Cognitive effects of estrogen have been considered in a number of large, randomized, double-blind, placebo-controlled trials. Most have involved older, postmenopausal women, and results of these provide little support for the view that estrogen-containing hormone therapy initiated after age 60 substantially affects mean cognitive performance over periods of time ranging up to 5 years. This conclusion appears particularly true for episodic memory, a cognitive domain in which impairments are associated with increased risk of Alzheimer's disease. Other domains have been less thoroughly assessed. For women undergoing surgical menopause, limited clinical trial evidence suggests that prompt initiation of estrogen therapy may benefit verbal episodic memory, at least over a period of several months. Among middle-aged women, observational studies indicate no important deleterious effect of the natural menopause transition on cognitive performance. Similarly, limited clinical trial evidence from middle-aged postmenopausal women implies no substantial effect of hormone therapy on episodic memory, at least over the short term. Unfortunately, no randomized clinical trials have addressed long-term cognitive outcomes of hormone therapy started during the menopausal transition or early postmenopause, a time hypothesized to represent a 'critical window' of opportunity. There is urgent need for research in this area, and at least two clinical trials now underway may eventually provide partial answers.