Avaliação do desempenho no vestibular/2003 projeção da classificação

Abstract

The Vestibular Contest belongs the classificatory selective process to candidates to the degree courses of Santa Maria's Federal University and each new accomplishment breaks records in number of registrations. The Program of Entrance to the higher education (PEIES) was institutionalized in 1995 and it is an alternative modality of vestibular exam, seriate, where candidate makes proofs at the end of each one of the three years of the Medium Teaching. That modality already made possible the entrance of eight groups in Santa Maria's Federal University, totaling 3697 classified candidates. At the end of every day of proofs, of the conventional vestibular exam, there was a great interest, on the part of candidates, in accompanying their relative classificatory situation in the vestibular exam, especially after the third day of proofs, when it will only be lacking the accomplishment of the proofs of Portuguese Language and Brazilian Literature, besides the proof of composition of the conventional vestibular exam, in the which should decide among the conventional vestibular exam and PEIES. To render support to the candidates it was grew a methodology to esteem the student's classification, with base in the available information: number minimum, medium and maximum of successes in each course, as well as the number of candidates, the number of vacancies in the course and the average. The method has for main objective to approximate the number of successes for a normal distribution, esteeming the parameters non available with base in the obtained information. The statistical consultant ship it esteemed the 81 candidates' classification, after the end of the selective process, a research was accomplished the approved lists close to in the vestibular exam and PEIES, verifying candidates' situation and the reliability of the method. The analyses of the employed methodological procedures are based fundamentally in theoretical revisions of the Normal distributions and Binomial and the Central Theorem of the limit of Lindergerg in the version of Moive-Laplace, and it collects of data. The method showed excellent productive performance.