There were four groups of participants: low-anxious (low anxiety-low defensiveness), repressors (low anxiety-high defensiveness), high-anxious (high anxiety-low defensiveness), and defensive high-anxious (high anxiety-high defensiveness). They were exposed to self-focus and other-focus conditions, and self-report, physiological, and behavioural measures of anxiety were recorded. The focus manipulation (self vs. other) was effective, with all four groups having higher self-reported anxiety in the self-focus condition. It was found that self-focused attention accounted for a significant amount of variation in self-reported anxiety. The findings were discussed within the theoretical framework proposed by Eysenck (1997).