| Abstract: | Validity of the GDS and the GDS-Short Form (GDS-SF) was investigated in relation to the Brief Psychiatric Rating Scale-Depression Factor (BPRS-D) with 138 elderly medical outpatients. Patients were at least 65 years of age, understood English, and had intact or mildly impaired cognitive ability. The GDS was administered once, and GDS-SF items were extracted from the parent scale for analysis. Significant Pearson correlation coefficients between the BPRS-D and the two forms of the GDS, supported concurrent validity of the GDS and GDS-SF: BPRSD/GDS=.83; BPRS-D/GDS-SF=.79, p<.0001. Although the correlation between the GDS and GDS-SF was .94 p<.0001, a t-test for dependent samples, indicated the BPRS-D/GDS correlation was higher than the BPRS-D/GDS-SF correlation: t(1, 138)=1.89, p<.05. ANOVA indication a significant difference in at least one of the groups identified by the GDS relative to BPRS-D classification of depression groups, F=88.6, p<.0001. Scheffe's test found a significant difference between the nondepressed group and the very mildly depressed group. Chi-square analysis also showed a significant difference between depressed and nondepressed groups in terms of expected frequency of classification by various GDS cutoff scores: GDS 11 yielded x^2 (1,138)=79.9; GDS 14 yielded x^2 (1,138)=69.7, GDS-SF 5 yielded x^2 (1,138)=54.5; GDS-SF yielded x^2 (1,138)=69.08, p>.000. Hit rates for cutoff scores were: GDS 11=89.9%; GDS 14=87.7%; GDS-SF 5=84%; GDS-SF 7=89.1%. All cutoff scores had a 68.8% base rate. The highest identification of depressed patients was by cutoff scores GDS 11 (35), and GDS-SF 5 (32). Thus, concurrent validity of the GDS and GDS-SF with the BPRS-D was supported. Utility of the GDS-SF cutoff of 5 was comparable to the GDS, although superior performance was observed by the GDS cutoff of 11. A discussion of results is presented, including limitations of the study and suggestions for future research. |
|---|