BACKGROUND AND OBJECTIVES: Mechanisms of engagement and disengagement of attention to emotional information are thought to contribute to the onset and maintenance of anxiety and depression, a conclusion based largely on findings in analogue subclinical samples. However, we argue that traditionally defined analogue samples can be misleading. Firstly, research has challenged the adequacy of conventional measures of subclinical traits by illustrating that supposedly distinct scales are highly inter-correlated and do not therefore measure independent constructs. Secondly, recent research in clinical groups has revealed results opposite to those expected from the analogue literature, suggesting speeded, rather than impaired, disengagement from threat. METHODS: We present analogue findings, from a sample of 70 healthy participants, allowing a purer distinction between the phenomenology of anxiety versus depression using the orthogonal traits of positive and negative affect to classify individuals. RESULTS: Using emotional peripheral cueing we found that, at short cue durations, dysphoric individuals' (those with low positive and high negative affect) attention to facial expressions was slowed by emotional compared to neutral invalid cues. LIMITATIONS: Limitations included a small sample size and limited generalisability due to sampling from a student population. CONCLUSIONS: The data suggest that, in line with the previous subclinical literature, dysphoric individuals are slow to disengage attention from emotional information at early stages of processing and are consistent with the possibility that patterns of orienting of attention might be qualitatively different in subclinical versus clinical populations.