	10 (2.0)	163 (19.4)
60 and older	25 (1.9)	3 (0.6)	22 (2.6)
Missing data	10 (0.7)	4	6
Gender
Male	404 (30.0)	140 (28.0)	264 (31.2)
Female	941 (70.0)	360 (72.0)	581 (68.8)
Education
Below Bachelor’s	603 (44.8)	179 (35.8)	424 (50.7)
Bachelor’s	647 (48.1)	272 (54.4)	375 (44.9)
Master’s	77 (5.7)	45 (9.0)	32 (3.8)
PhD	9 (0.7)	4 (0.8)	5 (0.6)
Missing data	9 (0.7)	0	9 (0.7)
Licensed
Yes	456 (33.9)	148 (29.7)	308 (36.6)
No	885 (65.8)	351 (70.3)	534 (63.4)
Missing data	4 (0.3)	1	3
Salary
RMB2000 and below	207 (15.4)	41 (8.2)	166 (19.7)
RMB2001–RMB4000	927 (68.9)	371 (74.6)	556 (66.1)
RMB4001–RMB6000	174 (12.9)	72 (14.5)	102 (12.1)
RMB6001 and beyond	30 (2.2)	13 (2.6)	17 (2.0)
Missing data	7 (0.5)	3	4
Mean (SD)	2933 (1.191)

SWO: social work organization; OO: other organization; SD: standard deviation.

Step 1: Exploratory factor analyses

The factor structures of the professional identity and job satisfaction scales were identified by Principal Components Analysis (PCA) and Principal Axis Factoring (PAF). To determine the appropriateness of the data set with which to conduct PCA and PAF, the Kaiser–Meyer–Olkin (KMO) and the chi-square value of Bartlett’s test were calculated, indicating that the data set was adequate (KMO = 0.836, p < 0.001; KMO = 0.837, p < 0.001, respectively). PCA was used to estimate the underlying number of factors with promax rotation (factors were inter-related). The criteria used to determine the factors and their items included the following: (a) a factor has an eigenvalue greater than 1, (b) an item has a factor loading greater than 0.4, (c) a factor has at least three items, and (d) an identified factor and retained items are interpretable in a theoretical context. We then identified two factors for professional identity (professionalism and public image) and two factors for job satisfaction (SWC: salary, welfare and career prospects; and HWW: human relationship, working content and working environment, shown in Table 2). With these two-factor solutions of professional identity and job satisfaction, we then conducted PAF with promax rotation to a pre-selected two-factor solution. The two factors of professional identity, comprising 10 items, explained 54.6 percent of the total variance in the structure matrix; the cumulative explanations were professionalism (37.4%) and public image (54.6%). No items were deleted because the indicator loading of all 10 items was higher than 0.4. The two factors of job satisfaction, comprising six items, explained 64.3 percent of the total variance in the structure matrix, and the cumulative explanations were SWC, 49.4 percent and HWW, 64.3 percent. No items were deleted because the indicator loading of all the six items was higher than 0.4.

Table 2.

Pattern matrix.^a

1. Professional identity	Two-factor
	Professionalism	Public image
I strive to protect social work professionalism	0.808
I am part of the social work sector	0.775
My work respects human rights	0.703
My work maintains social justice	0.653
Social work is a profession	0.630
I am professional	0.590
General public understand social work		0.888
General public have a good impression of social work		0.834
General public recognize social work as a profession		0.783
Social work has a professional status		0.572
2. Job satisfaction	Two-factor
	SWC	HWW
Salary	0.955
Welfare	0.947
Career prospects	0.627
Human relationships		0.866
Working content		0.634
Working environment		0.575

SWC: salary, welfare and career prospects; HWW: human relationship, working content and working environment.

Extraction method: Principal Axis Factoring.

Rotation method: Promax with Kaiser Normalization.

Step 2: Measurement model

In this analysis there are two indicators (professionalism and public image) of professional identity, two indicators (SWC and HWW) of job satisfaction and five indicators of turnover intention. Table 3 shows the descriptive statistics and correlation coefficients between these indicators.

Table 3.

Descriptive analysis and correlation coefficient (N = 1345).

	a	Mean	SD	1	2	3	4	5	6	7	8
Professional identity (PI)	0.837
1. Professionalism		3.95	0.599
2. Public image		3.176	0.858	.337**
Job satisfaction (JS)	0.832
3. SWC		2.632	0.951	.149**	.223**
4. HWW		3.186	0.794	.129**	.237**	.631**
Turnover intention (TI)	0.87
5. Thoughts of leaving		2.73	1.128	−.093**	−.231**	−.307**	−.321**
6. Feeling of no future		2.62	1.021	−.207**	−.246**	−.458**	−.381**	.618**
7. Looking for information		2.38	0.975	−.156**	−.164**	−.278**	−.257**	.588**	.619**
8. Leaving this organization		2.62	0.984	−.112**	−.133**	−.310**	−.275**	.472**	.569**	.569**
9. Leaving this profession		2.29	0.964	−.225**	−.177**	−.275**	−.288**	.545**	.573**	.652**	.534**

SD: standard deviation; SWC: salary, welfare and career prospects; HWW: human relationship, working content and working environment.

p < .05; **p < .01.

Fit indexes of the measurement model showed adequate model fit (χ²/df = 2.161; RMSEA = 0.029 (0.016–0.043); CFI = 0.996; TLI = 0.991; and SRMR = 0.016). Table 3 shows the estimates, standardized errors and standardized estimates of each factor in the three variables. The standardized estimate of factor loading from 0.50 to 0.88 is good because they demonstrate that each factor explains more than 50 percent variance of the variables.

Step 3: Structural model

Direct effect of professional identity on turnover intention

An initial structural model was tested on the direct effect of professional identity on social workers’ turnover intentions, shown in Table 4. The model fit was assessed as good. Factor loadings ranged from 0.53 to 0.87. The results showed that professional identity was significantly, negatively associated with turnover intention (β = −0.43, p < 0.001), which supported Hypothesis 1.

Table 4.

Measurement model of PI, JS and TI.

	Estimate	Standardized error	Standardized estimate
Professional identity (PI)
Professionalism	0.51***	0.02	0.50
Public image	1.0***	0.02	0.69
Job satisfaction (JS)
SWC	1.21***	0.03	0.80
HWW	1.0***	0.02	0.79
Turnover intention (TI)
Thoughts of leaving	1.0***	0.03	0.71
Feeling of no future	1.11***	0.03	0.88
Looking for information	0.84***	0.03	0.70
Leaving this organization	0.79***	0.03	0.65
Leaving this profession	0.78***	0.03	0.66

SWC: salary, welfare and career prospects; HWW: human relationship, working content and working environment.

p < .05; **p < .01; ***p < .001.

Mediation effect, moderation effect and demographic effect

The second structural model was tested by adding job satisfaction as a mediator, and the result is shown in Table 4. The model fit was assessed as good. Factor loadings ranged from 0.50 to 0.88. Additionally, the results showed that job satisfaction was a significant predictor of social workers’ turnover intentions (β = −0.45, p < 0.001). Social workers’ professional identity remained significantly associated with turnover intention after entering the latent construct of job satisfaction (β = −0.24, p < 0.001). Furthermore, professional identity was positively related to job satisfaction (β = 0.40, p < 0.001). Hence social workers’ satisfaction with a job had a partial mediating effect on the association between their professional identity and turnover intentions, which supported Hypothesis 2.

At the third stage, a structural model was tested by adding the type of organization as a moderator, shown in Table 5. A multiple groups analysis (Arbuckle, 2006) using Amos 20 was used to examine potential moderating effects. The results showed that the type of agency played a moderating role in the relationship between social workers’ professional identity and turnover intentions, indicated by the Z-score (2.582, p < 0.001); this result supported Hypothesis 3. The positive effect between social workers’ professional identity and their turnover intentions is more than three times as strong for social workers working in SWOs (β = −0.627, p < 0.001) compared with those working in OOs (β = −0.189, p < 0.01).

Table 5.

Goodness of fit indexes of measurement model and structural models.

Description of models	Measurement model	Initial structural model	Second structural model	Third structural model	Fourth structural model
	Three latent variables (PI, JS, TI)	The relationship between PI and TI	Adding JS as mediator	Adding the types of agency as moderator	Adding age and gender
χ²	34.577	14.047	34.577	106.973	308.175
df	16.0	6.0	16.0	48.0	96.0
RMSEA	0.029	0.032	0.026	0.021	0.029
CFI	0.996	0.998	0.996	0.993	0.977
TLI	0.991	0.992	0.991	0.985	0.952
SRMR	0.016	0.015	0.016	0.016	0.016

PI: professional identity; JS: job satisfaction; TI: turnover intention; RMSEA: root mean square error of approximation; CFI: comparative fit index; TLI: Tucker–Lewis Index; SRMR: standardized root mean square residual.

Finally, a fourth structural model was tested by adding age and gender as controlling variables, as Table 5 shows. The results revealed that age had a significant relationship with job satisfaction (β = −0.09, p < 0.001) and turnover intention (β = −0.09, p < 0.01), indicating that the older the social workers were, the higher job satisfaction and the lower turnover intention they had. Gender was also related to social workers’ turnover intentions (β = −0.07, p < 0.05). Compared with male social workers, female social workers had lower intentions of leaving.

Figure 1 shows the final version of the structural model depicting the ways in which social workers’ professional identity affects their turnover intentions, with job satisfaction as mediator, the type of organization as moderator, and age and gender as controlling variables. The model fit is good (CFI = 0.977, RMSEA = 0.029), which means the hypothesized model adequately fits the data – the real condition reported by social workers in Beijing.

Figure 1.

Social workers’ turnover intentions structural model (N = 1345).

Discussion

The data analysis showed that both professional identity and job satisfaction had a direct negative effect on social workers’ turnover intentions, and job satisfaction mediated the relationship between professional identity and turnover intention. For those who worked in SWOs, professional identity had a much stronger influence on turnover intention than for those who worked in OOs. Older social workers had more job satisfaction and lower turnover intentions, and female social workers had lower turnover intentions than their male counterparts. In summary, the results supported our hypotheses.

Our findings provide a new perspective to understand the crisis of high turnover rates among social workers in China. Professional identity, ignored for a long time by academics but proposed as a significant factor for turnover intention by this article, was proven to have an important impact on turnover intention and job satisfaction. It requires more attention from relevant stakeholders, including government, social work training institutes and service agencies, to nurture the intrinsic motivation of social workers.

Conforming to the theory of social identity, the results show that intrinsic professional identity can influence social workers’ perceptions of a job environment and can thereby exert a positive effect on job satisfaction. Our findings also support the assumption that the influence of professional identity occurs before that of job satisfaction. This indicates that professional identity is a more invisible but deeper factor affecting the retention of social workers. If a social worker is proud of being a social worker and identifies himself (or herself) as a social work professional, he (or she) would feel more satisfied with his or her job, even with poor remuneration and organizational environment, and would not be inclined to leave.

However, job satisfaction does matter, as other research has shown (Freund, 2005; Kim and Stoner, 2008). Its direct and mediating effect on turnover intention indicates that if some job characteristics could be improved, especially SWC, social workers would be less inclined to leave. As social identity theory shows, higher job satisfaction can lead to higher organization identity and reduced leaving intentions. This outcome requires capacity building of organizations and more government investment in social services.

Type of organization played an important role in explaining the influence professional identity exerted on turnover intention. This finding illustrates the extensive impact of political context with which organizations in China have to live. Different types of organizations receive different treatment from the government, resulting in different staff structures and challenges. As previously mentioned, there are more young graduates with social work degrees in SWOs. Their professional identity has influence on their choice of leaving, indicating the effectiveness of professional education in conveying social work values and giving workers the strength to resist the disadvantaged environment and negative public opinion.

Limitations and implications

We believe the results in this study are sound, but they are also incomplete. First, the generalizability of the findings to all social workers in other Chinese cities is limited. Although selected randomly, the sample was limited to social workers in Beijing. The findings should therefore be interpreted with caution.

Second, the study was methodologically limited by cross-sectional data in assessing structural models, including mediating models (Cole and Maxwell, 2003). Future studies should employ a longitudinal design to form causal relationships between professional identity and job satisfaction.

Third, other factors beyond the scope of this study, such as social workers’ burnout, their job commitment and organizational climate, may affect their turnover intentions (Freund, 2005; Kim and Stoner, 2008). However, this study investigated two main predictors of turnover intention (professional identity and job satisfaction) and R² for these predictors was comparatively high (34%).

Despite limitations, this article provides a convincing model to explain social workers’ turnover intentions. Relevant stakeholders could use this to understand the mechanism more clearly and to take action to prevent high turnover. Additionally, this article demonstrates that professional identity is an important antecedent to turnover intention and job satisfaction. Not only is this a new application of social identity theory in the field of social work, it also sheds light on the importance of the perception and image of the social work profession, which implies the need for improving education and recognition of social work in China.

Footnotes

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

Notes

Author biographies

Haiyan Jiang is postdoctoral researcher in the School of Social Development and Public Policy at Beijing Normal University.

Yean Wang is assistant professor in the School of Social Development and Public Policy at Beijing Normal University.

Ernest Chui is associate professor in the Department of Social Work and Social Administration at The University of Hong Kong.

Yuebin Xu is professor and dean in the School of Social Development and Public Policy at Beijing Normal University.

References

Arbuckle

(2006) Amos 7.0 User’s Guide. Chicago, IL: SPSS.

Auerbach

Schudrich

W.Z.

Lawrence

C.K.

Claiborne

McGowan

B.G.

(2014) ‘Predicting Turnover: Validating the Intent to Leave Child Welfare Scale’, Research on Social Work Practice 24(3): 349–55.

Beijing Statistical Information Net (2015) ‘Beijing Municipal Human Resources and Social Security Bureau and Beijing Municipal Statics Bureau: The Notice to Declare the Average Salary of Beijing Employee in 2014’. Available online at: http://old.bjstats.gov.cn/tjzn/mcjs/201506/t20150605_293863.htm

Berry

(2004) ‘Report from the General Social Care Council: Overseas Workers’, in Overseas Social Workers International Symposium, Sheffield, 10 November 2004.

Brown

(2000) ‘Social Identity Theory: Past Achievements, Current Problems and Future Challenges’, European Journal of Social Psychology 30: 634–67.

Burke

P.J.

Reitzes

D.C.

(1991) ‘An Identity Theory Approach to Commitment’, Social Psychology Quarterly 54(3): 239–251. Available online at: http://www.jstor.org/stable/2786653 (accessed 24 October 2016).

Cole

D.A.

Maxwell

S.E.

(2003) ‘Testing Mediational Models with Longitudinal Data: Questions and Tips in the Use of Structural Equation Modeling’, Journal of Abnormal Psychology 112(4): 558.

De Moura

G.R.

Abrams

Retter

Gunnarsdottir

Ando

(2009) ‘Identification as an Organizational Anchor: How Identification and Job Satisfaction Combine to Predict Turnover Intention’, European Journal of Social Psychology 39: 540–57.

Deaux

(2000) ‘Models, Meanings and Motivations’, in Capozza

Brown

(eds) Social Identity Processes, pp. 1–14. London: SAGE.

10.

X.T.

(2015) ‘Some Social Workers in Shenzhen Left to Dong Guan and Hui Zhou: The Turnover Rate Has Exceeded 22% Last Year’, Nanfang Daily, 30 January. Available online at: http://sz.southcn.com/content/2015-01/30/content_117407995.htm (accessed 24 October 2016).

11.

Eisenberger

Huntington

Hutchinson

Sowa

(1986) ‘Perceived Organizational Support’, Journal of Applied Psychology 71: 500–7.

12.

Ellemers

Kortekaas

Ouwerkerk

J.W.

(1999) ‘Self-Categorization, Commitment to the Group, and Group Self-Esteem as Related but Distinct Aspects of Social Identity’, European Journal of Social Psychology 29: 371–89.

13.

Etzioni

(ed.) (1969) The Semi-Professions and Their Organization: Teachers, Nurses, Social Workers. New York: Free Press.

14.

Freund

(2005) ‘Commitment and Job Satisfaction as Predictors of Turnover Intentions among Welfare Workers’, Administration in Social Work 29(2): 5–21.

15.

Fuller

J.B.

Barnett

Hester

Relyea

(2003) ‘A Social Identity Perspective on the Relationship between Perceived Organizational Support and Organizational Commitment’, The Journal of Social Psychology 143(6): 780–91.

16.

Gao

J.G.

Yan

M.C.

(2015) ‘Social Work in the Making: The State and Social Work Development in China’, International Journal of Social Welfare 24(1): 93–101.

17.

Goulder

A.W

. (1960) ‘The Norm of Reciprocity’, Americal Sociological Review 25: 165–7.

18.

Haslam

S.A.

(2004) Psychology in Organization: The Social Identity Approach, 3rd edn. London: SAGE.

19.

Kim

Stoner

(2008) ‘Burnout and Turnover Intention among Social Workers: Effects of Role Stress, Job Autonomy and Social Support’, Administration in Social Work 32(3): 5–25.

20.

Kline

(2011) Principles and Practice of Structural Equation Modeling. New York: Guilford Press.

21.

LeCroy

C.W.

Stinson

E.L.

(2004) ‘The Public’s Perception of Social Work: Is It What We Think It Is?’ Social Work 49(2): 164–74.

22.

Leung

C.B.

Wong

Y.C.

(2005) Report on Survey on Social Worker Image. Hong Kong, China: Hong Kong Institute of Social Service Development and Department of Social Work and Social Administration, The University of Hong Kong.

23.

Leung

T.T.F.

(2011) ‘The Work Sites as Ground of Contest: Professionalisation of Social Work in China’, British Journal of Social Work 42(2): 335–52.

24.

McFadden

Campbell

Taylor

(2015) ‘Resilience and Burnout in Child Protection Social Work: Individual and Organisational Themes from a Systematic Literature Review’, British Journal of Social Work 45(5): 1546–63.

25.

Miller

Garran

A.M.

(2008) Racism in the United States: Implications for the Helping Professions. Belmont, CA: Cengage Publishing.

26.

Ministry of Civil Affairs (2014) ‘Encouraging for China Association of Social Work Education’. Available online at: http://news.swchina.org/topvoice/2014/1105/18880.shtml (accessed 24 October 2016).

27.

Mor Barak

M.E.

Nissly

J.A.

Levin

(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): 625–61.

28.

(2004) ‘The Freedom of Association and the System of Non-Corporate Associations’, Global Law Review 3: 295–303.

29.

Shen

T.T.

Jiang

L.L.

(2014) ‘The Turnover Rate of Social Workers in Shenzhen Has Reached Up to the Warning Line’, Yang Cheng Daily, 19 March. Available online at: http://news.sina.com.cn/o/2014-03-19/083429742415.shtml (accessed 24 October 2016).

30.

Shim

(2014) ‘Do Organizational Culture and Climate Really Matter for Employee Turnover in Child Welfare Agencies? British Journal of Social Work 44: 542–58.

31.

Singer

(Host) (2013) ‘Social Workers and the Media: Interview with Maiken Scott (Episode 77)’, Social Work Podcast. Available online at: http://www.socialworkpodcast.com/2013/02/social-workers-and-media-interview-with.html (accessed 24 October 2016).

32.

Tajfel

Turner

J.C.

(1979) ‘An Integrative Theory of Social Conflict’, in Austin

W.G.

Worchel

(eds) The Social Psychology of Intergroup Relations, pp. 33–47. Monterey, CA: Brooks Cole.

33.

Tham

(2007) ‘Why are They Leaving? Factors Affecting Intention to Leave among Social Workers in Child Welfare’, British Journal of Social Work 37(7): 1225–46.

34.

The 2013 Social Work Day (Hong Kong) Organizing Committee (2013) Report on Survey on Social Workers Job Satisfaction, Future Development and Professional Identification. Hong Kong, China: The 2013 Social Work Day (Hong Kong) Organizing Committee.

35.

Tian

(2004) ‘Organizational Operation Under Non-Coordinate Constraint: A Framework of Research on the Relationship between the Charity Organizations and the Government in China’, Sociological Research 4: 64–75.

36.

Tian

L.L.

Jing

(2014) ‘Monthly Salary of Eighty Percent of Social Workers is below RMB 4600’, Southern Metropolis Daily, 1 July. Available online at: http://paper.oeeee.com/nis/201407/01/237298.html (accessed 24 October 2016).

37.

Tyler

T.R.

(1999) ‘Why People Cooperate with Organizations: An Identity-based Perspective’, in Sutton

R.I.

Staw

B.M.

(eds) Research in Organization Behavior, pp. 201–47. Greenwich, CT: JAI Press.

38.

Van Dick

Christ

Stellmacher

Wagner

Ahlswede

Grubba

(2004) ‘Should I Stay or Should I Go? Explaining Turnover Intentions with Organizational Identification and Job Satisfaction’, British Journal of Management 15: 1–10.

39.

(2013) ‘Non-Governmental Organizations in China: Developments and Challenges’, in Wang

Zheng

(eds) China: Development and Governance. Singapore: World Scientific.

40.

Ngai

(2011) ‘Moral Resources, Political Capital and the Development of Social Work in China: A Case Study of City J in Shandong Province’, British Journal of Social Work 43(8): 1589–610.

41.

Y.B.

Jiang

H.Y.

Leung

C.B.

(2017) ‘Who are Social Workers across Different Sectors in China: The Beijing Profile’, China Journal of Social Work (unpublished manuscript).

42.

Young

D.R.

(2000) ‘Alternative Models of Government–Nonprofit Sector Relations: Theoretical and International Perspectives’, Nonprofit and Voluntary Sector Quarterly 29(1): 149–72.

43.

Zhang

(2015) Social Workers’ Leaving Rate Has Ascended to 25%’, Beijing Evening News, 28 January. Available online at: http://news.xinhuanet.com/local/2015-01/28/c_127431704.htm (accessed 24 October 2016).