The hypothesis that age deficits in long-term episodic memory tasks are due to imparied metamemorial skills was investigated in a paired-associate task that required participants to predict the likelihood of recalling individual items. Both young and old adults were able to predict their ability to recall, and both age groups were equally sensitive to the effects of differences in the degree of relatedness between word-pair members on subsequent recall. However, both old and young adults were insensitive to the beneficial effects of using interactive imagery at encoding. It was concluded that differences in metamemorial skills are not responsible for age differences in memory.