Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) have contributed to test development and validation in counseling psychology, but additional applications have not been fully realized. The author presents an overview of the goals, terminology, and procedures of factor analysis; reviews best practices for extracting, retaining, and rotating factors in EFA; describes issues related to CFA, including potential hypotheses, how parameters are estimated, and current guidelines regarding sample size and the use of fit indices; and reviews more sophisticated applications of CFA, including multiple-groups analysis and testing of hierarchical models. Finally, the author recommends how to increase the application of factor analysis to counseling psychologists’ work in research, training, and practice settings.
1 March 2001 in Journal of Social and Clinical Psychology
This cross-sectional study investigated predictors of research productivity and science-related career goals in a sample of 267 doctoral students (representing a response rate of 5S%) from 15 randomly selected APA-accredited counseling psychology doctoral programs. A structural equation modeling procedure revealed that career goals and research productivity could be predicted by Holland personality type, perceptions of the research training environment, interest in research, and research self-efficacy. Students' gender and year in the doctoral program also contributed to this causal model as additional predictor variables, providing a very good fit to the data The present findings contribute to theories of research training by presenting a comprehensive examination of the major factors previously investigated in the literature as predictors of research productivity and science-related career goals within the context of a structural equation model
Measures of research self-efficacy have the potential to facilitate graduate training and mentoring, but the hypothesized factor structures of these measures have not been confirmed empirically. Moreover, the underlying dimensions of research self-efficacy across multiple measures are unknown. Graduate students in psychology programs (N = 1,004) completed three research self-efficacy instruments via a Web-based survey. Confirmatory factor analyses did not support the hypothesized factor structures. An exploratory factor analysis suggested that four dimensions of research self-efficacy (data analysis, research integration, data collection, and technical writing) underlie responses to items from the three instruments. The authors present possible reasons for the differences in factor structures across instruments and suggest how the assessment of research self-efficacy may facilitate the training and career mentoring of graduate students.
1 October 2009 in Journal of Counseling Psychology