Household travel survey data has been playing a key role in transportation planning. Over the years, many efforts have been made to investigate the methodology to supplement/replace household travel survey in travel demand modeling, e.g. transferring travel demand models, transferring or simulating household travel data. However, there are two key issues that had not been addressed in the literature: the contextual variability issue in data transferability studies and the sample size problem in the national household travel survey (NHTS) itself.

The first objective of this study is to identify the contextual effect on travel behavior and quantify how much behavioral variability is due to contextual variations. Hierarchical random-effect models are developed to quantify household travel features variability across metropolitan statistical areas (MSAs) in the United States. It is found that for most travel characteristics, the influence of contextual settings cannot be ignored. It is suggested that transferring the NHTS data directly across geographical areas may lead to incorrect travel forecasting model parameters. Therefore, the second part of the thesis focuses on applying model-based approaches (i.e., Small Area Estimation (SAE) methods) to generate household travel characteristics based on small sample household travel data and auxiliary area (i.e., census tract) information. Specifically, three SAE methods—the Generalized Regression Estimators (GREG) method, the Empirical Best Linear Unbiased Predictor (EBLUP) method, and the Synthetic method are applied to generate two household travel characteristics, at both the census tract and the household levels.

The proposed methods of dealing with household travel survey data in this research and the analysis findings will provide a useful tool for practitioners, planners and policy makers in transportation analyses. Understanding the living environment's influence on travel behavior shed lights on transportation decisions that involve the transportation-land use relationship, increasing mobility and accessibility for communities, and coping with changes of travel due to demographic change. My research has also shown that SAE methods are promising approaches to generate unbiased aggregate and disaggregate household travel characteristics data by incorporating population auxiliary information and local household travel survey data.