Longitudinal characteristics of the Center for Epidemiologic Studies--Depression Scale.

The internal structure of the Center for Epidemiologic Studies--Depression Scale was examined within the context of a three-year longitudinal study of 80 low-income parents of young children. The baseline, first-year, and second-year follow-up home interviews included the scale. Principal components factor analyses with varimax rotation yielded six factors for the baseline data, and five factors for the follow-up data with different items loading on different factors each year suggesting some instability of the factors. Cronbach alpha estimates of the internal consistency of the original factors showed satisfactory values for Depressed Affect, Positive Affect, and possibly Somatic, but not for Interpersonal. However, alpha for the total score was very strong. Positive Affect and the total score also showed stability over the three administrations. These results question the use of the original factor scores for the CES-Depression as subscales but support the use of the total score as a measure of depression.