Meta-analysis was used to examine the performance of depressed and Alzheimer-type dementia (DAT) patients on standard and experimental clinical tests of cognitive function. Deficits were found for depression on almost every psychological test. Relative to nondepressed controls, the average deficit was 0.63 of a standard deviation, but the magnitude of the effect varied with the type of test. DAT patients performed worse than depressed patients, with an average effect size of 1.21 standard deviations, but the size of the effect depended on the clinical test. Effect sizes for the comparison between depressives and controls were significantly affected by age, treatment setting, ECT use, severity of depression, and the source of diagnostic criteria, but not by the type of depression. Effect sizes in the comparison of depressives to DAT patients were influenced by age, the severity of depression, and ECT. Depressives performed proportionately worse than controls on tasks with pleasant or neutral, compared with unpleasant content, on speeded compared with nonspeeded tasks, and on vigilance tasks. However, there were no differences in the magnitude of effect size for tests using recall compared with recognition, using categorical compared with noncategorical word lists, on story compared with word comprehension, and using verbal compared with visual material. Relative to DAT patients, depressives performed no better on recall compared to recognition tasks, or verbal compared to visual material. The findings of the review are not consistent with the hypothesis that depression is associated with deficits in effortful processing. A model of psychological deficit in depression as a deficit in speed or attention has more promise. (JINS, 1997, 3, 631–651.)