A Validation Study of the Spanish Self-Directed Search Using Back-Translation Procedures

Abstract

Assessment of cultural differences in the structures of vocational interests assumes metric and translation equivalence of interest measures used for culturally diverse populations. The present study examines the adequacy of the Spanish translation of Holland's (1990) Self-Directed Search-Easy (SDS-E) using back-translation procedures and structural analyses. The most recent revision of the English-language SDS-E and, separated by a 1-week interval, the back-translated (Spanish to English) version of the SDS-E developed for this study were completed by 145 undergraduates. Analyses using the randomization test of hypothesized orders (Hubert & Arabie, 1987) indicated that Holland's model adequately fit both the SDS-E and back-translated scale (Bktran) scores of this sample and that the model fit the Bktran data significantly better than it fit the SDS-E data. For both instruments, Holland's model fit the scores of the men (n = 73) significantly better than it fit the women's (n = 70) scores. This observed gender difference accounted for the superior fit of the full Bktran data matrix to Holland's model, because a significant difference in model-data fit was observed only for the female subsample. These results suggest that the Spanish version is an adequate translation of the SDS Form E, but that Holland's model fits less well for women.