A Study on Self-Efficacy of Social Workers in Beijing, China

Background and Purpose: Social work development in China shows its localized characteristics, the work properties within two types of agencies, government sectors and organizations, are quite different. Meanwhile, according to Bandura's self-efficacy theory, the social workers' self-efficacy will affect their work attitude and performance. Therefore, the first objective of this study is to explore factors that affect social workers' self-efficacy. The second objective is to analyse the role of Agency-type in social workers’ self-efficacy.

Methods: The return rates of the various types of organizations were 75% for SWOs, 66% for COs and 51% for PIs. Until November 2014, 1345 valid questionnaires were collected from Beijing by self-administered questionnaire, with 500 (37.2%) coming from SWOs, 620 (46.1%) from COs and 225 (16.7%) from PIs. Pearson correlation coefficient and linear regression are used to analyze the correlation and predictive variables of social workers' self efficacy.

Results: Via Pearson correlation coefficient, Professionalism(β=0.253**), Education（β=0.15**）in Value-Factor and Working-hours（β=0.101**）, Salary（β=0.135**）in Utilitarian-Factor are significantly correlated. A multiple regression model is established to predict social workers’ self-efficacy with four variables, which include Professionalism, Education, Working-hours and Salary. Generally, regression model is significant (F=31.625，p<0.05, R²=0.088). Professionalism（β=0.23，t=8.4，p<0.001）,Education（β=0.1，t=3.5，p<0.001）and Working-hours（β=0.09，t=3.3，p=0.001） are significant predictor variables, Salary is not significant(β=0.05，t=1.8，p>0.05). In multiple regression analysis of different agencies, we come to a different conclusion, Professionalism (β=0.311**), Education（β=0.113**）and Salary（β=0.154**）show a significant correlation in SWOs. And, Working-hours (β=0.028, p>0.05) is not significant, while Professionalism (β=0.29, t=6.4, p<0.001), Education (β=0.06, t=1.28, p<0.001) are significant predictors. In Non-SWOs, Professionalism and Salary, and Education and Working-hours are negatively correlated. The figure of Professionalism (β=0.221**) and Salary (β=0.12**) decreases, while the figure of Education (β=0.16**) and Working-hours (β=0.137**) increases. Only Salary is not a significant predictor (β=0.04, t=1.1, p>0.05).

Conclusion and Implications: The competence values (β=0.006<0.05) between two types of agencies are compared by the independent t-test. We found that competence value of social workers in the SWOs is higher than that of the Non-SWOs. We have found that Agency-type is an important variable of competence. Pearson correlation shows there is a significant correlation between social workers’ Working-hours and their Competence (β=0.09**). By dividing working-hours to social-work-related and non-social-work-related, we have found that there is a significantly positive correlation between social-work-related Working-hours and social workers’ Competence (β=0.163, p<0.001). The study shows engaging in non-social-work will reduce the social worker's competence. Non-social-work-related working-hours of social workers from the survey show an upward tendency, which will undoubtedly hinder the promotion of social workers’ competence. With the analysis in Part III, Agency-type would affect the predictor effect of different factors on social workers’ self-efficacy. Therefore, if we want to improve the social workers’ self-efficacy, we should consider the role of Agency-type on self-efficacy, and take specific measures from the Value-Factor and Utilitarian-Factor.