Understanding Travel Behaviour and Factors Affecting Trip Rates



Understanding Travel Behaviour and Factors Affecting Trip Rates

Authors

K Jahanshahi, I Williams, X Hao, WSP Group, UK; R Cambery, Department for Transport, UK

Description

Making the use of the GB National Travel Survey dataset (NTS), this paper analyses the influence of various socio-economic and accessibility factors on trip rates, and whether these trip rates are stable over time.

Abstract

Making the use of the GB National Travel Survey dataset (NTS), this paper analyses the influence of various socio-economic and accessibility factors on trip rates, and whether these trip rates are stable over time.

The motivation of the research was:
? to examine whether the current variables in use within the UK Department for Transport's National Trip End Model (NTEM) are sufficient to explain the trip rates,
? to provide advice on how NTEM could be improved.

Various home based, non home based and escort trip purposes were individually analysed making innovative use of negative binomial regression, a method designed specifically for the analysis of count data. Negative Binomial is an extension to Poisson regression, which can account for greater than Poisson variation and is based on the negative binomial distribution. It has been shown that using Poisson regression for analysing trip rates may overestimate the importance of some of the variables within the model as the statistical tests can be biased.

The results suggest that the relative importance of variables varies between trip purposes. The demographic factors (i.e. age, gender, car ownership, household structure) and area type, which are in use within the current version of NTEM do have a significant effect for the majority of trip purposes on an individual's decision to make a trip. However, it has been found that the current NTEM model can be significantly improved by the inclusion of socioeconomic class and income variables. One interesting result is that income and socio-economic class are not highly inter-correlated within this model and act reasonably independently in influencing trip rates. However, if just one were to be selected for modelling purposes, the socio-economic class was shown to play a more important role than income group, for each individual trip purpose.

Another finding relates to the importance of the inclusion of interaction between variables. It is shown that in most cases, the effect of interaction between area type and socio-economic characteristics as well as those between socio-economic factors themselves (e.g. work status and gender) should be considered within the model in addition to the separate influence of each of these variables individually.

Regarding the effect of accessibility variables, the influences of area type and population density are proved to be highly correlated. Between the two, area type has shown a stronger role in explaining changes in trip rates.

NTEM predicts non home based trip rates as a direct multiplier of the number of home based trips. It also assumes that the same factors influence home based trips and escort trips as it represents both in one common trip purpose. These assumptions have been tested, with results suggesting that this model could be improved in two ways:

? When the number of home based trips are controlled, the socioeconomic and accessibility variables have independent and significant effects on explaining the number of non home based trips so their explicit inclusion would improve the modelling of non home based trips.
? The estimation of the number of escort trips is much improved by including both the characteristics of the escorter and the characteristics of the escortees.

No significant behavioural change has been identified in trip rates over time for any individual trip purpose. The apparent reductions in overall trip rates seem mainly to be the result of a combination of technical changes within the NTS survey methodology coupled with changes in the demographic pattern of the population.

Publisher

Association for European Transport