A wide variety of measures have been used to assess the effects of social support on physical and psychological well-being. Many instruments have incomplete information about reliability and validity. Other instruments appear to be reasonably valid, but involve a large number of questions and are therefore not suitable in population surveys. The main purpose of the study was to examine whether a psychometrically well documented, but time consuming instrument. The Interview Schedule for Social Interaction, could be abbreviated and simplified for the use in population surveys. By means of both item and content analyses the original instrument was shortened and both versions were tested in a study group of 83 middle-aged Swedish men. The distribution of scale scores were similar in the complete and in the abbreviated version and so was reliability, measured as internal consistency and split-half reliability. Validity was examined by comparing the two versions in relation to other measures of social integration, as well as personality and behaviour characteristics. Men with low social support were less socially and physically active, were more depressed and were less trustful than men with high support. Men who scored low on social support were also more often smokers, of lower social class, lower occupational level and lower education and had more complaints of ill health, mainly of cardiovascular nature. The discriminative capacity was as good for the complete as for the abbreviated version. Thus no obvious disadvantages in terms of reliability and validity could be demonstrated for the abbreviated version.