The measurement of social support in population surveys.

There is an increasing interest to include measures of social support in population surveys of chronic disease risk. The choice among a variety of measurement instruments, however, is difficult. We reviewed social support instruments, which were potentially applicable in population studies. Their conceptual framework, their reported predictive capacity and their psychometric properties were compiled and systematically compared. The convenience, clarity and applicability of the social support instruments were tested in a group of health care employees. Two distinct groups of instruments were identified, those describing quantitative aspects of social network and social interaction and those describing functions and adequacy of social support. The former instruments were more easily applicable, questions were more easily and quickly understood and answered. Furthermore, these instruments had been found to predict physical illness in prospective studies. Their psychometric properties, however, were unknown. Qualitative instruments, on the other hand, were found to include questions, not universally applicable to a general population. These instruments were mostly carefully tested for psychometric properties, but their illness predictive capacity was found to be less well examined. Thus, instruments which satisfied all the desirable requirements could not be identified.