Brazilian validation of the attitudes scale with regard to the AIDS: the Rasch model

Abstract

Introduction: Rasch analysis has been identified as a method of statistical analysis that estimates the level of difficulty of the issues and the level and ability of participants, separately and independently. Objective: To test the validity of the Brazilian version of an attitude scale with regard to HIV/AIDS, according to the Rasch model. Method: We conducted a cross-sectional study with students of a teacher-training course, using an attitude scale. Data analysis was performed using psychometric analysis by the Rasch method, using Big Steps statistical software. Results: We found that 23% of the questions were not in accordance with the model. Nonetheless, the instrument presented a rate of 92% for reliability, separation index of items by 4.76 logits, and good internal and external consistency. Conclusions: The scale demonstrated good reliability and correlation among the questions and can contribute to the deepening of discussions and reviews of the measures to prevent HIV/AIDS.