Evidências de validade e fidedignidade do teste de conhecimento numérico infantil: uma nova medida de cognição numérica

Abstract

Numerical cognition can be defined as the ability of the human brain to deal with the notion of quantity and with numbers. It is subdivided into two systems, one of which is innate and related to non-symbolic aspects, while the other involves symbolic aspects that are learned culturally over the years. Despite the importance of numerical cognition for cognitive development, there are few instruments with evidence of validity and reliability to assess this cognitive domain in early childhood education, both nationally and internationally. This work was organized into two studies, which were divided into two chapters. The first chapter presents an overview of studies on numerical cognition, its development throughout life and its neuropsychological bases. The relationship between language and this construct, cognitive predictors and a brief characterization of the most common numerical cognition dysfunction, Developmental Dyscalculia, are also discussed. Finally, a description of the main instruments for assessing numerical knowledge in national and international contexts was presented. It was therefore found that there is a lack of instruments that assess symbolic and non-symbolic processing of numerical cognition, allowing for a comparison between processing in a broad age range. Even more specifically, there is a lack of instruments that assess this construct in pre-school children in Brazil. It is worth noting that this chapter will be published in the future as a chapter in a book in the field. The second study is characterized as empirical research. The main aim of the research was to develop and obtain evidence of validity and reliability for the Children's Numerical Knowledge Test (TCNI), which is an instrument designed to assess symbolic and non-symbolic aspects of numerical cognition, targeting children aged five to eight. The study was written in the format of a scientific article and was therefore subdivided into the following topics: Introduction, Objectives, Method, Results, Discussion and Conclusion. The study involved 262 children aged between five and eight. Evidence of validity based on relationships with external measures was analyzed, such as concurrent criterion validity, discriminant validity, developmental norm validity and predictive criterion validity, as well as evidence of validity based on internal structure, through exploratory factor analysis. To assess evidence of reliability, the test-retest methodology was used, analyzing Cronbach's Alpha and Mc Donald's Omega. At the first meeting, the TCNI and the Raven's Colored Progressive Matrices Test were administered to all the participants. At the second meeting, half of the participants answered the phonological awareness tasks (to analyze discriminant validity) and the Numeracy Screener (to analyze concurrent criterion validity) and the other half of the participants took the TCNI again (to evaluate test-retest reliability). The instruments were applied individually and took place in a reserved room within the participants' own schools. After collection, the data was tabulated and analyzed using Jamovi Version 2.4 statistical software. The results showed that the TCNI is a promising instrument, since good evidence of validity and reliability was found, as well as filling a gap in the scientific literature, since the TCNI makes it possible to assess numerical cognition in Brazilian preschool children.