Cognitive models of insomnia suggest that selective attention may be involved in maintaining the disorder. However, direct assessment of selective attention is limited. Using the inducing change blindness (ICB) paradigm we aimed to determine whether there is attentional preference for sleep-related stimuli in psychophysiological insomnia (PI) relative to delayed sleep phase syndrome (DSPS) and good sleepers (GS). In the ICB task, a visual scene, comprising both sleep-related and neutral stimuli, 'flickers' back and forth with one element (sleep or neutral) of the scene changing between presentations. Therefore, a 2 x 3 totally between-participants design was employed. The dependent variable was the number of flickers it took for the participant to identify the change. Ninety individuals (30 per group) were classified using ICSD-R criteria, self-report diaries and wrist actigraphy. As predicted, PI detected a sleep-related change significantly quicker than DSPS and GS, and significantly quicker than a sleep-neutral change. Unexpectedly, DSPS detected a sleep-related change significantly quicker than GS. No other differences were observed between the two controls. These results support the notion that there is an attention bias to sleep stimuli in PI, suggesting that selective attention tasks such as the ICB may be a useful objective index of cognitive arousal in insomnia. The results also suggest that there may be an element of sleep preoccupation associated with DSPS. Results are discussed with reference to other experiments on attentional processing in insomnia.