Resumo

ANDRADE, Marcia; FERNANDES, Cristiano  e  SILVA, Pedro Luís do Nascimento. Impacto da conglomeração na estimação do coeficiente H da Escala de Mokken. Est. Aval. Educ. [online]. 2013, vol.24, n.55, pp.336-356. ISSN 0103-6831.

The purpose of this paper is to evaluate the impact of cluster sampling design on the estimation of coefficient H, which plays a key role in the construction and quality evaluation of a Mokken scale. For this purpose we considered as our reference population junior high school students, in public schools, in the urban area of the state of Rio de Janeiro, who had taken the Prova Brasil 2007. The responses to a set of 10 dichotomized items that measure the economic capital of the students’ families were used to calculate H. Repeated samples were selected from the same reference population using two sampling designs: simple random without replacement cluster sampling and simple random sampling without replacement. The results show a significant effect of cluster sampling design on the estimation of the Loevinger H coefficient. The possibility of using multilevel models must also be evaluated.

Palavras-chave : Scalability Coeficients; Non Parametric Item Response Theory; Complex Sampling Design; Socioeconomic Level.

        · resumo em Português | Espanhol     · texto em Português     · Português ( pdf )