The visual analogue scale (VAS) is a simple and frequently used method for the assessment of variations in intensity of pain. In clinical practice the percentage of pain relief, assessed by VAS, is often considered as a measure of the efficacy of treatment. However, as illustrated in the present study, the validity of VAS estimates performed by patients with chronic pain may be unsatisfactory. Two types of VAS, an absolute and a comparative scale, were compared with respect to factors influencing the reliability and validity of pain estimates. As shown in this study the absolute type of VAS seems to be less sensitive to bias than the comparative one and is therefore preferable for general clinical use. Moreover, the patients appear to differ considerably in their ability to use the VAS reliably. When assessing efficacy of treatment attention should therefore be paid to several complementary indices of pain relief as well as to the individual's tendency to bias his estimates.