Time of day is known to influence cognition differently across age groups, with young adults performing better later than earlier in the day and older adults showing the opposite pattern. Thus age-related deficits can be smaller when testing occurs in the morning compared with the afternoon/evening, particularly for tasks requiring executive/controlled/inhibitory processes. Stronger influences of time of day were therefore predicted on associative than on item recognition memory based on their differential requirements for demanding recollective (rather than familiarity) processes. In 2 experiments, participants were presented with unrelated word pairs and then tested on both item recognition (old/new item?) and associative recognition (intact/recombined pair?). In Experiment 1, young adults were tested either in the morning or in the evening; recognition memory was better when time of testing matched participants' morningness-eveningness preferences, and more so for associative than for item memory. In Experiment 2, young and older adults (evening and morning types, respectively) were tested both in the morning and in the evening; again, recognition memory was better at participants' preferred times of day, especially for associative memory. Consequently, age-related associative deficits varied considerably-indeed more than fourfold-from a nonsignificant 8% for testing in the morning to a substantial 35% for testing in the evening, suggesting that it is important to consider time of day effects in future studies of the associative deficit hypothesis. (PsycINFO Database Record