MODELING THE INDIVIDUAL CONSIDERATION OF
TRAVEL- RELATED STRATEGIES
Xinyu Cao
Department of Civil and Environmental Engineering
One Shields Avenue
University of California, Davis
Davis, CA 95616
voice: ( 530) 754- 7421
fax: ( 530) 752- 6572
e- mail: xycao@ ucdavis. edu
and
Patricia L. Mokhtarian
Department of Civil and Environmental Engineering
and
Institute of Transportation Studies
One Shields Avenue
University of California, Davis
Davis, CA 95616
voice: ( 530) 752- 7062
fax: ( 530) 752- 7872
e- mail: plmokhtarian@ ucdavis. edu
June 2003
This research is funded by the University of California Transportation Center.
i
TABLE OF CONTENTS
LIST OF TABLES AND FIGURES............................................................................................... ii
EXECUTIVE SUMMARY ........................................................................................................... iv
1. INTRODUCTION ...................................................................................................................... 1
2. THE DATA AND VARIABLES................................................................................................ 2
2.1 Data ............................................................................................................................... ... 2
2.2 The Travel- related Strategies............................................................................................ 3
2.2.1 Strategy descriptions.............................................................................................. 3
2.2.2 Identification of bundles of strategies.................................................................... 6
2.3 Explanatory Variables....................................................................................................... 8
3. MODELS OF CONSIDERATION OF EACH STRATEGY................................................... 12
3.1 General Specification and Interpretation Issues ............................................................. 12
3.2 Auto Improvement .......................................................................................................... 16
3.2.1 Buy a car stereo system........................................................................................ 20
3.2.2 Get a better car ..................................................................................................... 22
3.2.3 Get a fuel efficient car.......................................................................................... 24
3.3 Mobile Phone.................................................................................................................. 25
3.4 Work- Schedule Changes................................................................................................. 27
3.4.1 Change work trip departure time ......................................................................... 29
3.4.2 Adopt flextime ..................................................................................................... 30
3.4.3 Adopt compressed work week ............................................................................. 33
3.5 Hire Somebody to Do House or Yard Work................................................................... 36
3.5.1 The model with all respondents ........................................................................... 36
3.5.2 The model with only non- adopters ...................................................................... 38
3.6 Mode Change.................................................................................................................. 39
3.7 Home- based Work .......................................................................................................... 43
3.7.1 Buy equipment/ services to help you work from home ........................................ 46
3.7.2 Telecommuting .................................................................................................... 47
3.7.3 Start home- based business or put more effort into an existing one ..................... 50
3.8 Residential/ Employment Relocation .............................................................................. 54
3.8.1 Change jobs closer to home ................................................................................. 55
3.8.2 Move your home closer to work .......................................................................... 57
3.9 Alter Employment Status................................................................................................ 59
3.9.1 Work part- instead of full- time ............................................................................ 60
3.9.2 Retire or stop working ......................................................................................... 63
4. SUMMARY AND CONCLUSIONS ....................................................................................... 67
4.1 Overview of the Models.................................................................................................. 67
4.1.1 Overview of relationships of consideration to the contemporaneous explanatory
variables ........................................................................................................................ 72
4.1.2 Overview of relationships of consideration to prior adoption ............................. 74
4.2 Comparison between Hypotheses and Results ............................................................... 78
4.3 General Conclusions and Policy Implications ................................................................ 80
ACKNOWLEDGEMENTS.......................................................................................................... 82
REFERENCES ............................................................................................................................. 82
ii
LIST OF TABLES AND FIGURES
Table ES- 1. Summary of the Individual Models ( grouped by conceptual bundles) .............. ix
Table ES- 2. Relationships between Former Adoption and Current Consideration of
Strategies ( grouped by conceptual bundles) ................................................................ xiii
Table ES- 3. Relationships between Former Adoption and Current Consideration of
Strategies ( grouped by factor- based bundles).............................................................. xiv
Table ES- 4. Summary of Hypotheses and Results ............................................................... xv
Table 1. Demographic Characteristics of Sample Used in This Analysis .............................. 3
Table 2. Conceptual and Factor- based Bundles of the Travel- related Strategies ................... 7
Table 3. The Distribution of Former Adoption and Current Consideration of Strategies .... 14
Table 4. Models of Consideration of Auto Improvement Strategies ( Bundle 1).................. 18
Table 5. Model of Consideration of “ Buy a Car Stereo System” ......................................... 21
Table 6. Model of Consideration of “ Get a Better Car” ....................................................... 23
Table 7. Model of Consideration of “ Get a Fuel Efficient Car”........................................... 25
Table 8. Model of Consideration of “ Get a Mobile Phone” ( Bundle 2) ............................... 26
Table 9. Models of Consideration of Work- Schedule Changes ( Bundle 3) ......................... 28
Table 10. Model of Consideration of “ Change Work Trip Departure Time”....................... 30
Table 11. Model of Consideration of “ Adopt Flextime” ( all respondents) .......................... 31
Table 12. Model of Consideration of “ Adopt Flextime” ( only non- adopters) ..................... 32
Table 13. Model of Consideration of “ Adopt Compressed Work Week” ( all respondents) 34
Table 14. Model of Consideration of “ Adopt Compressed Work Week” ( only non- adopters)
............................................................................................................................... ....... 35
Table 15. Model of Consideration of “ Hire Somebody to Do House or Yard Work” ( all
respondents) .................................................................................................................. 37
Table 16. Model of Consideration of “ Hire Somebody to Do House or Yard Work” ( only
non- adopters) ................................................................................................................ 38
Table 17. Model of Consideration of “ Change from Driving Alone to Some Other Means”
( personal vehicle/ motorcycle commute mode users).................................................... 41
Table 18. Models of Consideration of Home- based Work ( Bundle 6)................................. 45
Table 19. Model of Consideration of “ Buy Equipment/ Services to Help You Work from
Home” ........................................................................................................................... 47
Table 20. Model of Consideration of “ Telecommute” ( all respondents).............................. 48
Table 21. Model of Consideration of “ Telecommute” ( only non- adopters)......................... 49
Table 22. Model of Consideration of “ Start Home- based Business or Put More Effort into
an Existing One” ( all respondents) ............................................................................... 51
Table 23. Model of Consideration of “ Start Home- based Business” ( only non- adopters)... 53
Table 24. Models of Consideration of Residential/ Employment Relocation ( Bundle 7) ..... 55
Table 25. Model of Consideration of “ Changing Jobs Closer to Home” ............................. 56
Table 26. Model of Consideration of “ Move Your Home Closer to Work” ........................ 58
Table 27. Models of Consideration of Altering Employment Status ( Bundle 8) ................. 60
Table 28. Model of Consideration of “ Work Part- instead of Full- time” ( all respondents) . 61
Table 29. Model of Consideration of “ Work Part- instead of Full- time” ( only non- adopters)
............................................................................................................................... ....... 63
Table 30. Model of Consideration of “ Retire or Stop Working” ( all respondents).............. 64
iii
Table 31. Model of Consideration of “ Retire or Stop Working ” ( only non- adopters) ........ 66
Table 32. Summary of the Individual Models ( grouped by conceptual bundles) ................. 68
Table 33. Relationships between Former Adoption and Current Consideration of Strategies
( grouped by conceptual bundles) .................................................................................. 76
Table 34. Relationships between Former Adoption and Current Consideration of Strategies
( grouped by factor- based bundles)................................................................................ 77
Table 35. Summary of Hypotheses and Results ................................................................... 79
Figure 1. Section E1 ( Adoption) from the Survey.................................................................. 4
Figure 2. Section E2 ( Consideration) from the Survey........................................................... 5
iv
EXECUTIVE SUMMARY
This report is one of a series of research documents produced by an ongoing study of
individuals’ adoption and consideration of travel- related strategies in response to congestion.
It is widely recognized that congestion has serious consequences for sustainable development.
Governments have been adopting a wide range of measures to alleviate congestion. However,
the limited effectiveness of these strategies has been puzzling policy makers. The gap between
policy assumptions and individuals’ behaviors is believed to greatly affect the effectiveness of
such strategies. Also, the dynamic nature of individuals’ response to congestion further
exacerbates the discrepancy between assumption and reality. Therefore, the primary goal of this
report is to develop disaggregate discrete choice models for the consideration of travel- related
strategies and examine any patterns emerging across the models, in order to better understand the
determinants of individuals’ consideration of each strategy, to improve predictions of the
effectiveness of proposed policies, and to help design more effective policies. In so doing we
also explore the relationship between the earlier adoption of a strategy and its reconsideration,
helping us to further understand the dynamic nature of individuals’ behavioral response to
congestion.
The data for this series of studies come from a 1998 mail- out/ mail- back survey of 1,904 residents
in three neighborhoods in the San Francisco Bay Area: Concord and Pleasant Hill representing
two different kinds of suburban neighborhoods comprising about half the sample, and an area
defined as North San Francisco representing an urban neighborhood comprising the remainder.
The questions in the survey were classified into 11 categories of variables: objective mobility,
subjective mobility, relative desired mobility, travel liking, travel attitudes, personality, lifestyle,
excess travel, adoption and consideration of travel- related strategies, mobility constraints, and
demographic characteristics. For this study, we chose to focus on commuting workers since they
contribute most heavily to peak- period congestion, and are likely to be the most active in the
adoption and consideration of travel- related strategies; the subset of 1283 cases that consists of
commuting workers with relatively complete responses to key questions is used in this analysis.
Binary logit models were developed for the consideration of each individual travel- related
strategy. Each dependent variable, consideration of the given strategy, was defined as a binary
variable, and the other variables were viewed as potential explanatory variables. Generally, the
significance level 0.05 was used to incorporate or release variables in the final “ best” model.
Specifically, based on our initial expectations, we developed binary logit models for the
consideration of each of 16 individual travel- related strategies, which can be conceptually
categorized as low- cost ( in the generalized sense) travel maintaining/ increasing, medium- cost
travel reducing, and high- cost location/ lifestyle change ( see Table ES- 1). Except for the model
of consideration of changing from driving alone to some other means, which is based on personal
vehicle/ motorcycle commute mode users, all other models presented in this summary are
estimated on the full sample ( for some strategies, the full report includes additional models
estimated only on non- adopters of those strategies, for reasons explained in Section 3.1).
Table ES- 1 ( Table 32 in the text) summarizes the variables significant in each model, with
positive and negative signs indicating the direction of effect for each variable. ρ 2 and adjusted
ρ 2 are used to measure the goodness of fit of these models. The ρ 2 s range from 0.183 for the
v
model of consideration of getting a better car, to 0.628 for the model of consideration of “ Move
your home closer to work”. The adjusted ρ 2 s for the models range from 0.158 to 0.615. Since
the market shares ( MSs) for several of the strategies were quite unbalanced, in those cases the
MS ρ 2 was already rather high. To measure the explanatory contribution of the true variables to
the models, we re- estimated the final models with the constant term fixed to zero and computed
the ρ 2 s again. The comparison between ρ 2 s for models with and without the constant term
shows that the true variables in the model always account for at least 87% of the information
explained by the full model, and carry at least 95% of the explanatory power of the model in
more than half of the cases. Thus, even when the statistical achievement of the full model does
not appear to be great compared to the MS model, its contribution to an understanding of the
relevant behavioral mechanisms can be substantial.
The key results of this report are as follows:
Objective mobility: Objective mobility variables are generally positively associated with the
consideration of the travel- related strategies presented in this study. The more an individual
travels for short distance, the more likely she is to consider the low- cost travel-maintaining/
increasing strategies. Whether the large amount of short- distance travel is by
necessity or by choice, the low- cost travel- maintaining/ increasing strategies offer appealing
options for making that travel more pleasant or productive. While both frequency and distance
of short- distance travel influence the consideration of the lower- cost strategies, it is logically
enough not the frequency but the distance of short- distance travel that has a more important
impact on the consideration of the medium- or high- cost travel- reduction strategies.
Subjective mobility: Generally, short- distance subjective mobility variables are positively
associated with the consideration of the travel- related strategies. The effect of subjective
mobility on the consideration of the travel- related strategies is quite similar to that of objective
mobility.
Relative desired mobility: The negative association of the relative desired mobility variables
with the consideration of the travel- maintaining/ increasing strategies was counter to our initial
expectation ( see Section 2.3): we thought that the more people want to increase their travel, the
more likely they would be to consider strategies that support traveling equal or greater amounts.
Instead, these strategies appear to be more desirable to those who want to decrease their travel,
as a way of making their undesired ( but perhaps necessary) current travel more palatable. On the
other hand, both effects may be at work and cancel each other out in many cases, which may
explain why only a few relative desired mobility variables are significant in this group of models.
By contrast, the effects of the relative desired mobility variables on the consideration of the
travel- reducing and major location/ lifestyle change strategies are bi- directional; that is, they may
positively or negatively affect the consideration. However, the positive coefficients of these
variables indicate competing preferences – the adoption of the strategies in these two bundles
would decrease the amount of commute travel, so as to be able to increase the amount of time
devoted to the desired activity/ travel. Worth noting is that individuals wanting less commuting
are more likely to seek medium- and high- cost adjustments ( telecommuting, residential and
employment relocation in this case) to reduce the commute.
vi
Travel liking: Liking short- distance travel for entertainment and liking long- distance travel
overall increase the probability of considering the travel- maintaining/ increasing strategies. In
general, however, the relative absence of travel liking variables from these models is noteworthy.
In some cases effects in opposite directions may be counteracting each other; in other cases the
effects of travel liking may be captured by related variables that are in the models.
Travel attitudes, personality and lifestyle: The attitude, personality and lifestyle factors that most
commonly, and positively, affect the consideration of the travel- related strategies are pro-environmental
solutions ( attitude), adventure seeker ( personality) and frustrated ( lifestyle).
Individuals advocating environmental protection are more likely than others to consider reducing
their commute and/ or minimizing solo driving to decrease their personal energy consumption
and impacts on the environment. Also, they are more likely to consider getting a fuel efficient
car to decrease their fuel consumption. The adventure seeker factor score has a positive impact
on the consideration of several different strategies in all three conceptual categories. The excess
travel indicator, which captures many of the characteristics of the adventure seeker factor, is
significant and positive for a seventh strategy. The frustrated factor score is significant in five
models. Individuals who are frustrated may view travel- related strategies as potentially one way
to increase their control and/ or life satisfaction.
Mobility constraints: Mobility constraints increase the probability of considering the travel-related
strategies in all three conceptual bundles. It is noteworthy that limitations on driving
during the day and vehicle availability are each significant in four models, and that these two
constraints are more likely to affect the consideration of the workstyle adjustments. This
suggests that a desire to shorten the commute is an important motivation for individuals with
such constraints to consider these travel- related strategies.
Demographics: Age- related variables ( age category and years lived in the U. S.) appear most
commonly in the models. Their generally negative effects indicate that older people are less
likely to consider most of these strategies. In these models, year of personal vehicle is only ( and,
logically, negatively) associated with the auto improvement strategies. Individuals having
dependent care are more inclined to acquire more temporal and/ or spatial flexibility to better
provide the necessary care. Higher personal and household incomes either directly or indirectly
have a positive impact on the consideration of travel- related strategies.
Former adoption of travel- related strategies: Apart from “ Change jobs closer to home”, the
former adoption of each of the remaining 15 individual strategies significantly affects the
consideration of the same strategy, as shown by the shaded cells in Table ES- 1. On one hand,
among the 15 strategies, the former adoption of getting a mobile phone, getting a better car, and
getting a fuel efficient car are negatively associated with their respective reconsiderations,
implying that the former adoption is still in force and the individual is enjoying the utility of such
an adoption. On the other hand, the former adoption of each of the other 12 strategies has a
positive impact on its reconsideration. Either the individual is enjoying and still wants to enjoy
the benefits from the former adoption, or such strategies are attractive again as circumstances
change. Given that these strategies are adopted once, it is natural that they would be adopted
repeatedly over a person’s working life. Whenever time since adoption of a strategy is
significant to the reconsideration of the same strategy ( specifically, for the five strategies C, D, F,
G, and O), it appears with the opposite sign to that of the binary former adoption variable,
vii
meaning reinforcement rather than counteraction of the former adoption variable. In addition,
the effects of three pairs of former adoption variables on the consideration of another strategy
( specifically, the binary adoption and time since adoption of strategies F on C, M on D, and G
on I) follow the same pattern as those of former adoption variables on the consideration of the
same strategy, indicating that the adoption of one strategy is more likely to trigger the
consideration of the other related strategy in the short term. As shown in the off- diagonal blocks
of Table ES- 2 ( Table 33 in the text), when the former adoption of a strategy is significant, its
dominant effect on the consideration of another strategy is positive: the former adoption of
strategy i increases the probability of considering strategy j. Table ES- 3 ( Table 34 in the text)
summarizes the effects of prior adoption, with the strategies grouped according to empirical
similarities ( see Section 2.2.2). It shows that complementary effects are obviously exhibited in
the home- based work bundle. The former adoption of each of the strategies in the alter
employment bundle does not affect the consideration of any other strategies studied here,
suggesting that working part- time and quitting work are likely to be the most radical and
exhaustive changes to cope with congestion. Although not as radical, mode change strategies are
also isolated in their nearly complete lack of influence on the consideration of other strategies
( with the exception, ironically, that changing to driving alone has a negative influence on the
consideration of changing to part- time work). Although the former adoption of changing jobs
closer to home does not significantly affect its reconsideration, it frequently appears with a
positive coefficient in models of the consideration of other strategies; conversely, the former
adoption of “ Move your home closer to work”, which is in the same bundle as the employment
relocation, is only significant in the model of its own reconsideration. This may imply that, in
contrast to a new residential location, some aspects of a new job ( e. g. a higher salary, increased
flexibility) offer individuals an opportunity to seek other kinds of changes, which, of course, may
not only be for transportation reasons.
Overall, the key findings provide evidence in support of most of our initial hypotheses. A more
detailed comparison of some of these hypotheses and results is summarized in Table ES- 4
( Table 35 in the text). Although a few unexpected relationships emerged and there are cases in
which our findings failed to support some hypotheses, the results were generally consistent with
our prior expectations.
In conclusion, the consideration of travel- related strategies is affected not only by the amounts of
travel that individuals actually did, but also by their subjective assessments, desires, and
affinities with respect to travel. This study helps us further understand the influences of these
mobility- related variables on the consideration of each strategy. However, the effects of
objective mobility, subjective mobility, relative desired mobility and travel liking are always
intertwined in individuals’ choice processes, which contributes to the substantial diversity of
their responses. Further, since it is objective mobility that is often the basis of public policy,
these relationships imply that individuals may not respond to public policies designed to adjust
their behaviors in the way that policy makers expected. An individual’s travel attitudes,
personality, and lifestyle play an important role in her consideration of travel- related strategies.
The frequent appearances of these factors further illustrate how different people respond to
congestion, and hence provide helpful information to better understand individuals’ diverse
behaviors. However, it is difficult for policy makers to acquire such information for various
reasons. An individual’s past experience greatly affects her consideration of travel- related
strategies. In the current study, there is evidence that ( 1) the former adoption of a strategy, and
sometimes the time since adoption as well, has an important impact on the consideration of the
viii
same strategy, with a positive association dominating; and ( 2) the adoption of one strategy
sometimes triggers the consideration of another related change in the short term. These findings
suggest that the effectiveness of public policies is impacted by individuals’ past experiences.
Finally, demographic characteristics may affect the response to public policies.
The single key theme that underlies the results of this study is that individuals’ responses to the
travel- related strategies analyzed here – many of them directly tied to public policies intended to
reduce vehicle travel – are influenced by a large variety of qualitative and experiential variables
that are seldom measured and incorporated into demand models. Although there are challenges
associated with that measurement and incorporation, those challenges are not insurmountable.
Devoting further efforts to understanding the role of these attitudinal, personality, lifestyle, and
experience variables will improve our ability to design effective policies and to accurately
forecast the response to policy interventions as well as natural trends.
ix
Table ES- 1. Summary of the Individual Models ( grouped by conceptual bundles)
Travel maintaining/ increasing
( Low cost)
Travel reducing
( Medium cost)
Major location/ lifestyle
change ( high cost)
Dependent Variable
Goodness- of- fit
A. Buy a car stereo system
B. Get a mobile phone
C. Get a better car
D. Get a fuel efficient car
E. Change work trip departure time
F. Hire somebody to do house or yard work
G. Adopt flextime
H. Adopt compressed work week
I. Change from driving alone to other means
K. Buy equipment to help work from home
L. Telecommute
M. Change jobs closer to home
N. Move your home closer to work
O. Work part- instead of full- time
P. Start home- based business
Q. Retire or stop working
N
1172
1263
1118
1155
1265
1238
1278
1278
987
1206
1253
1254
1269
1279
1277
1234
MS ρ 2
0.385
0.124
0.039
0.132
0.332
0.219
0.388
0.476
0.443
0.211
0.265
0.302
0.554
0.327
0.318
0.415
ρ 2
0.450
0.202
0.183
0.229
0.438
0.319
0.477
0.547
0.571
0.381
0.430
0.440
0.628
0.403
0.451
0.510
ρ 2 ( without the constant term)
0.450
0.191
0.167
0.220
0.415
0.298
0.435
0.510
0.565
0.360
0.418
0.440
0.614
0.354
0.446
0.468
Adjusted ρ 2
0.432
0.184
0.158
0.208
0.421
0.304
0.464
0.534
0.543
0.363
0.414
0.427
0.615
0.392
0.434
0.492
Explanatory Variable
Objective Mobility
Frequency of commuting ( SD) +
Frequency of work/ school- related travel ( SD) +
Frequency of grocery shopping travel ( SD) + +
Frequency of entertainment travel ( SD) +
Frequency of travel taking others where they need to go ( SD) +
x
A B C D E F G H I K L M N O P Q
Frequency of other purpose travel ( SD) +
Weekly miles in a bus ( SD) + +
Weekly miles in a train/ BART/ light rail ( SD) -
Total weekly miles ( SD) + +
Weekly miles of commuting ( SD) +
Weekly miles of grocery shopping travel ( SD) -
Weekly miles to eat a meal ( SD) + + +
Weekly miles of entertainment travel ( SD) + +
Weekly miles of travel taking others where they need to go ( SD) -
Commute time +
Commute distance +
Number of trips by personal vehicle ( LD) +
Number of trips by other means ( LD) +
Sum of log of miles for each trip by personal vehicle ( LD) -
Sum of log of miles for each trip by air ( LD) - +
Log total miles by personal vehicle ( LD) +
Subjective Mobility
Commute ( SD) + +
Travel for grocery shopping ( SD) + +
Travel for eating a meal ( SD) -
Travel for entertainment ( SD) + +
Take others where they need to go ( SD) + +
Travel by personal vehicle ( SD) + +
Travel by train/ BART/ light rail ( SD) -
Travel by personal vehicle ( LD) +
Relative Desired Mobility
Overall ( SD) - -
Commute ( SD) - - -
Work/ school- related travel ( SD) - +
Travel for grocery shopping ( SD) -
Travel for eating a meal ( SD) +
Take others where they need to go ( SD) +
Travel by bus ( SD) - +
Travel by walking/ jogging/ bicycling ( SD) +
Travel by air ( LD) + -
Travel Liking
Travel for eating a meal ( SD) +
Travel for entertainment ( SD) + + +
Travel by personal vehicle ( SD) -
Travel by train/ BART/ light rail ( SD) -
xi
A B C D E F G H I K L M N O P Q
Overall ( LD) + +
Work/ school- related travel ( LD) -
Travel for entertainment ( LD) +
Attitudes
Pro- environmental solutions factor score + + + + + + +
Commute benefit factor score -
Travel stress factor score + +
Pro- hi density factor score - -
Personality
Adventure seeker factor score + + + + + +
Loner factor score - -
Calm factor score -
Lifestyle
Frustrated factor score + + + + +
Family & community- oriented factor score + +
Status seeker factor score - +
Workaholic factor score +
Excess Travel
Excess travel indicator +
Mobility Constraints
Limitations on driving during the day + + + +
Limitations on driving on the freeway +
Limitations on flying in an airplane + +
Limitations on riding a bicycle + + +
Percent of time a vehicle is available - - - -
Demographics
North San Francisco -
Time living in the neighborhood +
Age - - +
Years lived in the U. S. - - + - - - - +
Female - +
Number of vehicles in the household - - + +
Year of personal vehicle - - -
Total workers in the household +
Household size -
Anyone in the household needing special care + + + + +
Household with single adult - - +
Household with two or more adults -
Household with two or more adults & children +
Sales occupation -
xii
A B C D E F G H I K L M N O P Q
Demographics
Service/ repair occupation +
Clerical/ administrative support occupation +
Production/ construction/ craft occupation -
Manager/ Administrator occupation +
Professional/ technical occupation +
Full- time worker + + +
Household income category +
Personal income category + +
Vehicle type is pickup +
Vehicle type is small + +
Strategy Adoption
Buy a car stereo system + +
Get a mobile phone - - -
Get a better car - + +
Time since getting a better car +
Get a fuel efficient car -
Time since getting a fuel efficient car +
Change work trip departure time + +
Hire somebody to do house or yard work + + -
Time since hiring domestic help - - - -
Adopt flextime + +
Time since adopting flextime - -
Adopt compressed work week + +
Time since adopting compressed work week +
Change from driving alone to some other means +
Change from another means to driving alone + -
Squared time since changing to driving alone +
Buy equipment to help work from home + + + +
Telecommute +
Change jobs closer to home + + + + + +
Time since changing jobs closer to home -
Move your home closer to work +
Work part- instead of full- time +
Time since working part- time -
Start home- based business + + +
Retire or stop working +
Time since retiring or stopping working +
SD = Short Distance LD = Long Distance
xiii
Table ES- 2. Relationships between Former Adoption and Current Consideration of Strategies ( grouped by conceptual
bundles)
Travel maintaining/ increasing
( Low cost)
Travel reducing
( Medium cost)
Major location/ lifestyle
change ( high cost)
Current Consideration
Former Adoption A B C D E F G H I K L M N O P Q
A. Buy a car stereo system + +
B. Get a mobile phone - - -
C. Get a better car - + +
D. Get a fuel efficient car -
E. Change work trip departure time + +
F. Hire somebody to do house or yard work + + -
G. Adopt flextime + +
J. Change from another means to driving alone + -
H. Adopt compressed work week + +
I. Change from driving alone to some other means +
K. Buy equipment to help work from home + + + +
L. Telecommute +
M. Change jobs closer to home + + + + + +
N. Move your home closer to work +
O. Work part- instead of full- time +
P. Start home- based business + + +
Q. Retire or stop working +
xiv
Table ES- 3. Relationships between Former Adoption and Current Consideration of Strategies ( grouped by factor- based
bundles)
Auto
Improvement
Mobile
Phone
Work
Schedule
Change
Hire
Domestic
Help
Mode
Change
Home-based
Work
Relocation Alter
Employ-ment
Current Consideration
Former Adoption A C D B E G H F I K L P M N O Q
A. Buy a car stereo system + +
C. Get a better car - + +
D. Get a fuel efficient car -
B. Get a mobile phone - - -
E. Change work trip departure time + +
G. Adopt flextime + +
H. Adopt compressed work week + +
F. Hire somebody to do house or yard work + + -
I. Change from driving alone to some other
means +
J. Change from another means to driving alone + -
K. Buy equipment to help work from home + + + +
L. Telecommute +
P. Start home- based business + + +
M. Change jobs closer to home + + + + + +
N. Move your home closer to work +
O. Work part- instead of full- time +
Q. Retire or stop working +
xv
Table ES- 4. Summary of Hypotheses and Results
Variable type General hypotheses Results
Objective
mobility
( 1) The more individuals travel, the more
likely they would be to consider all travel-related
strategies, including the travel-maintaining/
increasing ones.
( 1) Our findings support this hypothesis.
Subjective
mobility
( 1) A higher subjective mobility is positively
associated with the consideration of a wide
range of travel- related strategies.
( 1) Our findings support this hypothesis, similarly to
objective mobility.
Relative
desired
mobility
Individuals having a higher relative desired
mobility are ( 1) more likely to consider
travel- maintaining/ increasing strategies, and
( 2) less likely to consider travel- reducing
and major location/ lifestyle change
strategies.
( 1) Our findings are counter to this hypothesis,
indicating that these strategies are more favored by
those wanting to decrease their travel ( perhaps to
lighten the burden of undesired but necessary travel);
( 2) Our findings provide some support for this
hypothesis. However, competing preferences may
affect the direction of an individual’s consideration
( e. g., those wanting more travel are more inclined to
consider commute- reduction strategies).
Travel liking
The more individuals like travel, ( 1) the
more likely they would be to consider travel-maintaining/
increasing strategies, and ( 2) the
less likely they would be to consider travel-reducing
and major location/ lifestyle
changing strategies.
( 1) Our findings provide some support for this
hypothesis;
( 2) Our findings do not strongly support this
hypothesis, perhaps again due to competing
preferences.
Travel
attitudes
( 1) Individuals with attitudes favoring travel
would be more likely to consider travel-maintaining/
increasing strategies, while
( 2) those with attitudes not favoring travel
would be more likely to consider travel-reducing
and major location/ lifestyle change
strategies.
( 1)( 2) Our findings provide support for these
hypotheses although some travel attitude factors do
not often appear in the models, and others do not
appear at all.
Personality
( 1) The adventure seeker factor is positively
associated with the consideration of most
travel- related strategies.
( 1) Our findings support this hypothesis.
Lifestyle ( 1) The family/ community- oriented factor is
positively associated with the consideration
of travel- reducing and major location/ life
style change strategies;
( 2) Being frustrated is positively related to
considering a wide range of travel- related
strategies;
( 3) A positive score on the workaholic factor
positively affects the consideration of the
strategies beneficial to work;
( 4) Status seekers may be more inclined to
consider strategies involving material
acquisition.
( 1) Our findings provide some support for this
hypothesis;
( 2) Our findings support this hypothesis;
( 3) Our findings fail to support this hypothesis,
except for changing work trip departure time;
( 4) Our findings provide limited support for this
hypothesis.
Excess travel ( 1) Excess travel plays an important role in
the consideration of a wide range of travel-related
strategies.
( 1) Our findings fail to support this. However, the
effects of the excess travel indicator may be captured
by the adventure seeker factor and the mobility
variables.
xvi
( Table ES- 4. Continued)
Variable type General hypotheses Results
Mobility
constraints
( 1) Mobility constraints positively affect the
consideration of a variety of travel- related
strategies
( 1) Our findings support this hypothesis.
Demographic
( 1) Females are more likely to consider the
more costly, travel- reducing and major
location/ lifestyle change strategies;
( 2) Those in upper income categories are
more able and therefore more likely to
consider a wide range of travel- related
strategies.
( 1) Our findings fail to support this hypothesis,
although gender effects may be partly captured by
other variables in the models;
( 2) Our findings offer mixed ( direct and indirect)
support for this hypothesis.
Strategy
adoption
( 1) The former adoption of a strategy could
be either positively or negatively associated
with the consideration of other strategies;
( 2) The former adoption of a strategy
positively affects the consideration of the
same strategy;
( 3) The time since adoption of a strategy is
positively related to its reconsideration.
( 1) Our findings support this hypothesis;
( 2) Our findings generally support this hypothesis
although the effects of three strategies are counter to
it ( for logical reasons) and the effect of one strategy
is not significant;
( 3) Our findings fail to support this hypothesis.
Conversely, we found that the time since adoption of
a strategy appears with the opposite sign to that of its
former adoption.
Cao and Mokhtarian
1
1. INTRODUCTION
It is well known that congestion has become a major problem for urban and suburban residents.
The estimated annual cost of time lost due to congestion in the U. S. was put at $ 48 billion in the
mid- 1990s ( Arnott and Small, 1994). Beyond the loss of time, congestion has serious
consequences for energy consumption and the environment. Governments have been adopting a
wide range of policies to alleviate congestion. During the past two decades, Transportation
Demand Management ( TDM) strategies, such as increasing the cost of operating a private
vehicle, promoting public transit ridership, enhancing accessibility, advocating telecommunica-tion
alternatives and so on, have been a centerpiece of public policy. However, these strategies
have been of limited effectiveness. Most policies are focused on reducing vehicle miles traveled
( VMT) at peak periods, and policy makers assume that individuals will actively respond to these
policies in a manner that minimizes social costs. In reality, however, individuals tend to behave
in a way that minimizes their personal costs ( Salomon and Mokhtarian, 1997). This gap between
the assumptions on which policies are based and the behaviors with which individuals respond to
policy measures greatly affects the effectiveness of such strategies.
The dynamic nature of the individual’s response to congestion further exacerbates the
discrepancy between assumption and reality. A previous empirical study directed by the second
author found that an individual first tends to consider or adopt lower- impact, short- term
strategies ( such as buying a more comfortable car or changing work trip departure time), before
moving to higher- impact and/ or longer- term ones ( such as changing mode, telecommuting, or
relocating). There was also evidence that if dissatisfaction persists or returns an iterative process
is involved in the consideration of some strategies, with cycling back to the same or lower-impact
strategies often occurring ( Raney, Mokhtarian, and Salomon, 2000). Since it is the
higher- impact strategies that are often the focus of public policy, this pattern suggests that
generally individuals do not behave as policy makers expect. Moreover, the personal impacts
and distributional inequities of such strategies may make them less attractive, even criticized.
Therefore, for policy makers and plannners, understanding the determinants of the adoption and
consideration of travel- related strategies may contribute to improved predictions of the
effectiveness of proposed policies, and the design of more effective policies.
This study continues to explore the consideration of 17 specific alternatives. All of them may be
( but are not necessarily) adopted in response to congestion and all of them have travel
implications. It is part of the sequel to the previous study ( Mokhtarian, Raney, and Salomon,
1997; Salomon and Mokhtarian, 1997; Raney, et al., 2000) of a similar set of alternatives placed
in a questionnaire focused on telecommuting attitudes, preferences, and choices. The current
study has adopted several suggestions that the previous study offered for further research ( see
Section 4, Clay and Mokhtarian ( 2002) for details).
The first report in the current series ( Clay and Mokhtarian, 2002) presented a descriptive analysis
of relationships between the adoption or consideration, respectively, of each strategy in turn and
a variety of other variables. The key purpose of this report is to develop behavioral models
( specifically, binary logit models) for the consideration of each strategy and examine any
patterns that emerge across models. Although we collected data on both adoption and
consideration, we use “ consideration” rather than adoption of a strategy as the dependent
variable due to the cross- sectional nature of the available data. As described further in Section 2,
the survey used in this study obtained data on an individual’s past adoption of strategies, current
Cao and Mokhtarian
2
consideration of strategies, mobility- related variables, travel attitudes, personality, lifestyle,
demographics and other variables expected to affect congestion response. However, current
measures of attitudes, mobility, and the other variables are not necessarily appropriate indicators
of past adoption. Using them to estimate the models may either inappropriately reverse the roles
of cause and effect, or provide little explanatory power. Analysis of pairwise associations ( Clay
and Mokhtarian, 2002) confirms that the plausible direction of causality is often ambiguous with
respect to adoption. For consideration, by contrast, it is reasonable to expect current
measurements to help explain the likelihood of current consideration of various strategies.
One specific aspect of the key purpose of this study is to explore the relationship between the
prior adoption of a strategy and its reconsideration. The earlier empirical study suggested that
the previous adoption of some strategies would reduce the probability of considering the
strategies in the same bundle ( Raney, et al., 2000). In the present study, we wish to know
whether the previous adoption of a strategy is more likely to exclude its reconsideration or not,
and how the time since adoption of the strategy affects its reconsideration. Specifically, we
examine the role that previous adoption of a strategy plays in its reconsideration by developing
individual models of the consideration of each strategy, having its adoption and time since
adoption as explanatory variables among others. This exploration will help us to better
understand the dynamic nature of individuals’ behavior in this context.
The organization of this report is as follows. The next section will describe the data and
variables used in this analysis. Section 3 presents and interprets the binary logit models of
consideration of each individual travel- related strategy. Section 4 provides an overview of the
individual models and discusses some general conclusions based on the results.
2. THE DATA AND VARIABLES
2.1 Data
The data analyzed in this study come from a fourteen- page self- administered survey mailed in
May 1998 to 8,000 randomly selected households in three neighborhoods of the San Francisco
Bay Area. Half of the total surveys were sent to an urban neighborhood of North San Francisco
and the other half were divided evenly between the suburban cities of Concord and Pleasant Hill.
These areas were chosen to represent the diverse lifestyles, land use patterns, and mobility
options in the Bay Area. Approximately 2,000 surveys were completed by a randomly selected
adult member of the household and returned, for a 25% response rate. For this study, we chose
to focus on commuting workers since they will contribute most heavily to peak- period
congestion, and are likely to be the most active in the adoption and consideration of travel-related
strategies. The subset of 1,283 cases used in this analysis consists of commuting workers
with relatively complete responses to key questions.
Table 1 summarizes the sample distribution of key characteristics. The sample is relatively
balanced in terms of representation by neighborhood and gender. Higher incomes are
overrepresented compared to Census data.
As background to the variables described below, it should be noted that in the cover letter to the
survey, travel was defined as " moving any distance by any means of transportation – from
walking around the block to flying around the world." In questions relating to the amount of
Cao and Mokhtarian
3
travel conducted or desired by respondents, they were asked ( borrowing wording from the
American Travel Survey) to exclude " travel you do as an operator or crew member on a train,
airplane, truck, bus, or ship."
Most of the variables measured by the questionnaire can be grouped into 11 categories, which
are Objective Mobility, Subjective Mobility, Relative Desired Mobility, Travel Liking, Attitudes,
Personality, Lifestyle, Mobility Constraints, Excess Travel, Demographics and Travel- related
Strategies. The travel- related strategies, which are the focus of this study, are briefly described
in Section 2.2. The other variable categories are the subject of Section 2.3.
Table 1. Demographic Characteristics of Sample Used in This Analysis
Number Percent
Neighborhood Concord ( suburban) 294 22.92% ( n= 1,283)
Pleasant Hill ( suburban) 346 26.97%
North San Francisco ( urban) 643 50.11%
Gender Female 651 50.90% ( n= 1,279)
Male 628 49.10%
Employment status Full- time worker 1,080 84.18% ( n= 1,283)
Part- time worker 203 15.82%
Family status Single 319 24.86% ( n= 1,283)
2+ adults, no children 609 47.47%
1 adult, with children 34 2.65%
2+ adults, with children 321 25.02%
Personal income < $ 15,000 91 7.25% ( n= 1,255)
$ 15,000- 34,999 266 21.20%
$ 35,000- 54,999 386 30.76%
$ 55,000- 74,999 229 18.25%
$ 75,000- 94,999 126 10.04%
> $ 95,000 157 12.50%
Age 18- 23 42 3.27% ( n= 1,283)
24- 40 563 43.88%
41- 64 640 49.88%
> 65 38 2.97%
2.2 The Travel- related Strategies
2.2.1 Strategy descriptions
Figures 1 and 2 below reproduce the two pages of the survey dealing with the travel- related
strategies analyzed in this study. The questions under E1 asked about the adoption, and E2 about
the consideration, of 19 options having travel- related implications. The first column of boxes for
each question in both sections was coded as a binary variable, equal to 0 if the box was checked
( i. e., if the alternative was not adopted or considered), and 1 if one or more reasons for adoption
or consideration were checked. The time since adoption was coded as whole years ( rounded to
the nearest full year, with anything less than 6 months coded as zero).
Cao and Mokhtarian
4
Figure 1. Section E1 ( Adoption) from the Survey
Cao and Mokhtarian
5
Figure 2. Section E2 ( Consideration) from the Survey
Questions “ m” and “ n” had two parts each: “ Change jobs . . . closer to home” and “. . . farther
from home” ( referred to as “ m1” and “ m2,” respectively), and “ Move your home . . . closer to
work” and “. . . farther from work” (“ n1” and “ n2”). The format for these two questions, shown
in Figures 1 and 2, was designed to economize on vertical space. Unfortunately, it had the
Cao and Mokhtarian
6
unanticipated effect of confusing many respondents ( apparently leading them to think that they
needed to respond to only one member of each pair) and resulted in a disproportionately high
number of non- responses, particularly on the second half of each question. Given that 28.8% of
the sample was missing at least one of the four responses to the adoption and consideration
questions for the m2 and n2 alternatives, these variables were not used to screen out cases with
missing data, nor did we attempt to fill any missing data for them.
In previous analyses of these data, cases with missing responses on a number of key variables
were either removed or filled; this resulted in 1,904 cases containing relatively complete data for
variables other than the travel- related strategies. For this study, any case missing more than two
out of the 17 responses for either the adoption or consideration of the travel- related strategies
was removed, and stochastic data filling was used for the remaining missing responses ( see
Section 3, Clay and Mokhtarian ( 2002) for details). Finally, consistent with the focus of
previous analyses of these data on commuting workers ( in view of the observation that they tend
to have different travel patterns and attitudes than non- commuters or non- workers), cases were
removed if the respondent did not report working part- or full- time and commuting to work at
least once a month. This reduced the final usable dataset for this analysis to 1,283 cases.
2.2.2 Identification of bundles of strategies
To better understand how these travel- related strategies interact with travel attitudes,
demographics and other variables in our analysis, it is useful to group them into bundles based
on both conceptual and empirical similarities.
Similar to Mokhtarian, et al. ( 1997), two methods were used to develop bundles of travel- related
strategies, with the results shown in Table 2. First, variables were grouped conceptually into
three bundles based on the generalized cost ( including time, stress, and other impacts as well as
monetary cost) and the amount of lifestyle change associated with each travel alternative. Group
one includes low cost, travel- maintaining/ increasing strategies such as getting a more
comfortable car or purchasing a mobile phone. Group two includes more costly, travel- reducing
alternatives such as adopting a compressed workweek or telecommuting. The third group
consists of major location/ lifestyle changes such as quitting work, working part- time instead of
full- time, and moving home or work closer to each other.
In the second method, factor analysis of the responses was performed to identify bundle
groupings. Factor analysis identifies patterns of common variation among a group of variables
( the binary adoption and consideration variables, in this case), and as such groups our
alternatives based on the empirical affinities in responses to them. The bundles developed in this
analysis are a composite of the results of 36 different factor analyses. Factor analysis was
conducted for 3, 4, 5, and 6 factor solutions across the following groups: adoption for the entire
sample, adoption for commuters and full- time workers only, adoption for the entire sample
( excluding m2 and n2), adoption for commuters and full- time workers ( excluding m2 and n2),
consideration for the full sample, consideration for commuters and full- time workers only,
consideration for the full sample ( excluding m2 and n2), consideration for commuters and full-time
workers ( excluding m2 and n2), and combined adoption and consideration for the entire
sample ( excluding m2 and n2). The factor- based bundles that appear in Table 2 were the
groupings that most commonly appeared across all 36 factor analyses and conceptually made the
most sense.
Cao and Mokhtarian
7
Table 2. Conceptual and Factor- based Bundles of the Travel- related Strategies
Conceptual Bundle Groupings
Group 1. Travel maintaining/ increasing a. Buy a car stereo system
b. Get a mobile phone
c. Get a better car
d. Get a more fuel efficient car
e. Change work trip departure time
f. Hire someone to do house or yard work
g. Adopt flextime
j. Change from another means of getting to work to driving
alone
Group 2. Travel reducing h. Adopt compressed work week ( such as a “ 9/ 80” schedule)
i. Change from driving alone to work to some other means
k. Buy equipment/ services to help you work from home
l. Telecommute ( part- or full- time)
Group 3. Major location/ lifestyle change m. Change jobs closer to home
n. Move your home closer to work
o. Work part- time instead of full- time
p. Start home- based business or put more effort into an
existing one
q. Retire or stop working
Factor- based Bundles
Group 1. Auto improvement a. Buy a car stereo system
c. Get a better car
d. Get a more fuel efficient car
Group 2. Mobile phone b. Get a mobile phone
Group 3. Work- schedule changes e. Change work trip departure time
g. Adopt flextime
h. Adopt compressed work week ( such as a “ 9/ 80” schedule)
Group 4. Hire someone to do house or
yard work
f. Hire someone to do house or yard work
Group 5. Mode change i. Change from driving alone to work to some other means
j. Change from another means of getting to work to driving
alone
Group 6. Home- based work k. Buy equipment/ services to help you work from home
l. Telecommute ( part- or full- time)
p. Start home- based business or put more effort into an
existing one
Group 7. Residential/ employment
relocation
m. Change jobs closer to home
n. Move your home closer to work
Group 8. Alter employment status o. Work part- time instead of full- time
q. Retire or stop working
Eight bundles were identified from this process. Note that bundles two and four consist of only
one strategy each. In the previous related study ( Mokhtarian, et al., 1997), the strategy “ Get a
mobile phone” was grouped with the auto improvement bundle. For this analysis it is kept
separate based on factor loadings and the conceptual argument that mobile phones represent a
unique strategy in comparison to the purely auto- oriented solutions ( get a better car, get a more
fuel efficient car, and buy a car stereo system).
Cao and Mokhtarian
8
Bundle four, “ Hire someone to do house or yard work,” emerged as an independent factor in the
earlier study, and remains independent in this analysis for lack of conceptual ( or strong empirical)
linkage with other bundles in the study.
In a previous portion of this study ( Clay and Mokhtarian, 2002), descriptive analyses were
conducted for conceptual and factor- based bundles. Specifically, other variables in the dataset
were related not only to the adoption and consideration of individual strategies, but also to the
adoption and consideration of bundles of strategies. A bundle adoption variable was defined as 1
if any strategy in the bundle had been adopted, and 0 otherwise; bundle consideration was
defined similarly. In this study, although we build models for the consideration of individual
strategies, to organize the exposition we group the strategies into the factor- based bundles and
look for common patterns within each bundle. A parallel analysis ( Choo and Mokhtarian, 2003)
is estimating models of consideration of a strategy bundle rather than an individual strategy.
2.3 Explanatory Variables
Aside from the strategy adoption variables, the remaining explanatory variables fall into the ten
categories mentioned in Section 2.2. In this section we briefly describe each of those categories,
together with some hypotheses about their relationships to the consideration of our travel- related
strategies.
The three mobility categories and the travel liking category had similar structures. In each case,
measures were obtained both overall and separately by purpose and mode, for short- distance and
long- distance travel. Consistent with the American Travel Survey, long- distance trips were
defined as those longer than 100 miles, one way. The short- distance modes measured were:
personal vehicle, bus, Bay Area Rapid Transit ( heavy rail)/ light rail/ train, walking/ jogging/
bicycling, and other. The short- distance purposes measured were: commuting to work or school,
work/ school- related, grocery shopping, eating a meal, travel for entertainment, and taking other
people where they need to go. Long- distance measures were obtained for the personal vehicle
and airplane modes, and for the work/ school- related and entertainment/ social/ recreational
purposes.
Objective Mobility
These questions asked about distance and frequency of travel by mode and trip purpose, as well
as travel time for the commute trip. For short- distance trips, respondents were asked how often
they traveled for each purpose, with six categorical responses ranging from “ never” to “ 5 or
more times a week”. Frequency of trips by mode was not obtained ( a conscious design choice,
to reduce the burden on the respondent). Respondents were also asked to specify how many
miles they traveled each week, in total and by mode and purpose.
On one hand, reported estimations of typical travel, such as we obtained here, are not as reliable
as travel diary data. On the other hand, travel diaries can be criticized for generally
encompassing only a few days of travel and therefore potentially being unrepresentative at the
disaggregate level. Of course, these measures are respondents’ reports of the distance, frequency,
and time they are traveling, and hence are “ objective” only in the sense of referring to those
externally measurable quantities ( in contrast to the subjective measures of subjective and relative
Cao and Mokhtarian
9
desired mobility described below), rather than in the sense of actually being measured through
external observation.
For long- distance trips, pre- testing indicated that respondents would not be able to estimate
distances reliably. Thus, respondents were simply asked to tabulate how many trips they made
“ last year” for each mode- purpose combination ( personal vehicle/ work, personal vehicle/ enter-tainment,
etc.) to each of nine regions of the world. Those responses indicated number of trips
directly, and were also transformed to approximate measures of distance, through judgmental
average distances developed between the Bay Area and each of the nine world areas.
Two transformations of the long- distance objective mobility indicators are utilized in this report:
the natural log of the total miles plus one (“ log of miles”), and the summation of the natural log
of miles plus one ( to avoid evaluating the log of 0, which is negative infinity) for each
purpose/ mode combination (“ sum of log- miles”). The reason for performing a natural log
transformation was to reduce the weight of long trips, under the assumption that each additional
mile traveled would have a diminishing marginal impact ( i. e., each additional mile does not have
as strong an incremental effect as the previous mile). As shown by the example in Section 4.1.1
of Curry ( 2000), the two transformations differ in that sum of log- miles gives more weight to a
larger number of trips but traveling the same amount of miles, compared to log of miles.
The travel- related strategies discussed in this study represent some possible ways to cope with
congestion and a higher amount of travel. Thus, we would certainly expect a higher objective
mobility to be positively associated with the consideration of the travel- reducing and major
location/ lifestyle strategies of Table 2. The situation with respect to the travel- maintaining/
increasing strategies is not as clear. On one hand, it is possible that individuals with a higher
objective mobility want to cut their travel, and hence are less likely to consider an adjustment
that would maintain or increase their travel. However, the descriptive analyses in Clay and
Mokhtarian ( 2002) confirmed that those who actually did a lot of travel were more inclined to
consider even the travel- maintaining/ increasing strategies ( as well as the others), apparently in
order to make the travel they must do less costly and/ or more productive. Therefore, for the
models developed here, it is hypothesized that the more individuals travel, the more likely they
would be to consider all these strategies.
Subjective Mobility
We are interested not only in the objective amount an individual travels, but also in how that
amount of travel is perceived. One person may consider 100 miles a week to be a lot, while
another considers it minimal. For each of the same categories as for objective mobility ( overall,
purpose, and mode categories for short- and long- distance), respondents were asked to rate the
amount of their travel on a five- point semantic- differential scale anchored by “ none” and “ a lot”.
Similarly to objective mobility, we hypothesize that individuals perceiving that they do a lot of
travel will be more inclined to consider travel- reducing and major location/ lifestyle change
strategies. With respect to travel- maintaining/ increasing strategies, individuals with high
subjective mobility may either be less inclined to consider them because they do not want to
maintain or increase travel, or more inclined to consider them in order to make the extensive
travel they must do more comfortable or productive. Again, the findings in Clay and Mokhtarian
Cao and Mokhtarian
10
( 2002) support the latter expectation. Thus, we hypothesize that a higher subjective mobility is
positively related to considering a wide range of strategies.
Relative Desired Mobility
An individual may consider that she travels “ a lot”, but want to do even more. Thus relative
desired mobility refers to how much a person wants to travel compared to what she is doing now.
The structure of this question mirrors the structure for subjective mobility, with respondents
rating the amount of travel they want to do ( in each category) compared to the present, on a five-point
scale from “ much less” to “ much more”.
Individuals having a higher relative desired mobility want to increase their travel, thus they are
expected to be more likely to consider travel- maintaining/ increasing strategies and less likely to
consider travel- reducing and major location/ lifestyle change strategies. However, seemingly
counterintuitive results may occur for some individual strategies; for example, those desiring
more entertainment travel may be more likely to consider some commute- reduction strategies in
order to obtain more time for the desired activities.
Travel Liking
Whether a respondent who already travels a lot wants to reduce it or do even more is likely to
depend on how much she enjoys traveling. To directly measure the affinity for travel, the question
was asked, " How do you feel about traveling in each of the following categories? We are not
asking about the activity at the destination, but about the travel required to get there." Respondents
were then asked to rate each of the same categories as subjective mobility on a five- point scale
from “ strongly dislike” to “ strongly like”.
Despite our attempt to alert respondents to distinguish the destination activity from the travel, it
is likely that even many of those who actually read the instructions ( and more of those who did
not) were unsuccessful at doing so. Future studies should perhaps make this distinction even
more forcefully to the respondent; interactive interviews would be one mechanism for probing
answers and helping the participant to separate these components of the utility for travel.
Nevertheless, we believe that the responses to this question are essentially measuring the degree
of the respondent’s affinity for travel for its own sake, even if that measurement is imperfect.
Similar to relative desired mobility, a higher rating for travel liking indicates a positive utility of
travel. It is hypothesized that the more an individual likes travel, the more likely she would be to
consider travel- maintaining/ increasing strategies, and the less likely she would be to consider
travel- reducing and major location/ lifestyle changing strategies.
Attitudes
The survey contained 32 attitudinal statements related to travel, land use, and the environment, to
which individuals responded on the five- point Likert- type scale from “ strongly disagree” to
“ strongly agree”. Factor analysis was then used to extract the relatively uncorrelated
fundamental dimensions spanned by these 32 variables. Six underlying dimensions were
identified, using principal axis factoring with oblique rotation ( see Redmond, 2000 for details):
Cao and Mokhtarian
11
travel dislike, pro- environmental solutions, commute benefit, travel freedom, travel stress, and
pro- high density.
The different travel attitude factors we have measured can affect the consideration of each travel-related
strategy differently. Generally, a positive commute benefit or travel freedom factor score
indicates a utility of travel or lack of constraints on individuals’ travel, respectively, so they are
expected to be negatively associated with the consideration of strategies that reduce travel.
Conversely, a positive score on the other factors indicates some kind of disutility of travel or
anti- travel attitude, thus they are hypothesized to be positively associated with the consideration
of travel- reducing and major location/ lifestyle change strategies.
Personality
Respondents were asked to indicate how well ( on a five- point scale from “ hardly at all” to
“ almost completely”) each of 17 words and phrases described their personality. Each of these
traits was hypothesized to relate in some way to one’s orientation toward travel, or to reasons for
wanting to travel for its own sake. These 17 attributes reduced to four personality factors:
adventure- seeker, organizer, loner, and the calm personality.
Redmond ( 2001) hypothesized that those with a positive score on the adventure seeker factor
enjoy travel for entertainment more than for work, so they may be more likely to change their
commuting patterns. Clay and Mokhtarian ( 2002) found a strong positive association of this
factor with a number of different strategies, suggesting that to some extent adventure seekers
may value change or variety for its own sake. Thus, this factor is expected to be positively
associated with the consideration of most travel- related strategies. The impacts of the other
personality factors are less predictable, but we include them to explore the role they may play in
the consideration of travel- related strategies.
Lifestyle
The survey contained 18 Likert- type scale statements relating to work, family, money, status,
and the value of time. These 18 questions comprised four lifestyle factors: status seeker,
workaholic, family/ community- oriented and a frustrated factor.
The family/ community oriented factor score is expected to be positively associated with the
consideration of travel- reducing and major location/ lifestyle change strategies since these
strategies could save time for family and community activities. Being frustrated may be
positively related to considering a wide range of strategies because such people may believe that
a change would bring them greater satisfaction or control. A positive score on the workaholic
factor is expected to positively affect the consideration of the strategies beneficial to work, such
as telecommuting. Status seekers may be more inclined to consider strategies involving material
acquisition, such as getting a better car or a mobile phone.
Excess Travel
Thirteen statements asked how often ( on a three- point scale: “ never/ seldom”= 0, “ sometimes”= 1,
“ often”= 2) the respondent engaged in various activities that would be considered unnecessary or
excess travel. The excess travel indicator is the sum of the responses to these statements, ranging
Cao and Mokhtarian
12
from 0 for the respondent who never/ seldom did any of them to 26 for the respondent who often
did all of them. This variable can be considered an indicator of objective mobility, but also has a
psychological flavor indicating an enjoyment of travel beyond the purely utilitarian. The index
may represent a strong desire for travel generally, or a preference for discretionary travel which
may have a negative relationship with mandatory travel for such purposes as commuting and
taking others where they need to go. It is hypothesized that the index would play an important
role in the consideration of a wide range of strategies.
Mobility Constraints
In our study, mobility constraints are physical or psychological limits on travel. These
constraints may affect the amount an individual travels or her enjoyment of that travel. In our
survey, these constraints are measured by questions concerning limitations on traveling by
certain modes or at certain times of day ( with ordinal response categories “ no limitation”, “ limits
how often or how long”, and “ absolutely prevents”), and the availability of an automobile when
desired ( an oppositely- oriented measurement of mobility constraints). Mobility constraints are
expected to positively affect the consideration of a variety of strategies.
Demographics
Finally, the survey included an extensive list of demographic variables to allow for comparison
to other surveys and to Census data. These variables include neighborhood and vehicle type
dummies, gender, age, years in the U. S., education and employment information, household
information such as number of people in the household, family status, and personal and
household income.
Based on previous findings ( Mokhtarian, et al., 1997; Clay and Mokhtarian, 2002), females are
expected to be more inclined to consider the more costly, travel- reducing and major
location/ lifestyle change strategies; it is hypothesized that personal and household incomes
would be positively associated with consideration of a variety of strategies.
3. MODELS OF CONSIDERATION OF EACH STRATEGY
3.1 General Specification and Interpretation Issues
As illustrated in Figures 1 and 2, when a strategy was adopted or considered, the respondents
were asked to “ check all [ the reasons] that apply”. The last five columns of Sections E1 and E2
provide five reasons for making such a decision, of which only one is directly travel- related
(“ Reducing or easing travel”). Therefore, although each of these strategies has potential travel
consequences, all of them could be adopted or considered for reasons unrelated to transportation
concerns. For example, there are many possible reasons for taking a new job that just happens to
reduce the commute. Clay and Mokhtarian ( 2002) analyzed the resulting responses and found
that changing from driving alone to some other means is the only strategy for which “ Reducing
or easing travel” is the most commonly cited reason. However, it should be noted that while we
deliberately avoided a response bias in favor of the travel reason by placing it fourth ( just before
“ other”) in the set of five reasons, there is in fact a response bias in the opposite direction.
Although respondents were invited to check as many reasons as applied, many would have
stopped after checking the first relevant reason. Even when they were willing to check multiple
Cao and Mokhtarian
13
reasons, they may not always have realized the importance of transportation to their choices. For
example, a respondent could have selected “ family related” recalling that the alternative was
adopted to allow more time with family, but not immediately recognizing that the additional time
with family was obtained by reducing the amount of time spent driving. This logic holds true for
many of the reasons selected, given that the list of travel- related alternatives was designed to
comprise mostly strategies that could ease or reduce the impact of driving. Thus, the role of
transportation in these choices is most likely understated. In any case, whether adopted or
considered for transportation reasons or not, it is still worth studying behavior with respect to
these strategies since many of them are promoted as transportation policy, and all of them do
have travel impacts.
Binary logit models were developed for the consideration of each individual travel- related
strategy. Each dependent variable was defined as 1 if the strategy was considered by a
respondent and as 0 if not. The explanatory variables were selected from the strategy adoption
variables presented in Section 2.2.1 and the variables described in Section 2.3. In all, more than
180 variables were considered to potentially have some explanatory power in the models. The
logistic regression function of SPSS was used to estimate the models, due to its automated
specification refinement capabilities. In particular, the forward likelihood ratio method was
adopted to refine the initial experimental specifications in which all explanatory variables were
allowed to enter. Generally, the significance level 0.05 was used to incorporate or release
variables in the final “ best” model. However, a few marginally significant variables were kept in
the final models when they provided some interesting or insightful information. Conversely, in
several cases, explanatory variables had to be excluded from the specification due to their
appearance with counterintuitive signs.
A critical survey design feature affected model development for several of the travel- related
strategies. As explained in Section 2.2.1, respondents were asked in Question 1 of Part E of the
survey to indicate which alternatives they had adopted, and in Question 2 ( after stating “ even if
you have already made some of these choices, you could be thinking about making a similar
change again, or considering new options”) to indicate which they were seriously considering.
For reference later on, Table 3 presents the corresponding distribution of former adoption and
current consideration of each strategy. This design leaves two serious ambiguities. First, some
respondents who had adopted, and were still engaged in, a particular strategy ( such as
telecommuting), may have felt uncomfortable indicating they “ were not seriously considering”
something they were in fact actually doing. On the other hand, we have no way of ascertaining
whether a strategy such as telecommuting, once adopted, remained in place or not – it is well-established
that many telecommuting engagements are temporary ( Varma, et al., 1998).
The result of these two situations is that when someone who has previously adopted a certain
strategy indicates she is currently considering it, we are not certain whether she is actually
currently doing it, or whether she has previously discontinued the strategy and is now
considering it again. Naturally these two groups of people could be quite different in terms of
explanatory variables. Similarly, “ non- considerers” who have previously adopted a strategy
comprise at least two distinct groups: those who are not considering it because their prior
adoption is still in force, and those who are not considering it because they have previously
discontinued the strategy and do not wish to re- visit it at this time.
Cao and Mokhtarian
14
Table 3. The Distribution of Former Adoption and Current Consideration of Strategies
Consideration
No Yes
Adoption Adoption
Strategy No Yes No Yes Sample size
a. Buy a car stereo system 590 ( 46.0%) 505 ( 39.4%) 72 ( 5.6%) 116 ( 9.0%) 1,283 ( 100%)
b. Get a mobile phone 492 ( 38.4%) 411 ( 32.0%) 263 ( 20.5%) 117 ( 9.1%) 1,283 ( 100%)
c. Get a better car 252 ( 19.7%) 552 ( 43.0%) 181 ( 14.1%) 298 ( 23.2%) 1,283 ( 100%)
d. Get a fuel efficient car 558 ( 43.5%) 360 ( 28.0%) 210 ( 16.4%) 155 ( 12.1%) 1,283 ( 100%)
e. Change work trip departure time 704 ( 54.9%) 353 ( 27.5%) 88 ( 6.9%) 138 ( 10.7%) 1,283 ( 100%)
f. Hire someone to do house or yard work 753 ( 58.7%) 233 ( 18.2%) 138 ( 10.7%) 159 ( 12.4%) 1,283 ( 100%)
g. Adopt flextime 909 ( 70.9%) 181 ( 14.1%) 99 ( 7.7%) 94 ( 7.3%) 1,283 ( 100%)
h. Adopt compressed work week ( such as a “ 9/ 80” schedule) 1,041 ( 81.1%) 90 ( 7.0%) 110 ( 8.6%) 42 ( 3.3%) 1,283 ( 100%)
i. Change from driving alone to work, to some other means 955 ( 74.4%) 183 ( 14.3%) 93 ( 7.2%) 52 ( 4.1%) 1,283 ( 100%)
j. Change from another means of getting to work, to driving alone 1,081 ( 84.2%) 142 ( 11.1%) 42 ( 3.3%) 18 ( 1.4%) 1,283 ( 100%)
k. Buy equipment/ services to help you work from home 776 ( 60.5%) 202 ( 15.7%) 122 ( 9.5%) 183 ( 14.3%) 1,283 ( 100%)
l. Telecommute ( part- or full- time) 931 ( 72.6%) 88 ( 6.9%) 148 ( 11.5%) 116 ( 9.0%) 1,283 ( 100%)
m. Change jobs closer to home 772 ( 60.2%) 268 ( 20.9%) 174 ( 13.5%) 69 ( 5.4%) 1,283 ( 100%)
m2. Change jobs farther from home 775 ( 79.9%) 144 ( 14.9%) 37 ( 3.8%) 14 ( 1.4%) 970 ( 100%)
n. Move your home closer to work 1,015 ( 79.1%) 149 ( 11.6%) 91 ( 7.1%) 28 ( 2.2%) 1,283 ( 100%)
n2. Move your home farther from work 930 ( 88.8%) 77 ( 7.4%) 31 ( 3.0%) 9 ( 0.8%) 1,047 ( 100%)
o. Work part- time instead of full- time 918 ( 71.6%) 139 ( 10.8%) 145 ( 11.3%) 81 ( 6.3%) 1,283 ( 100%)
p. Start home- based business or put more effort into an existing
one
990 ( 77.2%) 62 ( 4.8%) 148 ( 11.5%) 83 ( 6.5%) 1,283 ( 100%)
q. Retire or stop working 1,081 ( 84.3%) 23 ( 1.8%) 166 ( 12.9%) 13 ( 1.0%) 1,283 ( 100%)
Cao and Mokhtarian
15
In some cases, the nature of the strategy is such that, once it is adopted, it remains in force until it
is re- adopted, so to speak. For example, “ Changing work trip departure time” cannot be
unadopted without effectively re- adopting it. For those strategies, the ambiguities described
above are not a problem. For each of the remaining strategies, however, we chose to estimate
two models: one on the full dataset, and one on non- adopters only. Specifically, for the
following seven strategies, we developed models based on only non- adopter data, as well as on
the full dataset: “ Hire somebody to do house or yard work”, “ Adopt flextime”, “ Adopt
compressed work week”, “ Telecommuting ( part- or full- time)”, “ Work part- instead of full- time”,
“ Start home- based business or put more effort into an existing one”, and “ Retire or stop
working”. However, estimating the models with only non- adopters is not an ideal solution either.
Analyzing only the non- adopter models would be unsatisfactory since we wish to understand the
behavior of adopters as well as non- adopters ( particularly since adopters can comprise up to
30.6% of the sample for these strategies). However, we believe that a comparison of the full-data
and non- adopter models will be fruitful, with both the similarities and the differences
between them being instructive. In such a comparison, it should be kept in mind that, for the
non- adopter models ( unlike the full- sample models), adoption, time since adoption of the given
strategy and its quadratic term must of necessity be excluded as potential explanatory variables.
It is appropriate to comment in general on the inclusion of the adoption and time since adoption
variables in the models estimated on the full sample. As mentioned in Section 1, some evidence
suggests that individuals first tend to consider or adopt lower- impact strategies, moving to
higher- impact ones if dissatisfaction still persists or returns, and there is a weaker tendency for
them to cycle back to lower- impact strategies if dissatisfaction recurs after they have adopted a
higher- impact one. On the other hand, if the adoption of a strategy has met individuals’ needs,
its adoption may decrease the probability of considering the other strategies. Therefore, the
former adoption of a strategy could be either positively or negatively associated with the
consideration of other strategies. By contrast, for most of the strategies we are studying, we
expect that the former adoption of a strategy positively affects the consideration of the same
strategy. Either the individual is enjoying and still wants to enjoy the benefits from the previous
adoption, or such strategies are attractive again as circumstances change. Given that they are
adopted once, it is natural to expect them to be adopted repeatedly over a person’s working life.
Moreover, we initially expected the time since adoption of a strategy to be positively related to
its reconsideration. That is, the longer ago an individual adopts a strategy, the more likely she is
to consider the same strategy. The time since adoption variable posed a difficulty with respect to
the treatment of non- adopters, however. Non- adopters had to be given a value for this variable
in order for them ( and this variable) to be included in the full- sample models. The standard
practice of setting a variable to zero for cases for which it was not applicable was unsatisfying in
this situation, however. Setting time since adoption to zero for non- adopters lumped non-adopters
together with very recent adopters ( having nearly zero time since adoption), whereas in
reality, one might expect those two groups to be quite different ( perhaps even opposite) in their
propensity to consider the same alternative ( with non- adopters far more likely to consider a
strategy than recent adopters).
To reflect the expectation that consideration of a strategy would generally increase with time
since adoption, with non- adopters being most likely of all to consider it, we experimented with a
“ synthetic” time since adoption variable for each strategy. For non- adopters we set time since
adoption of that strategy equal to the longest time since adoption of that strategy found in the
Cao and Mokhtarian
16
sample, plus an arbitrary inflation factor of 20%. That is, for all non- adopters, time since
adoption of a given strategy was defined to be 1.2 times the longest time since adoption in the
sample. But the models containing these synthetic variables were unsatisfactory – difficult to
interpret and producing coefficients with counterintuitive signs. In retrospect, our hypothesis
that the propensity of non- adopters to consider a strategy would be similar to that of a long- ago
adopter was probably too simplistic: in many cases individuals may not have adopted a strategy
precisely because of a disinclination toward it that still persists and makes them unwilling to
consider it.
Ultimately then, we abandoned the synthetic time since adoption variable, and returned to the
original variable that was defined as zero for non- adopters. We interpret this variable as the
interaction or product of the binary adoption variable and time since adoption, and hence as
representing the impact of time since adoption for adopters. We also included a squared time
since adoption variable for each strategy, to allow for non- linear effects. One could imagine that
the propensity to consider a strategy might be highest for intermediate times since adoption:
recent adopters of course may be less likely to consider it again, but also, an adoption long ago
and not more recently may signify rejection of the strategy for whatever reasons ( it was not
deemed effective, it is no longer deemed appropriate or desirable or available), and hence a
lower propensity to consider it.
As a final comment with respect to specification and interpretation, the richness of our set of
variables makes it unrealistic to assume them to be totally independent of each other. Depending
on the context of a specific model, a certain explanatory variable entering the model may be not
only representing itself but also acting as a proxy for other variable( s). For example, the number
of vehicles in the household may be an indicator of a mobility constraint and/ or income
characteristics. This potential multi- faceted nature of the variables made the interpretation
process particularly interesting.
The following sections group the 17 individual strategies according to the eight factor- based
bundles described in Section 2.2.2. In each section, we first present a table summarizing the
directions of effects in each individual model comprising the bundle, and make some
observations on the bundle as a whole. We then discuss the individual models in brief,
accompanied by tables presenting the actual coefficient estimates and other information. In
addition, for the seven strategies mentioned above, the models with only non- adopters will be
presented, following the models with all respondents.
3.2 Auto Improvement
The auto improvement bundle contains three low- cost, short- term, and travel- maintaining
adjustments: “ Buy a car stereo system”, “ Get a better car” and “ Get a fuel efficient car”.
Although individuals cannot reduce their actual amount of travel by adopting these strategies,
they can obtain a more comfortable, satisfying and functional travel environment when frustrated
by congestion ( Salomon and Mokhtarian, 1997). Thus, the likely consequence of these strategies
is to reduce the costs of travel without reducing objective mobility. Table 3 shows that auto
improvement strategies are extensively accepted by commuters. In this sample, about 82% of
respondents have ever adopted one or more of the auto improvement strategies, and about 48%
of respondents have been considering one or more strategies in this bundle. Both adoption and
consideration of strategies in this bundle rank first among all the bundles discussed in this study.
Cao and Mokhtarian
17
An intrinsic enjoyment of travel itself and the benefits of commuting may partly explain this
popularity. First, travel is not absolutely a derived demand, but sometimes it is desired for its
own sake ( Mokhtarian and Salomon, 2001). Individuals wanting to travel for its own sake are
likely to do a lot of traveling, and therefore likely to want to make that travel even more
enjoyable by adopting one or more of these strategies. Thus, as pointed out by Clay and
Mokhtarian ( 2002), these strategies may be both an effect and a cause of a positive utility for
travel. Second, commuting is not entirely a waste of time; people may attribute some benefits to
commuting. For example, commute time may act as a transition between home and work roles,
individuals could conveniently link other errands to the commute trip, and so on. Several studies
suggest that people want to engage in some commuting, and that not everyone wants to reduce
their current commute time ( e. g., Redmond and Mokhtarian, 2001). Therefore, the utility of
commuting specifically may further enhance the desirability of these auto improvement
strategies. On the other hand, results from Clay and Mokhtarian ( 2002) suggest that these
strategies may also be adopted to ameliorate the burden of travel that must be done. Thus, since
both likers of travel and dislikers of travel are motivated ( for different reasons) to adopt these
strategies, their popularity is not surprising.
As shown in Table 4, one variable, year of personal vehicle, is common to all three strategies in
this bundle. The higher the model year, of course the newer the vehicle. The possession of a
newer personal vehicle suggests that individuals may have already created a relatively
comfortable travel environment, and thus will have little motivation to consider further auto
improvements. Therefore, it is reasonable that year of personal vehicle is negatively associated
with the consideration of each of the auto improvement strategies.
Except for year of personal vehicle, the model of consideration of buying a car stereo system
does not share any variable with either that of getting a better car or that of getting a fuel
efficient car, while there are seven variables that are common to the latter two models. This
suggests that the processes of consideration of getting a better car and of getting a fuel efficient
car show substantial commonality. However, it should be kept in mind that the definition of a
better car was intentionally left ambiguous, with respondents answering according to their own
ideas. Different people could have different criteria for classifying a car as “ better”, such as
being larger, more luxurious or even more fuel efficient, but at a minimum, it is better in some
function( s) than their current one. Generally, when individuals consider getting a better/ fuel-efficient
car, the car could be either a brand new one, or a used one but better ( e. g. newer) than
their current one. Therefore, for either of these two strategies, the perceived benefits could be
mainly focused on providing a more comfortable, reliable means of travel, reducing the out- of-pocket
costs of travel, and/ or improving the environment ( since newer, more fuel- efficient cars
also generally pollute less).
Individuals liking short- distance travel for entertainment are more likely to get a better/ fuel-efficient
car. A higher liking of short- distance travel for entertainment is positively associated
with doing more, or with the desire to do more, of such travel. Thus, such individuals not only
like travel for entertainment, but do a fair amount of it already, and also want to increase it.
Therefore it is plausible that they are more likely to consider these two strategies, either for
obtaining a more comfortable travel environment or for reducing the costs of such travel, or both.
Especially for those who are entertainment- oriented, the vehicle itself may constitute a form of
entertainment. It is also noteworthy that a related variable, the actual frequency of short- distance
entertainment travel, is positively significant to the consideration of buying a car stereo system.
Cao and Mokhtarian
18
Table 4. Models of Consideration of Auto Improvement Strategies ( Bundle 1)
Buy a car
stereo system
Get a better
car
Get a fuel
efficient car
N 1172 1118 1155
MS ρ 2 0.385 0.039 0.132
ρ 2 0.450 0.183 0.229
Adjusted ρ 2 0.432 0.158 0.208
Variable
Objective Mobility
Frequency of entertainment travel ( SD) +
Frequency of other purpose travel ( SD) +
Weekly miles in a train/ BART/ light rail ( SD) -
Total weekly miles ( SD) +
Weekly miles to eat a meal ( SD) +
Sum of log of miles for each trip by air ( LD) -
Subjective Mobility
Travel for grocery shopping ( SD) +
Take others where they need to go ( SD) +
Travel by train/ BART/ light rail ( SD) -
Travel by personal vehicle ( LD) +
Relative Desired Mobility
Overall ( SD) -
Travel by bus ( SD) -
Travel Liking
Travel for entertainment ( SD) + +
Overall ( LD) + +
Attitudes
Pro- environmental solutions factor score +
Pro- high density factor score - -
Personality
Adventure seeker factor score +
Lifestyle
Frustrated factor score + +
Status seeker factor score -
Mobility Constraints
Limitations on driving during the day +
Demographics
Age -
Female -
Year of personal vehicle - - -
Years lived in the U. S. -
Household with single adult - -
Service/ repair occupation +
Personal income category +
Vehicle type is small + +
Strategy Adoption
Buy a car stereo system +
Get a better car -
Time since getting a better car +
Get a fuel efficient car -
Time since getting a fuel efficient car +
Hire somebody to do house or yard work +
Cao and Mokhtarian
19
( Table 4. Continued)
Buy a car
stereo system
Get a better
car
Get a fuel
efficient car
Strategy Adoption
Time since hiring domestic help -
Adopt compressed work week +
Squared time since changing to driving alone +
Change jobs closer to home + +
Time since changing jobs closer to home -
SD = Short Distance LD = Long Distance
Similarly, those liking long- distance travel overall are more likely to consider these two
strategies. Liking long- distance travel overall is strongly correlated with liking ( 0.409) and
desiring more ( 0.259) long- distance travel by personal vehicle, although less strongly than its
correlation with liking ( 0.538) and desiring more ( 0.368) long- distance travel by air. The
connection to long- distance vehicle travel is clear: someone liking and desiring more of such
travel is inclined to consider a better/ fuel- efficient vehicle for making such travel even more
enjoyable. The potential connection to long- distance air travel is more subtle, but still plausible.
First, even air travel generally involves airport ground access in a personal vehicle, and in any
case individuals doing a lot of flying are likely to be doing above- average amounts of traveling
in general, and personal vehicle traveling in particular. The person who likes long- distance
traveling overall, and wants to do more of it, may be especially motivated to consider a fuel-efficient
car, so as to minimize – as far as possible – the environmental impacts of the travel she
wants to do. Thus, the liking for long- distance overall travel variable may be a marker for a
complex constellation of variables and relationships.
The pro- high density factor score is negatively associated with the consideration of getting a
better/ fuel- efficient car. This land- use factor is based on attitudes about residential density and
about proximity to services, and a positive pro- high density factor score may indicate those who
have an aversion to travel by auto and prefer travel by walking or transit ( Redmond, 2000).
Therefore, those with a higher score for this factor may be less likely to consider getting a car at
all ( whether better or fuel- efficient or not). Conversely, individuals who are frustrated are more
likely to consider these two strategies. Individuals who are frustrated are traveling less than
others and wanting to travel more ( Choo, Collantes, and Mokhtarian, 2001), so they are likely to
consider getting a better/ fuel- efficient car to support increased travel. Also, they may view these
two strategies as ways to increase control and/ or life satisfaction ( through an improved travel
environment and/ or saved travel costs), or at least to provide a welcome diversion from their
difficulties.
Being in a single- adult/ no- children household is negatively related to considering getting a
better/ fuel- efficient car. In this sample, such respondents tend to have much lower household
incomes, and tend to live in North San Francisco. Individuals with lower household incomes
may be less able to afford a newer car, while residents in an urban area ( similarly to the effect of
the pro- high density variable) may be less inclined to travel by car in general, in view of the
walking and transit options available. In addition, children in the household may trigger one to
consider getting a better/ fuel- efficient car; however, single respondents obviously lack such
motivation. Logically enough, those who currently drive a small car are more likely to consider
Cao and Mokhtarian
20
getting either a better car to “ trade up” to a more comfortable means of travel, or a fuel efficient
car to continue saving monetary costs of travel.
Interestingly, the previous adoption of changing jobs closer to home is positively associated with
the consideration of getting a better/ fuel- efficient car, and the time since changing jobs closer to
home decreases the probability of considering getting a fuel efficient car. Although there are
many stimuli to affect a change in job, in most cases, such a change involves a higher personal
income or a higher status. Therefore, plausibly, individuals are more likely to consider getting a
newer car as a symbol of their advancement after they get a better salary or a higher position; the
more recently they have changed jobs, the more likely they are to consider getting a fuel efficient
car ( that generally replaces an older car). It is noteworthy that the former adoption of each of
these two auto improvement strategies affects its reconsideration in the same way: the former
adoption of each strategy has a negative impact on the consideration of the same strategy; and
the longer ago individuals have adopted each strategy, the more likely they are to reconsider it.
The former relationship suggests that the previous adoption may be still in force, and thus
individuals are less likely to reconsider it. But the car previously obtained becomes obsolete
and/ or mechanically unreliable as more time elapses, which would motivate individuals to
reconsider getting a newer car. In contrast, the former adoption of buying a car stereo system
increases the probability of its reconsideration. The implication is that people generally either
never buy a car stereo system or do so repeatedly. Given that it has once been adopted, it is
natural to expect that a car stereo system will be reconsidered, especially when individuals are no
longer satisfied with their current ones or when they are getting another car.
Adventure seekers are more likely to consider buying a car stereo system, perhaps to enhance the
entertainment value of traveling. Similarly, when we developed the model of consideration of
getting a fuel efficient car, we found that the adventure seeker factor score is one indicator of
liking for long- distance travel overall, and that it is positively associated with the consideration
of getting a fuel efficient car. It makes sense that adventure seekers are more likely to consider
auto improvement strategies. However, when we tried to include both variables in the
specification for the fuel- efficient car model, only the liking for long- distance travel overall was
significant. Therefore, we retained it rather than the adventure seeker factor score in the final
model.
3.2.1 Buy a car stereo system
The purpose of installing a car stereo system is to make the time spent in the vehicle more
enjoyable, or to cater to a certain kind of personality. Table 5 presents the model of
consideration of buying a car stereo system. The proportion of information in the data explained
by the model, ρ 2 , is 0.450. The proportion of information in the data explained by the market
share model, MS ρ 2 , is 0.385. This means that all explanatory variables other than the constant
term only explain 6.5 additional percentage points of information in the data. However, the final
model re- estimated with the constant term constrained to equal zero also resulted in a ρ 2 of
0.450 ( the difference is extremely small, about 0.00002), meaning that the true variables ( all
explanatory variables other than the constant term) carry essentially the full explanatory power
of the model.
Cao and Mokhtarian
21
Table 5. Model of Consideration of “ Buy a Car Stereo System”
Variable Estimated coefficient p- value
Constant 0.282 0.845
Objective Mobility
Frequency of entertainment travel ( SD) 0.0863 0.008
Frequency of other purpose travel ( SD) 0.0696 0.018
Weekly miles in a train/ BART/ light rail ( SD) - 0.00346 0.046
Sum of log of miles for each trip by air ( LD) - 0.0190 0.016
Subjective Mobility
Travel for grocery shopping ( SD) 0.269 0.008
Relative Desired Mobility
Overall ( SD) - 0.278 0.025
Personality
Adventure seeker factor score 0.316 0.003
Mobility Constraints
Limitations on driving during the day 1.689 0.000
Demographics
Age - 0.404 0.005
Female - 0.456 0.011
Year of personal vehicle - 0.0356 0.009
Service/ repair occupation 0.963 0.004
Strategy Adoption
Buy a car stereo system 0.433 0.015
Adopt compressed work week 0.731 0.004
Number of observations 1172
Log likelihood at 0 ( LL( 0)) - 812.369
Log likelihood of market share ( MS) model ( LL( MS)) - 499.215
Log likelihood at convergence ( LL( final)) - 446.454
MS ρ 2 [ 1 - ( LL( MS)/ LL( 0))] 0.385
ρ 2 [ 1 - ( LL( final)/ LL( 0))] 0.450
ρ 2 ( without the constant term) 0.450
Adjusted ρ 2 { 1 - [ LL( final)- # of coefficients]/ LL( 0)} 0.432
SD = Short Distance LD = Long Distance
Six mobility variables are significant in this model. Both frequency of short- distance travel for
entertainment and that of short- distance travel for other purposes are positively associated with
the consideration of buying a car stereo system. As mentioned above, individuals could make
their travel more pleasant by listening to the radio ( or tapes, CDs), and thereby reduce the
disutility of travel time. Thus, it makes sense that those who actually engage in a lot of travel are
more likely to consider this strategy. Moreover, individuals with a higher frequency of short-distance
travel for entertainment may indicate those who enjoy an “ audio- rich” environment, and
thus are more likely to consider buying a car stereo system. Travel for other purposes means
discretionary travel to a great extent; for individuals doing a lot of such travel, a sound system
increases the enjoyment of travel itself. Similarly, individuals who perceive that they do a lot of
travel for grocery shopping are more likely to consider this strategy, presumably because this
strategy could make travel time more tolerable. Also, in this sample, a higher subjective
mobility of travel for grocery shopping partly indicates those who have children in the household;
such individuals may be more likely to consider buying a car stereo system to cater to diverse
Cao and Mokhtarian
22
family needs. On the other hand, short- distance weekly miles in a train is negatively associated
with the consideration of buying a car stereo system. Individuals traveling longer distances by
train presumably spend less time in cars, so it is logical that they are also less likely to consider
this strategy. Those who actually do a lot of long- distance travel by air are also less likely to
consider this strategy. In this sample, greater distances of or higher frequencies of long- distance
travel by air are somewhat associated with disliking travel by personal vehicle ( correlation with
short- distance travel liking: - 0.113; correlation with long- distance travel liking: - 0.168) and/ or
not wanting to do more travel by personal vehicle ( correlation with short- distance relative
desired mobility: - 0.099; correlation with long- distance relative desired mobility: - 0.121).
Therefore, this relationship is quite plausible. The relative desire for short- distance travel overall
has a negative impact on the consideration of buying a car stereo system. Wanting to travel less
overall short- distance means individuals are burdened by the travel they are currently doing, so
they are more likely to consider this strategy to make that travel more pleasant.
Individuals having limitations on driving during the day are more likely to consider buying a car
stereo system, presumably to mitigate travel stress and make the travel more comfortable. Age
negatively affects the consideration of this strategy. It is logical that younger people are more
likely to consider installing a sound system in their vehicles. Likewise, men are more likely to
consider buying a car stereo system. In this sample, men are significantly associated with a
cluster of personality and lifestyle factors; for example, being impatient or aggressive, frustrated,
adventure and status seeking. Therefore it is plausible that men are more likely to consider this
strategy. Having a service/ repair occupation is also positively associated with the consideration
of buying a car stereo system, which is natural since such occupations often involve a lot of
work- related travel.
Interestingly, the former adoption of a compressed work week is positively associated with the
consideration of buying a car stereo system. One purpose of adopting a compressed work week
is to reduce the total number of commute trips. A car stereo system serves as a complement of
this commute– reduction strategy, making commuting more enjoyable. It is plausible that
consideration of such a travel- maintaining strategy would follow the adoption of a travel-reduction
strategy, to further ameliorate the disutility of the remaining travel that must be done.
Furthermore, for the particular travel- reduction strategy of a compressed work week, a car stereo
system may act as a cushion to relieve the incremental stress of the longer workday ( 9 or 10
hours instead of the usual 8).
3.2.2 Get a better car
In this sample, the consideration rate of getting a better car is 37.3%, the largest among all
strategies presented in this study. The personal vehicle is the dominant means of passenger
travel in the U. S. The Federal Highway Administration ( 1997) found that travel by private
vehicle accounted for 86% of all person trips and 91% of all person miles in 1995. Therefore, it
is natural that the consideration rate of getting a better car ranks first.
Table 6 presents the model of consideration of getting a better car. The proportion of
information in the data explained by the model, ρ 2 , is 0.183, smallest among all the models.
The proportion of information in the data explained by the market share model, MS ρ 2 , is 0.039.
This means that all explanatory variables other than the constant term explain 14.4 additional
Cao and Mokhtarian
23
percentage points of information in the data. The final model re- estimated without the constant
term resulted in a ρ 2 of 0.167, meaning that the true variables shoulder 87% of the full
explanatory power of the model. In the model, the constant term is positive and significant,
meaning that the average impact of the unobserved variables is in the direction of considering
getting a better car.
Table 6. Model of Consideration of “ Get a Better Car”
Variable Estimated coefficient p- value
Constant 6.235 0.000
Objective Mobility
Weekly miles to eat a meal ( SD) 0.0137 0.009
Subjective Mobility
Take others where they need to go ( SD) 0.264 0.000
Relative Desired Mobility
Travel by bus ( SD) - 0.216 0.002
Travel Liking
Travel for entertainment ( SD) 0.269 0.008
Overall ( LD) 0.195 0.023
Attitudes
Pro- high density factor score - 0.244 0.010
Personality
Frustrated factor score 0.284 0.001
Demographics
Year of personal vehicle - 0.0957 0.000
Years lived in the U. S. - 0.0149 0.009
Household with single adult - 0.382 0.029
Personal income category 0.247 0.000
Vehicle type is small 0.383 0.020
Strategy Adoption
Get a better car - 1.048 0.000
Time since getting a better car 0.156 0.000
Hire somebody to do house or yard work 0.438 0.021
Time since hiring domestic help - 0.0988 0.001
Squared time since changing to driving alone 0.0126 0.017
Change jobs closer to home 0.304 0.050
Number of observations 1118
Log likelihood at 0 - 774.939
Log likelihood of MS model - 744.427
Log likelihood at convergence - 633.214
MS ρ 2 0.039
ρ 2 0.183
ρ 2 ( without the constant term) 0.167
Adjusted ρ 2 0.158
SD = Short Distance LD = Long Distance
Weekly miles to eat out is positively related to considering getting a better car. More travel for
eating out implies that an individual is social- oriented, with a lifestyle focused outside the home;
therefore she is more likely to consider getting a better car to make that “ on- the- go” lifestyle
more comfortable. Similarly, those perceiving that they do a lot of travel for taking others where
Cao and Mokhtarian
24
they need to go are more likely to consider this strategy. These individuals tend to be
family/ community- oriented, and they also like such travel. So their consideration of a better car
may imply the desire for a larger, more comfortable car, or a car with more amenities to
accommodate their passengers, or the need for a safer, more reliable means of transportation.
Logically, those desiring more travel by bus are less likely to consider getting a better car.
Years lived in the U. S., acting as a proxy for age, is negatively associated with the consideration
of getting a better car, which is logical since younger people are more likely to be starting out
with a lower- end car. Individuals with higher personal incomes are more likely to consider
getting a better car, as expected. Interestingly, the previous adoption of hiring domestic help is
positively associated with the consideration of getting a better car; the more recently individuals
have hired domestic help, the more likely they are to consider this strategy. Individuals hiring
domestic help tend to be those who have higher household incomes and personal incomes.
Therefore, it is logical that they tend to consider getting a better car. These relationships also
suggest that they may want or need to engage in more travel after adopting this time- buying
strategy. The squared time since changing from another mode to driving alone positively affects
the consideration of getting a better car. That is, the longer ago individuals changed to driving
alone, the more likely they are to consider replacing their old cars.
3.2.3 Get a fuel efficient car
Table 7 presents the model of consideration of getting a fuel efficient car. The proportion of
information in the data explained by the model, ρ 2 , is 0.229. The proportion of information in
the data explained by the market share model, MS ρ 2 , is 0.132. This means that all explanatory
variables other than the constant term explain 9.7 additional percentage points of information in
the data. The final model re- estimated without the constant term resulted in a ρ 2 of 0.220,
meaning that the true variables carry 96% of the full explanatory power of the model.
Sixteen variables are significant in the model. Individuals who do a lot of short- distance travel
overall are more likely to consider getting a fuel efficient car. Since short- distance total weekly
miles is dominated by miles traveled by personal vehicle, it is not surprising that such individuals
would want a fuel efficient car, to reduce the costs of travel. For the same reason, those
perceiving that they do a lot of long- distance travel by personal vehicle are more likely to
consider getting a fuel efficient car. Similar to short- distance weekly miles in a train discussed
in Section 3.2.1, individuals perceiving that they do a lot of travel by train tend to be those who
are not auto- dependent, so it is plausible that they are less likely to consider getting a fuel
efficient car.
Individuals advocating environmental protection are more likely to consider getting a fuel
efficient car to reduce their personal energy consumption and impacts on the environment.
Status seekers view the automobile as a status symbol. Since most fuel efficient vehicles are
compact or small, and since there is a tradeoff between fuel efficiency and engine performance,
it is plausible that status seekers are less likely to consider getting a fuel efficient car ( or at least a
car for which fuel efficiency is stressed as a selling point).
Cao and Mokhtarian
25
Table 7. Model of Consideration of “ Get a Fuel Efficient Car”
Variable Estimated coefficient p- value
Constant 4.529 0.000
Objective Mobility
Total weekly miles ( SD) 0.00120 0.001
Subjective Mobility
Travel by train/ BART/ light rail ( SD) - 0.167 0.019
Travel by personal vehicle ( LD) 0.115 0.049
Travel Liking
Travel for entertainment ( SD) 0.203 0.049
Overall ( LD) 0.209 0.018
Attitudes
Pro- environmental solutions factor score 0.470 0.000
Pro- high density factor score - 0.231 0.031
Lifestyle
Frustrated factor score 0.271 0.002
Status seeker factor score - 0.233 0.013
Demographics
Year of personal vehicle - 0.0820 0.000
Household with single adult - 0.579 0.001
Vehicle type is small 0.579 0.001
Strategy Adoption
Get a fuel efficient car - 0.563 0.006
Time since getting a fuel efficient car 0.0968 0.000
Change jobs closer to home 0.780 0.000
Time since changing jobs closer to home - 0.126 0.007
Number of observations 1155
Log likelihood at 0 - 800.585
Log likelihood of MS model - 694.633
Log likelihood at convergence - 617.280
MS ρ 2 0.132
ρ 2 0.229
ρ 2 ( without the constant term) 0.220
Adjusted ρ 2 0.208
SD = Short Distance LD = Long Distance
3.3 Mobile Phone
The mobile phone bundle contains only one individual strategy. Table 8 presents the model of
consideration of mobile phones. The proportion of information in the data explained by the
model, ρ 2 , is 0.202. The proportion of information in the data explained by the market share
model, MS ρ 2 , is 0.124. This means that all explanatory variables other than the constant term
explain 7.8 additional percentage points of information in the data. The final model re- estimated
without the constant term resulted in a ρ 2 of 0.191, meaning that the true variables carry about
95% of the full explanatory power of the model.
Cao and Mokhtarian
26
A cursory review of the model indicates that the consideration of a mobile phone is greatly
affected by objective mobility. Eight objective mobility variables are significant in this model:
six with positive signs, and the other two being negatively associated with the consideration of
mobile phones. The positive association is quite natural: the more one travels, the more useful it
becomes to have mobile communication capabilities. The two negative coefficients relate to
weekly miles of grocery shopping travel and taking others where they need to go. In both cases,
the frequency of travel for that purpose is also in the model, with the expected positive sign.
Thus, the negative effect of the distance variables partly modifies the direct positive effect of the
frequency variables. Generally, the combined impact of the frequency- distance pair of variables
in a given case is still positive. Specifically, the combined impact of frequency and distance for
grocery shopping is positive for three- quarters of the sample, and the impact of taking others
where they need to go is positive for 57.8% of the sample. In any case, it is plausible that the
perceived utility of a mobile phone would be higher for a person making many trips than for one
making fewer trips covering the same or longer distance, because of the increased uncertainty
and scheduling complexity associated with making many trips.
Table 8. Model of Consideration of “ Get a Mobile Phone” ( Bundle 2)
Variable Estimated coefficient p- value
Constant - 2.341 0.000
Objective Mobility
Frequency of work/ school- related travel ( SD) 0.0572 0.004
Frequency of grocery shopping travel ( SD) 0.0927 0.002
Frequency of travel taking others where they need to go ( SD) 0.0739 0.006
Total weekly miles ( SD) 0.00104 0.006
Weekly miles of grocery shopping travel ( SD) - 0.0178 0.025
Weekly miles to eat a meal ( SD) 0.0185 0.001
Weekly miles of travel taking others where they need to go ( SD) - 0.0173 0.002
Sum of log of miles for each trip by air ( LD) 0.0122 0.033
Subjective Mobility
Travel for entertainment ( SD) 0.149 0.039
Travel by personal vehicle ( SD) 0.158 0.009
Mobility Constraints
Limitations on driving on the freeway 0.645 0.019
Demographics
Age - 0.391 0.000
Anyone in household needing special care 0.973 0.004
Strategy Adoption
Buy a car stereo system 0.284 0.034
Buy a mobile phone - 0.978 0.000
Number of observations 1263
Log likelihood at 0 - 875.445
Log likelihood of MS model - 766.457
Log likelihood at convergence - 698.661
MS ρ 2 0.124
ρ 2 0.202
ρ 2 ( without the constant term) 0.191
Adjusted ρ 2 0.184
SD = Short Distance LD = Long Distance
Cao and Mokhtarian
27
Similarly, both subjective mobility effects are positive. If individuals perceive that they do a lot
of short- distance entertainment travel and travel by personal vehicle, they are more likely to
consider mobile phones to utilize their travel time effectively and to coordinate with other people.
Individuals having limitations on driving on the freeway are more likely to consider obtaining a
mobile phone, perhaps to alleviate higher- than- average fears about safety, or travel stress in
general.
Two demographic variables enter the model. The negative sign of the age variable indicates
younger people are more likely to consider mobile phones – a logical result for a technological
innovation still in its infancy at the time the data were collected ( 1998). For those who have
anyone in the household needing special care, mobile phones could provide direct and timely
communications with the family whenever they are working or traveling. Thus, the positive
coefficient of this variable is logical.
The former adoption of a car stereo system has a positive impact on the consideration of mobile
phones. Both are considered travel- maintaining strategies, and may complement each other. On
the other hand, prior adoption of a mobile phone has a strongly negative impact on considering
the same strategy, which is natural since the prior adoption is probably still in force.
3.4 Work- Schedule Changes
Three strategies, “ Change work trip departure time”, “ Adopt flextime” and “ Adopt compressed
work week”, were grouped into the bundle of work- schedule changes. These strategies share
some common characteristics, such as an adjusted commute schedule, likely avoidance of peak
period congestion, possible impacts on the household and so on. However, “ Change work trip
departure time” is different from the latter two strategies in some important ways. For example,
the latter strategies require support from the employer before they can be adopted ( Salomon and
Mokhtarian, 1997). Perhaps for this reason, the model for “ Change work trip departure time”
only shares one explanatory variable with the models for the other two strategies, while the latter
two models have four common variables.
Table 9 summarizes the individual models in Bundle 3. Four out of the five objective mobility
variables appearing in any of these models have positive impacts on the consideration of
changing work trip departure time. Higher objective mobility, especially with respect to
commuting, exposes an individual to greater travel stress and congestion, and changing work trip
departure time is a logical way to try to reduce such exposure. But objective mobility has a
weaker role in the consideration of the other two strategies, with only one objective mobility
variable significant in the model for compressed work week. This suggests that other factors
play a more important role in the consideration of the latter two strategies in this bundle.
However, the perceived amount of commuting is significant for the consideration of these two
strategies, consistent with expectations.
Cao and Mokhtarian
28
Table 9. Models of Consideration of Work- Schedule Changes ( Bundle 3)
Change work trip
departure time
Adopt flextime Adopt compressed
work week
N 1265 1278 1278
MS ρ 2 0.332 0.388 0.476
ρ 2 0.438 0.477 0.547
Adjusted ρ 2 0.421 0.464 0.534
Variable
Objective Mobility
Frequency of commuting ( SD) +
Frequency of grocery shopping travel ( SD) +
Weekly miles in a bus ( SD) +
Weekly miles of commuting ( SD) +
Commute time +
Subjective Mobility
Commute ( SD) + +
Take others where they need to go ( SD) +
Relative Desired Mobility
Overall ( SD) -
Travel by air ( LD) +
Travel Liking
Travel for entertainment ( SD) +
Attitudes
Pro- environmental solutions factor score + +
Commute benefit factor score -
Personality
Adventure seeker factor score + +
Lifestyle
Family & community- oriented factor score +
Workaholic factor score +
Mobility Constraints
Limitations on flying in an airplane +
Limitations on riding a bicycle +
Percent of time a vehicle is available -
Demographics
Number of vehicles in household -
Years lived in the U. S. -
Anyone in household needing special care +
Household with two or more adults and
children
+
Sales occupation -
Full- time worker + +
Strategy Adoption
Change work trip departure time +
Time since hiring domestic help -
Adopt flextime +
Time since adopting flextime -
Adopt compressed work week +
Change jobs closer to home + +
SD = Short Distance LD = Long Distance
Cao and Mokhtarian
29
In addition to the subjective amount of commute travel, four other variables are significant in
two of the three models in this bundle. Since congestion deteriorates air quality and increases
gas consumption, individuals advocating environmental protection are more likely to consider
work- schedule changes, specifically the formal employer- based alternatives of flextime and
compressed work week. Adventure seekers may be considering adopting work schedule change
strategies in order to enjoy higher commute speeds or to save time for more highly desired
activities. Generally, full- time workers commute at rush hour every weekday, so they will have
more motivation to consider these strategies. It is interesting that adoption of the higher- cost,
longer- term strategy of “ Change jobs closer to home” is positively associated with the lower- cost
employer- based work schedule change strategy. Either the job change did not reduce the
commute to a satisfactory level, or the effectiveness of the move diminished over time, either
way causing the individual to cycle back to considering lower- cost strategies ( Raney, et al.,
2000). It is also possible that the new employer is more supportive of alternative work schedules
than the previous one was.
Consistent with expectation, the previous adoption of each work schedule change strategy is
positively associated with the consideration of the same strategy.
3.4.1 Change work trip departure time
Fourteen variables are significant in the model of consideration of “ Change work trip departure
time”. The proportion of information in the data explained by the model, ρ 2 , is 0.438. The
proportion of information in the data explained by the market share model, MS ρ 2 , is 0.332.
This means that all explanatory variables other than the constant term explain 10.6 additional
percentage points of information in the data. The final model re- estimated without the constant
term resulted in a ρ 2 of 0.415, meaning that the true variables carry about 95% of the full
explanatory power of the model.
Similar to objective mobility, higher subjective mobility ( for taking others where they need to go,
in this case) makes this strategy attractive, therefore an individual is more likely to consider it.
Those who like short- distance entertainment travel may consider this strategy to better
coordinate their schedule with after- work entertainment activities, or to avoid congestion so as to
have more time for such activities. Individuals who desire less short- distance travel overall are
more likely to consider this strategy, presumably also to save travel time and reduce stress by
avoiding congestion. Conversely, individuals who enjoy higher benefits of their current
commute definitely find this strategy less appealing.
The positive coefficient of the workaholic factor indicates that the more priority an individual
gives to work, the more likely she would be to consider this strategy, probably to stay longer at
work. This result duplicates a finding in the previous study ( Raney, et al., 2000).
Mobility constraints on flying or bicycling may indicate individuals who feel anxious about
travel or have physical constraints on traveling; this strategy could help them avoid the anxiety
experienced during peak period commuting and thus make the commute more comfortable.
Years lived in the U. S. a