Introduction: The term “generativity” was used for the first time by Erikson (1950). It refers to the state of adulthood in the life cycle, which implies procreativity, productivity and creativity, it also impulses the development of one’s own identity. There are several ways to measure generativity such as interviews, direct observation, case studies, or self-report questionnaires. The most widely used method is self-report, and among the available scales, the Loyola Generativity Scale (LGS) is one of the salient scales. lgs is a self-report measure composed of twenty items that measure a general factor of generativity. Objective: This study examined the factorial validity and internal consistency of the adaptation of the Loyola Generativity Scale to Spanish for its use with Dominican teachers. Method: Two competitive factor models were tested based on previous existing literature, and a third model with method effects associated to negatively worded items was added, in order to find the best fitting solution for this sample. In order to examine the factorial validity of the three models, three Confirmatory Factor Analysis (CFA) were performed, and internal consistency and composite reliability indices were also calculated. Results: In general, the three models showed good psychometric properties. However, the third model that considered a general factor of generativity together with a method effect factor showed the best fit for this sample. Conclusion: Implications for an adequate measurement of generativity are discussed. More research is needed to examine if these properties are stable on different samples from different populations.

El Cuestionario de Salud General (General Health Questionnaire, GHQ) fue diseñado para medir la morbilidad psiquiátrica menor evaluando el funcionamiento “sano” y la aparición de nuevos síntomas molestos. Entre sus versiones, la de 12 ítems es una de las más usadas. La validez y fiabilidad del GHQ-12 ha sido puesta a prueba en muestras provenientes de distintas poblaciones. En la versión española, los estudios han llegado a distintas conclusiones de estructuras de uno, dos y hasta tres factores. El objetivo de la esta investigación es presentar evidencia adicional de la validez factorial de la versión española del GHQ-12, a través de modelos factoriales competitivos. Tres muestras de trabajadores (N = 525, 414 y 540) fueron utilizada para poner a prueba una serie de modelos sustantivos encontrados previamente en la literatura española e internacional. Los resultados mostraron que los modelos multidimensionales tenían correlaciones inter-factores de moderadas a altas (desde .29 a .76), pero no lo suficiente para poner en peligro la validez discriminante de los modelos. Los modelos que presentaron el mejor ajuste fueron la solución origina de Graetz (1991) y la solución exploratoria de tres factores ofrecida por Rocha et al. (2011), una solución multidimensional de tres factores con correlaciones entre ellos. La conclusión es que hay una estructura multidimensional de tres factores que subyace a los ítems del GHQ-12.