Apresentação e evidências iniciais de validade do “Teste de Abordagem de Aprendizagem : identificação do pensamento contido em textos, segunda versão (TAP-Pensamento 2)”

Abstract

The methodologies available for studying the theory of approaches to learning have demonstrated important limitations: the phenomenographic method presents the problem of judges' bias and difficulty in applying it to large samples; the method of using the self-report questionnaires presents the problem of the respondent's biases; and the SLAT-Thinking performance test presents an issue related to the high probability of answering correctly by chance, which can lead to many false positives. This work presents the advances that can be obtained from the use of an instrument that was recently developed from SLAT-Thinking as a way to deal with these limitations, the performance test SLAT-Thinking 2. The structure of the new test is presented, demonstrating the theoretical basis used to create its answer options and the constitution of form A and form B of the test. The instrument's advantages over previous methods are demonstrated through the analysis of the results of its application to a sample of high school students, composed of 129 students from a public school in Belo Horizonte. Based on the responses of these students, the rates of correct answers within the structure of SLAT-Thinkng 2 are compared with the rates of correct answers that students would have achieved had the test structure been similar to that of SLAT-Thinking. We observed that form A of SLAT-Thinking 2 was able to prevent, on average, 31% of false positives that would be characterized as correct responses in the structure of SLAT-Thinking and that form B was able to prevent 26%. In addition, the incorrect answers made by the students were categorized based on errors typical of the use of the superficial approach in reading and analyzing texts. This categorization made it possible to investigate the categories of errors prevalent in the test and to compare the responses of different sexes and school grades. The predominant error category was the same for both form A and form B of the test and the categories of error predominant for items varied. Regarding the comparisons of sexes and school grades, no relevant differences were found. The implications of these results are discussed. We conclude that the results found with the application of the instrument are promising and demonstrate that SLAT-Thinking 2 can be a valuable asset for the use of researchers and professionals in the educational field.