This paper presents an empirical investigation of age and cognitive ability as predictors of computerized information retrieval. Upon the basis of age-related changes in cognitive ability, hypotheses were generated relating to the effects of database structure (linear, hierarchical, or network) and node selection method (explicit or embedded menu). In keeping with previous research in other areas of human-computer interaction, there was a significant main effect of age, with older subjects performing more slowly. However, interactive effects of ‘question block’ indicated that older subjects were at a particular disadvantage in the early stages of task performance. Age differences in processing speed and/or psychomotor skill appeared to be a particularly important factor. Whilst the effects of cognitive ability were generally weak, spatial memory and logical reasoning scores were negatively correlated with information retrieval response times. Although interactions were not significant, trends in the data for all dependent measures suggested that older subjects may be at a disdvantage when using a network structure.