Pur-pose: The purpose of the study was to check the generality of the structure of social support sources as measured by the Multidimensional Scale of Perceived Social Support (MSPSS) and to examine the instrument's psychometric properties in Turkey. A previous study in Turkey showed that the definitions of support sources may differ across cultures and an attempt was made in the present study to incorporate definitions to make the scale more meaningful in Turkey. Method: Three samples (psychiatry, surgery, and normal), each consisting of 50 subjects, were selected from a hospital in Turkey. No significant demographic differences were present between the samples. Factor analysis. Cronbach's I, correlations with other validated scales, and group comparisons were used for the analyses. Results: The original three support sources (family, friends, significant other) were found. However, in this version of the MSPSS the Family and the Significant other subscales incorporated definitions more meaningful in our samples. Internal consistency was high and correlations with the other scales showed adequate validity, particularly for the psychiatry and the surgery samples. The psychiatry sample reported the lowest levels of support. Conclusion: The generality of support sources, with some differing definitions, was supported. The psychometric properties of the MSPSS were adequate for it to be used in Turkey. However, social support may be especially relevant for groups with psychological/medical problems, and support from close people (particularly the family) may be especially predictive of health.