Validação de um instrumento de avaliação de desempenho de coordenadores de curso pela teoria da resposta ao item

• Autor(a) principal: Mello, Luciany Abreu de
• Data de Publicação: 2014
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Manancial - Repositório Digital da UFSM
• dARK ID: ark:/26339/001300000ftn4
• Texto Completo: http://repositorio.ufsm.br/handle/1/4700
• topic: Avaliação de desempenho; Competências do coordenador de curso; Teoria da resposta ao item; Performance evaluation; Powers of the course coordinator; Item response theory; CNPQ::CIENCIAS SOCIAIS APLICADAS::ADMINISTRACAO

In recent decades, a new theory called Item Response Theory (IRT) has been progressively studied and successfully applied in the construction and analysis of tests. The IRT proposes a set of mathematical models that relate one or more latent traits of an individual with the probability of this to give right answer to an item, this relation is expressed in such a way that the higher the skill, the greater the probability of getting the item (ANDRADE et al. 2000). This bias, the present study aimed to construct and validate an instrument for evaluating the performance of undergraduate engineers, as perceived by the students, from the Item Response Theory. Data collection was conducted through a questionnaire consisting of twenty-six closed items based on Likert scale with responses ranging into four levels of agreement. The instrument assessed the technical and behavioral skills of the undergraduate program coordinator grouped into four dimensions: interpersonal / leadership, responsibility, communication and productivity / commitment for data analysis, we resorted to the exploratory factor analysis and factor analysis of full information . For the validation of the questionnaire used the Logistic Model Parameters 2 (ML2) proposed by TRI. The results indicate twenty items with good ability to discriminate the response of the individual, giving quality to items. The remaining six items, indicated as unsatisfactory should be reworked, retested and validated in future studies, as well as their dimensions.