Avaliação comportamental de leitura e escrita : estudos transversais de um instrumento baseado na equivalência de estímulos

Abstract

Reading and writing skills can be understood as a network of equivalence relations between stimuli (for example, written words, dictated words and images) and stimuli and responses (naming, writing, etc.). This conception has been the basis for assessment tools and teaching programs. The Behavioral Assessment of Reading and Spelling (BARS) is an instrument of 15 online tasks that involves matching-to-sample, nomination, reading and writing. The present study consists of two cross-sectional studies carried out based on the database from BARS in which were analyzed the results of 2388 participants, aged 6 to 12 years. The evaluations happened between 2009 and 2017, in 16 cities in 7 states of Brazil. The first study aimed to (I) characterize the network of relationships measured in BARS, (II) to investigate the relationship between the skills assessed by the instrument and reading and writing in dictating condition and (III) to verify the theoretical proposition of integration of networking skills. To assess the explanatory relationship between the skills, the Generalized Additive Models (GAM) was used. To verify the proposition of the integration of skills, an integration index was calculated for each student. This index assesses how similar the performances in different tasks were. All skills assessed showed a significant relation with reading and writing. The two abilities that directly related to reading the most (greater magnitude) were matching written words with dictated words and consonant naming. Matching written words with pictures was the skill that related the most with spelling to dictation. It was found that the integration between skills were higher among students with better overall results in the assessment, as expected by the stimulus equivalence paradigm. Future studies can investigate the effects of teaching different reading and writing skills in order to better understand the learning process. The second study aimed to conduct analyzes in order to compare the participants' performance by gender and age group, with ways to understand how such variables have been related to the high or low performance in ACoLE. The cross-sectional analysis showed that the girls performed better in word reading tasks and constructed response spelling. A better average performance was also observed for older students, except for those above ten years old. Both results are in line with what the literature in the area has pointed out about students with reading and writing difficulties.