The purpose of this investigation was to address the need to clarify the factorial measurement properties of the Social Physique Anxiety Scale (SPAS). Data collected from 760 female participants (who, on average, were young adults) were randomly placed into one of two samples to facilitate double cross-validation analyses. Calibration confirmatory factor analyses of three plausible models identified in research reports were conducted using structural equation modeling procedures. Subsequent cross-validation revealed a model with two first-order factors subordinate to one second-order factor to be unambiguously the most adequate among competing models. This model also exhibited a good fit both in calibration and in cross-validation with all incremental fit indexes exceeding the desirable .90 criterion. These results challenge initial validation study contentions that the SPAS is unidimensional.