Cross-cultural validity and measurement invariance of the social physique anxiety scale in five European nations

Abstract

The cross-cultural generalizability of the social physique anxiety scale (SPAS) was evaluated using confirmatory factor analysis (CFA) in five European nations: Britain, Estonia, Spain, Sweden, and Turkey. Motl and Conroy's (2000) methods were used to develop modified versions of the scale within each sample based on the original 12-item version. Pending the satisfactory fit of the CFAs of the modified models within each sample, it was expected that the measurement parameters and mean values of these models would be equivalent across samples in multisample CFAs. An eight-item version of the SPAS exhibited a good fit with data from the British, Estonian, and Swedish samples, and a seven-item version fitted the data well in the Spanish and Turkish samples. The eliminated items were also influenced by a method effect associated with the item wording. Multisample analyses revealed that factor loadings were equivalent across samples. Tests of latent means revealed that British and Spanish participants reported the highest levels of SPA, with Estonian participants reporting the lowest. Results indicate that the SPAS is generalizable across these cultures, although subtle variations existed in the Spanish and Turkish samples. Researchers are advised to follow these procedures to develop a valid version of the SPAS appropriate for their sample. © 2007 Blackwell Munksgaard.