Literacy in mathematics and reading and its association with cultural behaviors. A LCRA application

PORCU, MARIANO;TEDESCO, NICOLA

2013-01-01

---

Abstract

Considering the 2010 wave of the national sample survey on the math and reading literacy of fifth grade pupils in Italy, this paper aims to identify latent classes of subjects from a multi-way table of categorical items. We analyze the responses provided by pupils on a questionnaire in order to collect information on their sociocultural backgrounds and their perception of the school climate. Fitting a Latent Class Regression Analysis model pupils are classified into three classes based upon membership probabilities (posterior probabilities) estimated directly from their response pattern to the items of the questionnaire and considering as covariates their proficiency in mathematics and reading.