Scoring nonresponse on the Mini-Mental State Examination.

Alternative procedures for coding nonresponse on the Mini-Mental State Examination as error or as correct) results in different classification of 13% of a random sample of 1931 subjects aged 60 and over. Comparison of responders' and nonresponders' ability to perform activities of daily living, and examination of the relative difficulty level of omitted items indicates that, in epidemiological surveys, scoring nonresponse as error is more likely to be correct.