Inclusion of a mental status examination in research and clinical screening instruments lengthens the protocol, thereby adding to the difficulty of using instruments addressing all relevant issues under study without taxing participants. The purpose of the present study was to evaluate the possibility of substituting a subset of items from the widely used Mini-Mental State Examination (MMSE) for the entire examination in order to reduce the time needed to screen for cognitive status. Study data came from a health study of 783 community-dwelling, white females, 65 years of age and over, selected randomly from a 20-census-tract area of northeast Baltimore. Results indicate that seven MMSE items can be used to reliably predict total MMSE scores. Because of an age X education interaction in the prediction of total MMSE scores, four age X education-specific predictive equations were developed. These four equations are most useful for predicting continuous MMSE scores rather than for categorizing individuals according to impaired versus unimpaired status. A short form of the MMSE and four age X education-specific scoring equations are presented and their potential utility discussed.