An empirical investigation of Coombs' unfolding theory was conducted in which basic assumptions of the theory were tested. The primary concern was to investigate the effects upon individuals' preference judgements when the presented set of scaled stimuli was changed. Two types of judgemental tasks were used to assess these possible contextual effects: rank order and paired comparisons, and two different sets of scaled stimuli were presented. A reliability measure was also taken as apparent context effects could merely be due to lack of reliability. The obtained data supported the basic assumptions of Coombs' theory, and subjects' responses were highly reliable, thus little confounding should have occurred. The change of context effects, however, were contrary to the prediction of Coombs' theory. Strong evidence for these context effects was found for both sets of stimuli and both kinds of tasks implemented. Implications of these findings are discussed and doubts are raised concerning the usefulness of the unfolding theory in the context of attitude scaling.