Habitual travel behaviour: Evidence from a six-week travel diary ROBERT SCHLICH & KAY W. AXHAUSEN Institute of Transportation, Traffic, Highway- and Railway-Engineering (IVT); Swiss Federal Institute of Technology (ETH), Zürich, Switzerland Key words: habits, travel behaviour, travel diary, variability Abstract. This paper introduces different methods to measure similarity of travel behaviour addressing the question of how repetitious travel behaviour actually is. It compares empirical results of the different methods based on the data from a six-week travel diary. In general, the results show that the day-to-day behaviour is more variable if measured with trip-based methods instead of methods based on time budgets. Furthermore, it is confirmed that the similarity declines if the method captures more of the complexity of the travel pattern. It is also shown that travel behaviour is neither totally repetitious nor totally variable. Even for the whole observation period, it is demonstrated that two days always have some common elements. Additionally, it is found that the different methods yield the same pattern of variability for different types of day. Travel behaviour is clearly more stable on work days. Similar results for all methods are also obtained concerning the question of how long the minimum period of observation should be. All measures show that the period should not be less than two weeks if one aims at measuring variability. 1. Introduction A frequent assumption in travel behaviour research is that travel behaviour consists mainly of routines as travellers prefer to repeat those activity patterns with which they were satisfied in the past. This assumption is based on utility maximisation theory, which assumes that humans try to perform activities as efficiently as possible. Out of a set of given possibilities, they choose the activity whose implementation will realise the highest expected utility. This set of possible activities is however limited due to temporal or spatial con- straints or obligations (Hägerstrand 1970). It is reasonable to assume that humans perform a high proportion of actions regularly because those con- straints and obligations do not change every day. For example, one can assume that commuting trips resemble each other concerning mode choice, route choice or departure times due to nearly stable constraints for these trips. Besides, it is unlikely that humans will judge their activities anew every time and attach a new subjective utility to each possible activity or activity pattern. They will rather repeat an activity pattern that offered them a satisfying experi- ence without carefully judging the alternatives. Transportation 30: 13–36, 2003  2003 Kluwer Academic Publishers. Printed in the Netherlands.