Abstract Anxiety is often paired with sleep disturbances and both interact in a quite complex manner. Sleep (and vigilance) problems are often included in the descriptive definition or in the diagnostic criteria for anxiety disorders. Nevertheless, if anxiety may cause sleep disturbances, it is also known that sleep deprivation may produce symptoms which fall within the symptom complex of anxiety. In this paper, some of the methodological issues encountered when studying sleep and anxiety are discussed. Polygraphic recordings of sleep in anxious patients have consistently shown an increased sleep latency and, quite often, also exhibited a reduced sleep time, a reduced total sleep time, less slow-wave sleep, a greater arousal index and an increased duration of wakefulness during sleep. From our own study, we also report anomalies of the first night cycle in anxious poor sleepers who are otherwise indistinguishable from normal controls (with regard to the ‘classical’ sleep parameters). We have also observed the large interindividual variability of numbers of sleep parameters in anxious people. The question of a potential heterogeneity of the studied groups with regard to their clinical presentation as well as their sleep profile has been raised through our research as well. It is apparent that strategies for exploring the source of the potential heterogeneity of anxiety disorders are still needed.