Emotional disorders such as anxiety are frequent and are of great importance because of its interference with psychosocial functioning. The study of anxiety-related factors has shown that there are a large number of variables that affect its origin, development, maintenance and alteration. However, appropriate instruments are needed for credible research.Therefore, the validation of one of the most widely used anxiety scale at the international level is of paramount importance. Thus, the purpose of this research was to conduct a validation study of the Beck Anxiety Inventory (BAI) in Portugal. BAI has been analyzed with the rating scale model to evaluate their psychometric quality, for which we have tested the response categories and estimated the parameters of the model, its precision, the degree of data-model fit, the unidimensionality and the absence of differential item functioning. We compared BAI scores with those of other evaluation instruments of anxiety used in Portugal.The sample consists of 1160 adults of both genders (427 men and 733 women) aged 18-82 years (mean age 33.39 years [SD = 11.85]). We used a demographic questionnaire designed for this research, with questions about gender, age, place of residence, ethnicity, educational level, religion, marital status, number of children and occupation. The instruments for measuring anxiety were:Concerning the BAI, the results indicate that its system of four categories is appropriate; that the reliability of the items is very high; that data-model fit is reasonable; that the overall reliability of the people is not excellent, although is adequate, with a value similar to Cronbach's alpha of .79. The results also show that the subjects anxiety level is very low and that most people are below the range of the variable in which the items are located, indicating that it is not designed to measure with high precision people with low levels of anxiety. The analyses also indicate that the measure can be treated as unidimensional.