In a study of preconditions of responsiveness, we examined the variances obtained by two different ways of responding to questionnaire items. Subjects were 168 first-year medical students. They all completed a questionnaire of nine items. The original Likert-like scoring was (partially) replaced by visual analogue scoring in three quarters of the questionnaires. For three items the analogue form had a significantly greater variance than the discrete form, but no discrete item had a significantly greater variance than its analogue counterpart. Assuming that a greater variance was indicative for a greater responsiveness, our findings provide some support for the preferred use of analogue scales over discrete response categories in the measurement of changes over time.