The development and validation of the math interest behavioral indicators scale

The goal of the paper was twofold. First, we intended to show that a prevalent approach to assessing domain-specific interests and attitudes through generalized items (I like math) has serious flaws, especially when applying to children in large-scale surveys. We demonstrated these flaws with the Scale of attitudes to math used in Trends in International Mathematics and Science Study (TIMSS). Second, we showed an alternative approach to assessing interests through behaviorally anchored items developed in the framework of the interest development theory. The new Math Interest Behavioral Indicators Scale (MIBI) proved not only excellent psychometric characteristics, but it also discriminated children who were from advanced math programs and who were not, what the TIMSS scale did not do. The importance of theory for scale development is discussed.