Elderly people, the majority of whom completed education at age 14, comprise an increasingly significant proportion of the population. Neurological disorders which affect language processing are particularly associated with older age. To detect mild or subtle deficits, assessment is required to extend beyond decontextualized units which permit only literal interpretation. A lack of elderly normative data limits the usefulness of such measures. The influences of age, education and gender on comprehension were assessed in adults aged between 40 and 88. Clear advantage for those educated beyond minimum school leaving age and disadvantage for those over age 75 was evident in all measures: discourse, metaphor and inference. In particular the scores of very elderly subjects whose education was completed at the then minimum age of 14 show similarity to the published test data for pathological populations. Age and education are relevant to interpretation of language test performance of those with known pathology but also to everyday communication.