21 2Análise Psicométrica de Versões Curtas do Big Five Inventory em Universitários Mexicanos 
Home Page  


Avaliação Psicológica

 ISSN 1677-0471 ISSN 2175-3431

PRIMI, Ricardo; ALMEIDA, Leandro S.; NAKANO, Tatiana de Cassia    CAMPOS, Carolina Rosa. Standardization of the of Reasoning Tests Battery (BPR-5) using Item Response Theory. Aval. psicol. []. 2022, 21, 2, pp.127-139. ISSN 1677-0471.  https://doi.org/10.15689/ap.2022.2102.20136.01.

The Reasoning Tests Battery (BPR-5) is widely used in Brazil for the assessment of intelligence. It has three different forms: children (1st to 6th grade of elementary school), Form A (7th to 9th grade of elementary school) and Form B (high school and higher education). This study describes the steps followed to create a common metric across the forms. It aimed to: (a) calibrate the items of the three forms using Rasch model, link items and equate subjects' scores across forms using the anchoring of common items method, (b) update the norms by expanding the representativeness of the samples by producing norms for different combinations of age, education and sex, and (c) describe the developmental patterns of the BPR-5 subtests across a wide age range, from 6 to 52 years. We present two studies, the first reporting the calibration of item and person parameters with the Rasch model and a good fit to the model. The second illustrated the use of multiple regression analysis to create norms for the psychological tests considering the variables age, education and gender as predictors of the BPR-5 scores. These three variables had significant effects explaining 8% (Abstract Reasoning), 13% (Verbal Reasoning), 12% (Spatial Reasoning), 8% (Numerical Reasoning) and 22% (Mechanical Reasoning) of the variance.

: intelligence; multiple regression; equating; BPR-5; norms; Item Response Theory; Rasch Model.

        · | |     · |     · ( pdf )

 

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License