expectation. One might expect that those households look for space and a relatively quiet area
to raise children. On the other hand a lot of opportunities (e.g. schools, sports) are available in
a big city, possibly leading to a relatively lower dislike of living in such a big city. Some other
effects, namely the positive signs of the coefficients for a larger municipality and for living in
an apartment can be seen as self-selection effects.
In contrast to a big city, the coefficient for living in a small town (less than 10.000
inhabitants) is positive. The sign becomes even more positive for people currently living in a
bigger municipality. Thus, respondents living currently in a larger municipality on one hand
have a lower dislike for living in a big city compared to those living in smaller municipalities
but also relatively like living in a small town to a higher extent. This means however, that
people currently living in a municipality with more than 50.000 inhabitants relatively dislike
living in a medium sized city more than people currently living in a smaller municipality. This
again can be regarded as a self-selection effect. As for larger municipalities, respondents
driving a car on gasoline also value living in a small town relatively higher. Finally,
respondents within the lowest income class seem to value living in a small town lower. In
combination with the earlier described effect of respondents with lower incomes dislike living
in a big city to a higher extent, points to the direction that respondents within the lowest
income class seem to prefer to live in a medium sized city.
Looking at the trip related factors (i.e. travel cost and travel time) several heterogeneity
effects can be observed. Respondents living in a region in Holland suffering from traffic
congestion problems seem to value travel cost more negatively than people living outside
these regions. Besides that, respondents living in an apartment value travel cost less
negatively as is the case for people who have the possibility to work at home. As expected,
people with a higher income value travel costs less negatively, leading to a higher VOT for
higher income classes. Furthermore, respondents with departure time constraints seem to
value travel time less negatively.
A special situation occurs for respondents with longer actual travel times. These respondents
have a significant other coefficient for both travel time and cost. As can be seen in table 11,
the travel time and travel cost coefficients are less negative for higher actual travel times.
However on basis of this result the conclusion of respondents with a higher actual travel time
(above 50 minutes) being less sensitive for travel costs or travel time cannot be drawn.
23