Factor Analyses of the Family Assessment Device

Authors

  • TY A. RIDENOUR Ph.D., M.P.E.,

    1. Dr. Ridenour, Postdoctoral Fellow, and Dr. Wendy Reich, Associate Professor, are affiliated with the Department of Psychiatry, Washington University School of Medicine, 40 N. Kingshighway, Suite 4, St. Louis MO 63108. Address correspondence to Ty A. Ridenour, Ph.D. at this address or by e-mail: rident@epi.wustl.edu.
    Search for more papers by this author
  • JAMES DALEY Ph.D.,

    1. James G. Daley, Ph.D., Assistant Professor, School of Social Work, Southwest Missouri State University, Springfield MO 65804-0095.
    Search for more papers by this author
  • WENDY REICH Ph.D.

    1. Dr. Ridenour, Postdoctoral Fellow, and Dr. Wendy Reich, Associate Professor, are affiliated with the Department of Psychiatry, Washington University School of Medicine, 40 N. Kingshighway, Suite 4, St. Louis MO 63108. Address correspondence to Ty A. Ridenour, Ph.D. at this address or by e-mail: rident@epi.wustl.edu.
    Search for more papers by this author

  • This project was funded in part by an NIMH Postdoctoral Training Grant, MH17104. The authors thank three anonymous reviewers for comments on an earlier draft of the manuscript.

Abstract

The Family Assessment Device (FAD) operationalizes the McMaster Model of Family Functioning, which has been used in numerous studies, translated into seven languages, and is regarded as one of the most researched family assessment tools available. However, exploratory and confirmatory factor analyses using the 7-by-7 matrix of subscale correlations from the original validity study on the FAD (Epstein, Baldwin, & Bishop, 1983) indicated that the FAD subscales overlap substantially and do not assess unique dimensions of family functioning. Results of our study suggest that the conservatively best use of the FAD is using the General Functioning subscale as a summary score. A model that fits the data marginally better than the General Functioning score and a Measurement Error model, however, consisted of “Collaboration” and “Commitment” latent factors. These results illustrated the need for more extensive validity research on the FAD, because interpretation of the factors and subscales had to rely heavily on face validity.

Ancillary