Based on these findings, the BCAT appears to be more sensitive than the BIMS in predicting cognitive level for nursing home residents.

We discuss the merits and shortcoming of this study, the utility of MemPics™ for providing meaningful engagement in long-term care residents with mild to moderate dementia, and ideas for future research.

IADL dependence can occur even in MCI. Testing suggestive of MCI should alert clinicians to further investigate the older adult's profile of cognitive and functional limitations to highlight targets for caregiver support and promote independence by "right-sizing" community or facility resources.

The Brief Cognitive Assessment Tool (BCAT) is a new screening measure for cognitive dysfunction that emphasizes contextual memory and executive control functions. A total of 104 older adults referred for neuropsychological evaluation were recruited from assisted-living facilities. Psychometric analyses confirmed strong evidence for reliability, construct validity, and predictive validity. The BCAT's utility for identifying dementia versus mild cognitive impairment was excellent, with a sensitivity of .99, a specificity of .77, and an area under the receiver-operating characteristic (ROC) curve of .95. Executive control, contextual memory, and attentional capacity items were the best predictors of diagnostic category and of instrumental activities of daily living.

Forty-three elderly women at two low-income city nutrition sites participated in a study of social ties and social support. As predicted, discrete facets of support were differentiated from among a variety of social network and support vuriables. Age was negatively related to network size, amount of emotional support, and time spent with confidantes and relatives. With age controlled, the best predictors of life satisfaction were network size, frequency of church attendance, and proportion of network members considered intimates. Results are discussed with regard to the multidimensionality of the social support construct.There is a growing consensus that social ties play an important role in maintaining health, morale, and well-being (Broadhead, Kaplan, James, Wagner, Schoenbach, Grimson, Heyden, Tibblin, & Gehlbach, 1983). Although the claims made for social support are impressive, a problem limiting advances in our understanding of this area is that the dominant research style employed in bast studies has emphasized predictive validity. Research was directed more toward demonstrating the connection of social support to health outcomes than toward understanding the complexities of the support concept. The disadvantage of this approach is that predictive validity is maximized at the expense of con-