Methods: To examine the applications of SEM in social work research, 51 journal articles were reviewed by research topic and type of SEM analysis. Also, the following SEM principles were evaluated in each article: model specification, data screening, identification, estimation, model evaluation, and model modification.

Results: Across 51 journal articles, 24% dealt with children and youth; 20% dealt with substance/alcohol abuse; 16% dealt with mental health; 12% dealt with family issues; 9% dealt with social economic development; 8% dealt with health; 6% dealt with aging; and 5% dealt with minority issues. Across the same journal articles, 22% used confirmatory factor analyses and 78% used full models that combined measurement and structural models. Further, 18% used longitudinal data and 10% used multiple group comparison. Although most articles presented the above principles, several articles missed explanations about data screening and rationale for model modification, which left the reader questioning the validity of the results.

Implications: SEM analyses have provided comprehensive approaches to estimate the impacts of psychosocial factors in social work research. In terms of quantitative methodology, guidance for adequate applications of SEM analyses is necessary to help researchers confirm and develop theoretical models in social work.