Standardization data for the Wechsler Memory Scale-Revised (WMS-R) were used to generate confidence intervals for the statistical abnormality of differences between scales. Tables are provided for evaluation of material-specific rates of forgetting on the Logical Memory and Visual Reproduction subtests, adequacy of overall retention as reflected by contrasts between the General and Delayed Memory Indexes, and test-retest changes on the WMS-R scales. Confidence intervals derived from reliability and abnormality data are compared. The clinical significance of a discrepancy between scales may be better addressed by reference to information about the latter probability.The Wechsler Memory Scale (Wechsler, 1945) provided a single summary score or Memory Quotient that consolidated measures of attention, verbal, and nonverbal recall, but did not differentiate various aspects of memory. The Wechsler Memory Scale-Revised (WMS-R) was improved by the inclusion of separate indexes summarizing Visual Memory, Verbal Memory, General Memory, Attention/Concentration, and 30-min Delayed Recall tasks (Wechsler, 1987).Comparisons between the patient's scores on these scales may provide useful diagnostic information. For example, Butters et al. (1988) and Wechsler (1987) reported that scores on the Attention/Concentration Index were significantly higher than were General and Delayed Memory Index scores in Alzheimer's and amnesic patients. A similar pattern characterized the residual memory deficits of patients who had sustained moderate to severe closed head injuries (Crossen & Wiens, 1988;Wechsler, 1987). The Delayed Memory Index was significantly lower than was the General Memory Index in amnesia and closed head injury patients (Butters et al., 1988;Crossen & Wiens, 1988;Wechsler, 1987), indicating an abnormal rate of forgetting in these conditions.Contrasts between the immediate and delayed recall trials on the Logical Memory and Visual Reproduction subtests may also provide useful clinical information. Consistent with studies using the earlier versions of these subtests (Russell, 1988), when delayed and immediate recall conditions are compared, it has been found that patients with disease processes such as Alzheimer's dementia, alcoholic Korsakoff's syndrome, and Huntington's disease retain less information than normal after 30 min (Butters et al., 1988). Differential rates of forgetting may

Patterns of intrasubtest scatter in the WAIS‐R protocols of patients (n = 32) with Alzheimer's disease were compared to those of normal elderly controls (n = 32). The Alzheimer's patients showed more randomly dispersed item failures on some subtests, but normal controls showed more intrasubtest variability on other measures. Rates of correct diagnostic classification based on scatter measures were only slightly better than chance despite the presence of prominent anomia, memory impairment, construction apraxia, and significant decline from premorbid intellectual level in demented patients. In contrast, demographically based estimates of intellectual loss produced accurate diagnostic classification in 81 % of the cases. The incremental validity of qualitative scatter analysis in the evaluation of suspected Alzheimer's disease appears to be minimal.

No abstract