Meta-Analysis of Turnover Intention Among Child Welfare Workers

Background and Purpose: High rates of turnover among public child welfare (CW) workers has a detrimental influence on social workers, clients and their families, agencies, and states. Although the severity and prevalence of the problem are well understood, it was only recently that research on turnover among CW workers became more popular. Our understanding of the phenomenon, however, is still lacking and requires quantitative research synthesis. To address this research gap, the purpose of this study is to conduct a meta-analysis of the existing literature on turnover intention among child welfare workers.

Methods: The inclusion criteria to select articles for this study included: 1) used turnover intention as an outcome; 2) had predictors as the antecedents of turnover intention; 3) used Pearson product moment correlation coefficients, which assessed the relationship between the predictor variables and turnover intention; 4) involved study samples of current front-line public child welfare workers; 5) did not include any interventions or manipulations in the study; 6) were conducted in the U.S.; 7) were either published or unpublished research articles between 1990 and 2013 (as of Feb, 2013). Extensive literature search (i.e., databases, relevant journals, and reference lists) returned 144 studies and, among them, 22 studies were used in the final meta-analysis. Hunter and Schmidt (2004)’s approach was used to combine results of these studies and such artifacts as sample size and reliability of variables were corrected.

Results: Overall, predictors related to CW worker attitudes and perceptions (e.g., organizational commitment and job satisfaction) had the highest influence on turnover intention among CW workers. On the other hand, demographic predictors, such as age, race, and gender, showed small or negligible effects on turnover intention. Among work-related predictors, stress-related predictors and sub-factors of burnout had medium to high influence on turnover intention while predictors related to decision-making showed medium effect sizes. Among predictors in work environment category, various types of support predictors had varying influence on turnover intention while such variables as perceptions of fairness and policy had relatively high effect sizes of around .4.

Conclusions and implications: Several commonly-studied factors with proven validity, such as organizational commitment, stress, job satisfaction, professional commitment, and organizational climate, emerged as some of the strongest predictors. Other variables, such as perceptions of fairness, safety concern, and policy, were also shown to be strong predictors, but have received relatively less attention. While caseload is commonly thought of as one of major drivers of turnover, this study showed that caseload had little effect on turnover intention of CW workers. This study highlights several key areas of further research. First, job performance and economy-related factors are rarely utilized in studies of turnover intention among CW workers and therefore, deserve greater attention. Second, studies with private CW workers were relatively small compared to studies with public CW workers. Third, there has been a conceptual confusion of turnover intention measures, which future research can help to further clarify. The theoretical and practical implications highlighted by this study are also discussed.