STATISTICAL COMPETENCIES IN THE TRAINING OF MATHEMATICS TEACHERS IN 2017 ENADE: AN IRT APPLICATION Sandra Cristina Martini Rostirola, Elisa Henning, and Ivanete Zuchi Siple Santa Catarina State University, Brazil sandra.rostirola@ifc.edu.br The National Students' Performance Examination [Exame Nacional de Desempenho dos Estudantes– Enade] is a large-scale assessment instrument for Brazilian higher education programs whose results, along with other quality indicators, allow for an evaluation of the country's higher education. This work analyzed 2017 Enade questions for the mathematics degree program regarding the statistical competencies of future educators. The quantitative approach methodology allowed analyzing 10,869 participants' responses through Item Response Theory (IRT) using the Three-Parameter Logistic Model. The results indicate evidence of weakness in questions related to probability and statistics regarding t levels of difficulty and discrimination in addition to reflecting discrepancies between the statistical content in the official test descriptors and those found in the questions. INTRODUCTION The National Students' Performance Examination [Exame Nacional de Desempenho dos Estudantes–Enade] is one of the assessment instruments that allows measuring the quality of Brazilian higher education, both for traditional classrooms and for distance education. Enade evaluates the performance of students in the last period of the undergraduate programs regarding the syllabus defined in each program's curriculum and the development of competencies and skills required for a deeper general and professional training. (Instituto Nacional de Estudos e Pesquisas Educacionais Anísio Teixeira [INEP], 2022). This evaluation has been carried out annually since 2004, and the included educational areas change each year in a rotation system. In 2017, it included the areas of Engineering, Information Technology, and Teaching Degrees, including the degree in mathematics. The mathematics degree program aims to help the development of mathematics educators' competencies, including curriculum, evaluation, methodological aspects, and the selection and production of teaching materials. Pedagogical knowledge, ability to solve problems, and ability to establish mathematical models and conjectures for content areas such as statistics are included among the teaching competencies. The latter is the main interest in this study because understanding of statistics structure and concepts allows one to understand and be part of the social environment in which he or she is inserted, with information interpretation, data analysis, and decision making providing an opportunity to experience citizenship. To measure the future mathematics teacher's competencies, Enade is based on a competency matrix that guides the skills to be assessed through the examination. This study analyzed the 2017 Enade's questions for the mathematics degree program and future educators' statistical competencies, measuring the quality of the test and of each item (question). The 2017 Enade examination consisted of 40 questions, five of which were essays and 35 of which were multiple-choice questions with five alternatives—one correct answer and four incorrect distractors. From that set of questions, two—items 20 and 21—comprise this study's analysis base because they are related to statistics. This work aims to assess the statistical competencies of the mathematics degree program's students who took the 2017 Enade. It should be noted that the mathematical model employed in Enade is Classical Test Theory (CTT). However, for the purpose of a wider scope, the items were analyzed using Item Response Theory (IRT). METHODS The adopted methodology is the quantitative approach from a documental perspective using Enade’s 2017 test for the mathematics teaching degree program to analyze two questions (items) related to the contents of probability and statistics, as described by the Reference Curriculum in the Summary Report (INEP, 2018). The test consisted of 40 questions, five essays and 35 multiple-choice questions. The data regarding the answer pattern from 10,869 participants were collected through microdata from INEP. All participants who answered at least one item from the multiple-choice questions were included. Classical Test Theory (CTT), probability models of Item Response Theory (IRT), and the Nominal Response Model (NRM) were applied to the response pattern set of the 35 objective items. ICOTS11 (2022) Contributed Paper - Refereed (DOI: 10.52041/iase.icots11.T14D4) Rostirola, Henning, & Siple In S. A. Peters, L. Zapata-Cardona, F. Bonafini, & A. Fan (Eds.), Bridging the Gap: Empowering & Educating Today’s Learners in Statistics. Proceedings of the 11th International Conference on Teaching Statistics (ICOTS11 2022), Rosario, Argentina. International Association for Statistical Education. iase-web.org ©2022 ISI/IASE