We compared performance of younger and older human participants to that of younger and older dogs on tasks that evaluate object discrimination, egocentric spatial ability, object recognition, spatial memory, and cognitive flexibility. Our goal was to determine whether (i) tasks sensitive to advanced age in dogs are also age-sensitive in humans; (ii) the pattern of task difficulty observed in dogs mirrors that observed in humans; (iii) dogs and humans use similar strategies to solve equivalent tasks. Dogs performed more poorly than humans on reversal tasks that evaluate cognitive flexibility. We suggest that dogs, most notably older dogs, use different strategies than healthy humans when solving these tasks. Humans appear to test a priori hypotheses to solve the task at hand. As a consequence, expectations about the complexity of the task being tested can impair human performance. By contrast, dogs appear to rely more heavily on either simpler hypotheses, or associative trial and error learning, which probably accounts for their difficulty in learning non-matching tasks. Dogs also show perseverative responding, which hinders the acquisition of reversal tasks.