The aim was to construct an updated and balanced version of a previous Swedish social attitude scale (Sidanius, 1976), and to assess the reliability and construct validity, employing a sample of 445 students, aged 14–17, from schools within metropolitan Stockholm. The results showed acceptable precision and internal consistency and reasonably high test‐retest reliability for the scale as a whole. The a priori attitude domains (political‐economic conservatism, religion, punitiveness, social inequality, ethnocentrism, and environment) were identified through use of confirmatory factor analysis. Most domains showed acceptable reliability. Differing results for the various age groups indicated that the present version of the scale should not be used among subjects below the age of 15. Copyright © 1990, Wiley Blackwell. All rights reserved