RESUMO
BACKGROUND: In the XXI century anxiety disorders have become the most prevalent in Mexico, excessive worry is one of the first features that allows its identification. Have a valid and reliable instrument to assess the pathological worry is essential to identify the disorder from the beginning. METHOD: The aim of this study was to analyze the psychometric properties of the Penn state worry questionnaire (PSWQ) in the four different versions used in clinical contexts in Spanish-speaking countries: the original scale (PSWQ-16), the direct form of the scale (PSWQ-16D) and two abbreviated versions (PSWQ-11 y PSWQ-8). A total of 2,267 participants were given those versions of the questionnaire. RESULTS: Our results suggest that the original scale (16 items) fits to two related factors model. However, the analysis of the PSWQ version with all the items in its direct form and short versions (11 and 8 direct items), together with other arguments indicate that it is more convenient to conceive a one dimensional construct. Besides high internal consistency and test-retest reliability, and adequate concurrent and discriminant validity. CONCLUSIONS: Results suggest use of the short versions (11 and 8 direct items) which shows a one-dimensional structure and the best goodness of fit indices. Results are discussed and future research are suggested.