The reliability and validity of methods to assess social networks, social support and control were investigated in a population of 12,009 females and males born between 1926 and 1945 (the "Malmö Shoulder and Neck Study"). This study demonstrated an overall reliability with kappa coefficients between 0.70 and 0.47, but the reliability was more varying among females and lower in the youngest age group. The analysis of the construct validity indicated that the different indices measure different aspects of the psychosocial environment, but both theoretical and methodological problems were identified, when the validity of multidimensional concepts are to be determined. The validity of such indices can best be judged by combining quantitative and qualitative methods. Potential validity problems must be kept in mind when these indices are used in epidemiological research. The results from the reliability analysis call for repeated assessments and the sample size must be adjusted vis-a-vis the reliability.