In Study 1 subjects high and low in trait anxiety were asked to identify the colour of anxiety-related, anxiety-matched neutral, happiness-related and happiness-matched neutral words. There were two types of stimulus presentation: blocked trials, where stimuli were blocked with respect to their valence, and mixed trials, where stimuli were mixed with respect to their valence. Subjects high in trait anxiety took longer to identify the colour of anxiety-related compared to matched neutral words for blocked-trial presentation only. In Study 2 subjects were exposed to either a positive or a negative mood-manipulation followed by a mixed-trial Stroop task. The mood manipulation procedure was successful in producing predicted changes in self-reported state anxiety. Analyses showed that high-trait-anxiety subjects exhibited interference effects consistent with the induced mood. No such effects were observed for the low-trait-anxiety subjects.