An observation method for assessing chronic pain in back pain and rheumatoid arthritis has been developed during the last decade in the U.S.A. This study examined the validity and reliability of the method for a non-chronic back pain population from another culture. Two trained observers recorded pain behaviors displayed by 61 Swedish subjects during a 10 min, videorecorded, standardized sequence of maneuvers. Subjects were asked to make ratings of their pain intensity, depression, helplessness and disability level. A physical examination was performed by an orthopedist to obtain objective medical information about the subjects. Results showed that interrater as well as test-retest reliability were satisfactory. Correlations between pain behavior and other measures of pain, e.g., intensity ratings, medication intake, and spinal mobility, were statistically significant but somewhat lower than expected. The results indicate that the behavioral observation method provides reliable and valid information about non-chronic back pain among Swedish females. However, some modifications in the standardized sequence of maneuvers and the definitions of pain behaviors may be necessary to improve the utility of the method in this population.