Regresión multinomial en el análisis de factores asociados en el consumo de alcohol en estudiantes adolescentes

Abstract

The present study shows the relevance of the implementation of the multinomial logistic regression model (MLR) for risk analysis and identification of risk factors for a condition or event such as alcohol consumption in students. A total of 395 young people residing in Portugal and belonging to two schools in the country were analyzed, considering their mathematics qualifications, information about their family history and inclusive their academic data. This study concludes that the increase in the scale of alcohol consumption is related to the fact of studying in a particular school, whether living under the guardianship or not of the parents influences the probability of having a greater scale of consumption. The likelihood ratio and Akaike criteria have been used to determine the variables that influence alcohol consumption