Forty-three consecutive patients were interviewed on the eve of elective gynaecological surgery to determine the effect of anxiety on the granting of informed consent to participate in an hypothetical study. Anxiety was assessed using the Spielberger state-trait anxiety inventory and 10-cm linear analogue scale. A standardised explanation of an hypothetical premedication study was given and the patients' consent requested. Results were grouped for those who granted (n = 33) and those who withheld (n = 10) consent: anxiety scores for the latter were significantly higher (p less than 0.01). It is concluded that patients with high pre-operative anxiety levels are more likely to withhold consent for inclusion in premedication studies than are those who are less anxious. Seeking informed consent would introduce bias into studies of anxiolytic premedication.