Handling Intrahousehold Correlations in Modeling Travel: Comparison of Hierarchical Models and Marginal Models

Abstract

In this study, the necessity for treating intrahousehold correlation was investigated by analyzing two travel behavior indexes, travel time and travel distance, for three important travel motivations (commuting, shopping, and leisure). Data stemming from the 2010 Belgian National Household Travel Survey were used in the analysis. Two model approaches that accommodated intrahousehold correlation were compared, namely, the generalized linear mixed model (GLMM) and the generalized estimating equation (GEE) model. Both model approaches showed that high levels of intrahousehold correlation were present, and therefore the use of models that took into account intrahousehold correlation was strongly recommended. Results indicated that this requirement was the most urgent for noncommuting trips. Moreover, the results showed that the GLMM and the GEE model yielded comparable estimates in the case of normally distributed data. Furthermore, evidence was provided that the more the estimates of the intrahousehold correlation provided by the two approaches differed, the less the homogeneity of the parameters was ensured. In this regard, if one has to choose between the GLMM and the GEE model, the negative consequences of choosing an inappropriate covariance model in the case of GLMM especially favor the selection of the GEE model. Further research is needed to compare the two approaches in the context of nonnormally distributed travel behavior data.