The purpose of this study was to test empirically a conceptually derived multidimensional formulation of social support. Analyses were based on responses of 313 expectant couples who varied in age, income, and education. A questionnaire, which included the Support Behaviors Inventory (SBI), was given in the latter half of pregnancy. A four-phased analysis was conducted: discriminate validity testing of the multidimensional support structure, further subscale analysis, factor analysis of the structure of the SBI, and reformulation of the SBI based on results of the factor analyses. Results did not demonstrate independence of measurement with items that were developed to represent a priori social support dimensions of emotional, material, informational, and appraisal support. The factor analysis in particular revealed that the scree test by Cattell (1966) demonstrated a large, dramatic discontinuity in eigenvalues and suggested that there was only one systematic factor. These results implied a dominant construct of social support in pregnant couples that organizes at the broad level the perceived degree of experienced support during pregnancy. This broader concept was so dominant in the current study, explaining 48% of the variance in partner support and 61% of the variance in others support, that the idea of multidimensionality was not confirmed. These data suggest an ongoing need to scrutinize carefully and validate empirically the hypothesized multiple dimensions of social support proposed in a number of recent support instruments.