International Choice Modelling Conference, International Choice Modelling Conference 2015

Font Size: 
Triggers of behavioral change
Konstadinos Goulias, Jae Hyun Lee, Adam Wilkinson Davis

Last modified: 11 May 2015


Triggers of behavioral change

Jae Hyun Lee

Adam Wilkinson Davis

Konstadinos G. Goulias


Van Acker and Witlox (2009) in their overview article about land use, activity-travel, and attitudes/perception argue persuasively that one should analyze the relationships among behavior, social- demographic changes, and the built environmental jointly to account for complex relationships.  In a similar vein of development, Mokhtarian and Cao (2008) acknowledge the significant relationships between land use characteristics and travel behavior and caution us to be aware of potential pitfalls in data analysis techniques that are not designed to discover causality.  In fact, they conclude that a longitudinal data collection design has the potential to reveal the causal mechanism underlying the observed behavior, can protect us from finding spurious relationships, this type of design has (of course) information to study time precedence, and offers strength in helping discover association among variables. 


All this points to the advantages of having individual and household longitudinal data (e.g., a travel diary of the same household members repeatedly answered for many years) and longitudinal data measuring the attributes of the geographical space surrounding the residences of these households.  This database exists in the United States and it is called the Puget Sound Transportation Panel (PSTP). PSTP spans a long period from 1989 to 2002 with data that are harmonized and suitable for longitudinal analysis (Goulias et. al., 2003).   This enabled the use of multi-equation multivariate statistical techniques to unravel causality in behavioral variables at multiple levels (Goulias, 2002) as suggested by Mokhtarian and Cao (2008).  More recently the PSTP database was enriched with longitudinal land use information that allows examination of residential relocation jointly with demographic changes and their influence on activity and travel behavior.  In one analysis using nine waves of PSTP and age as the time progression variable, Dalal and Goulias (2013) found significant differences in the initial status and growth trajectories for households that anticipate a life cycle change (e.g., a new child in the household and the loss of an adult). In an exploratory analysis using all ten waves of PSTP and elapsed time from the beginning of the panel as the time progression variable, Goulias et al., (2014), used Mixed Markov Latent Class models to illustrate a few basic longitudinal heterogeneity properties of the behavioral data of household variables such as daily amount of time in out of home activities, daily amount of time traveling to work, number of trips traveling alone by car, and number of trips traveling with relatives.  Key discovery in this type of analysis is the identification of households that follow different trajectories in their behavioral evolution.


In this paper, we expand our previous analysis of the ten waves of PSTP to explore the longitudinal relationships among the variables in Figure 1 that include a variety of household composition variables, different aspects of activity-travel behavior, and examples of land use variables such as density and diversity of activity opportunities surrounding homes.  The analysis here is using data from the two-day PSTP travel diary and 230 households participating in all ten waves of the panel.


The main objective of this analysis is to identify what triggers behavioral change in activity and travel and to quantify and compare the size of these triggering effects using different methods.  Our search for triggers includes the birth of a child, child leaving the household, and the entry or exit from the labor force of a household member.  In the land use changes we identify moves from high density and diversity environments to lower density and the opposite. The methods used to analyze the data include longitudinal Mixed Markov Latent Class and Mixed Latent Growth models that allow to not only account for behavioral heterogeneity at each time of observation but to also test for the existence of multiple latent trajectories of change and the role played by triggers in the shaping of these trajectories. . 



Please see PDF file accompnying this abstract 


Figure 1 Average Trends of Panel Participants








Dalal P. and K.G. Goulias (2013) Preparation and Adaptation Processes in Travel Behavior Dynamics: A Latent Growth Modeling Approach. Geotrans Report 2013-01-01. Santa Barbara, CA.


Goulias, K. G. (2002). Multilevel analysis of daily time use and time allocation to activity types accounting for complex covariance structures using correlated random effects. Transportation, 29(1), 31-48.


Goulias, K. G., Kilgren, N., & Kim, T. (2003). A decade of longitudinal travel behavior observation in the Puget Sound region: sample composition, summary statistics, and a selection of first order findings. In 10th International Conference on Travel Behavior Research, Moving through nets: The physical and social dimensions of travel, Lucerne.


Goulias K. G., J.H. Lee, and A.W. Davis (2014) Longitudinal Mixed Markov Latent Class Analysis of the 1989 to 2002 Puget Sound Transportation Panel Data.  Paper presentation at the 94th Annual Meeting of the Transportation Research Board, Washington, D.C., January 11-15, 2015. Also published as GEOTRANS Report 2014-8-04, Santa Barbara, CA.


Veronique Van Acker et Frank Witlox, « Why land use patterns affect travel behaviour (or not) », Belgeo, 1 | 2009, 5-26 (


Mokhtarian, P. L., & Cao, X. (2008). Examining the impacts of residential self-selection on travel behavior: A focus on methodologies. Transportation Research Part B: Methodological, 42(3), 204-228.



Conference registration is required in order to view papers.