There is increasing evidence that blood-injection-injury (BII) phobia is qualitatively different from the other specific phobias in the sense that phobic distress takes the form of disgust rather than (threat-induced) fear. Following this, we tested the relative importance of harm and disgust-related associative biases in BII-fear. High (n=25) and low (n=27) fearful individuals saw a series of fear-relevant (blood-related) and fear-irrelevant (rabbit and flower) slides which were randomly paired with either a harm-related outcome, a disgust-related outcome, or nothing. Preexperimentally, participants expected blood-related slides to be followed by both disgust- and harm-relevant outcomes. These selective preexperimental outcome expectancies were readily corrected during the experiment. Neither low nor high fearful participants showed a postexperimental covariation bias. The absence of differential effects between high and low fearful participants does not support the idea that disgust- or harm-relevant associative biases play a role in the maintenance of BII-fears. The results corroborate the previous finding of Pury and Mineka [1997. Covariation bias for blood-injury stimuli and aversive outcomes. Behaviour Research and Therapy, 35, 35-47] that people are generally liable to selectively associate BII-stimuli with aversive outcomes.