Mensuração de Atitude: Proposição de um Protocolo de Elaboração de Escalas

Abstract

This theoretical essay aims to study how scales are developed and through which procedures they can be considered valid and suitable for use as legitimate scientific instruments. In this perspective, this paper's objective was to develop a protocol for constructing scales to measure attitude. The proposed protocol is configured as a logical meeting of steps based on theorists such as Allport and Hartman (1925), Thurstone (1928), Likert (1932), Campbell and Fiske (1959)and Bock (1972), which permeate all aspects of drafting scales, including construct definition, the choice of the scale itself, item preparation, scale purification, and finally its validation. At the end of the study, we present a protocol for the preparation of specific scales to measure attitude that differs from existing protocols in Churchill (1979), Rossiter (2002) and DeVellis (2003). This is in order to unite for the first time a set of promising techniques, primarily the objective delineation of the construct using focus group methodology, the proposition of an inherently dichotomous scale, scale purification via item response theory (IRT), and predictive validity.