Equating Design

The method we used for equating was the common-item nonequivalent random group design, in which the participants were submitted to distinct forms of the instrument with some items applied to the total sample (the last design showed in Figure 1.1 of Kolen and Brennan, 2014). Hence, the children from the selected schools were randomly assigned to Form A or B. Additionally, all participants were exposed to a text and a set of related items to equate difficulty (anchor text items). The anchor items were used to compose both the raw scores and the final estimated theta scores of the examinees; therefore, these items were classified as internal to the scores (Kolen and Brennan, 2014).