ABSTRACT
The development of the GDS began with a team of geriatric mental health specialists selecting 100 items believed to be useful for distinguishing between depressed and nondepressed older adults. An effort was made to include items that covered the range of depressive symptoms, including memory loss, cognitive complaints, somatic complaints, and self-image. A yes/no response format was chosen based on the experiences of the authors with multiple-response formats. They observed that multiple-response formats often confuse older patients. The 100 items were administered to 47 male and female participants who were over the age of 55 and were either community dwelling with no complaints of depression or hospitalized for depression. The authors used a bootstrapping strategy to select items; that is, those items evidencing the best correlation to the total score were retained for further validation. The rationale for this strategy was that the 100 items generated would provide the best measure of the geriatric depression construct. A decision to select 30 items was made, presumably, to minimize fatigue effects. The item-to-total correlations ranged from .47 to .83. Interestingly, the 12 items related to somatic concerns were not selected based on this procedure; that is, the item-to-total correlations for these items were not in the top 30. Table 12.1 contains the 30 items that compose the GDS.
