Who Likes Traveling?
Models of the Individual’s Affinity for Various Kinds of Travel
David T. Ory
Department of Civil and Environmental Engineering
and
Institute of Transportation Studies
University of California, Davis
Davis, CA 95616
voice: ( 415) 378- 9102
fax: ( 530) 752- 6572
e- mail: dtory@ ucdavis. edu
and
Patricia L. Mokhtarian
Department of Civil and Environmental Engineering
and
Institute of Transportation Studies
University of California, Davis
Davis, CA 95616
voice: ( 530) 752- 7062
fax: ( 530) 752- 7872
e- mail: plmokhtarian@ ucdavis. edu
Research Report UCD- ITS- RR- 04- 20
July 2004
This research is funded by the DaimlerChrysler Corporation and the National Science
Foundation Integrative Graduate Education and Research Traineeships ( IGERT) program.
ii
TABLE OF CONTENTS
DOCUMENTS PRODUCED BY ATTITUDES TOWARDS MOBILITY PROJECT........... iv
LIST OF TABLES AND FIGURES.................................................................................. viii
ACKNOWLEDGEMENTS................................................................................................. x
EXECUTIVE SUMMARY ................................................................................................. xi
1. INTRODUCTION ....................................................................................................... 1
1.1 Background ............................................................................................................. 1
1.2 What are the Sources for a Liking of Travel? .......................................................... 3
1.3 Data......................................................................................................................... 6
2. VARIABLES............................................................................................................... 7
2.1 The Dependent Variables: Travel Liking ................................................................. 7
2.2 The Potential Explanatory Variables ..................................................................... 11
3. MODELS.................................................................................................................. 19
3.1 General Specification Issues................................................................................. 19
3.2 Discussion of Variables Appearing in Multiple Short- Distance Models ................. 21
3.3 Overall Short- Distance Travel ............................................................................... 26
3.4 Commute to Work/ School ..................................................................................... 28
3.5 Short- Distance Work/ School- Related.................................................................... 31
3.6 Short- Distance Entertainment/ Recreation/ Social .................................................. 33
3.7 Short- Distance Personal Vehicle........................................................................... 35
3.8 Short- Distance Bus ............................................................................................... 37
iii
3.9 Short- Distance Rail ............................................................................................... 38
3.10 Short- Distance Walk/ Jog/ Bicycle ........................................................................ 40
3.11 Discussion of Variables Appearing in Multiple Long- Distance Models................ 41
3.12 Long- Distance Overall......................................................................................... 43
3.13 Long- Distance Work/ School- Related .................................................................. 44
3.14 Long- Distance Entertainment/ Recreation/ Social................................................. 45
3.15 Long- Distance Personal Vehicle ......................................................................... 46
3.16 Long- Distance Airplane ....................................................................................... 48
4. SUMMARY AND DISCUSSION .............................................................................. 49
4.1 Summary and Implications .................................................................................... 49
4.2 Comparison of Hypothesized Bases for Travel Liking and Model Results............ 51
4.3 Directions for Future Research ............................................................................. 51
REFERENCES ............................................................................................................... 54
A. APPENDIX: DESCRIPTIVE STATISTICS OF VARIABLES ............................... 57
iv
DOCUMENTS PRODUCED BY ATTITUDES TOWARDS MOBILITY PROJECT
Journal Articles Produced by this Project to Date
Cao, Xinyu and Patricia L. Mokhtarian ( forthcoming) How do individuals adapt their
personal travel? A conceptual exploration of the consideration of travel- related
strategies. Transport Policy.
Cao, Xinyu and Patricia L. Mokhtarian ( forthcoming) How do individuals adapt their
personal travel? Objective and subjective influences on the consideration of travel-related
strategies. Transport Policy.
Choo, Sangho and Patricia L. Mokhtarian ( 2004) How do people respond to congestion
policies? Exploring the individual consideration of travel- related strategy bundles. Under
review for publication.
Choo, Sangho and Patricia L. Mokhtarian ( 2004) What type of vehicle do people drive?
The role of attitude and lifestyle in influencing vehicle type choice. Transportation
Research A 38( 3), 201- 222.
Choo, Sangho, Gustavo O. Collantes, and Patricia L. Mokhtarian ( forthcoming) Wanting
to travel, more or less: Exploring the determinants of a perceived deficit or surfeit of
personal travel. Transportation.
Clay, Michael J. and Patricia L. Mokhtarian ( 2004) Personal travel management: The
adoption and consideration of travel- related strategies. Transportation Planning and
Technology 27( 3) ( June).
Collantes, Gustavo O. and Patricia L. Mokhtarian ( 2002) Qualitative subjective
assessments of personal mobility: Exploring the magnifying and diminishing cognitive
mechanisms involved. Under review for publication.
Handy, Susan L., Lisa Weston, and Patricia L. Mokhtarian ( 2003) Driving by choice or
necessity? The case of the soccer mom and other stories. Paper presented at the 82nd
annual meeting of the Transportation Research Board, Washington, DC, January, draft
available on conference CD- ROM.
Handy, Susan L., Lisa Weston, and Patricia L. Mokhtarian ( 2004) Driving by choice or
necessity? ( later version of 2003 paper) Manuscript under review for publication.
Mokhtarian, Patricia L., Ilan Salomon, and Lothlorien S. Redmond ( 2001) Understanding
the demand for travel: It's not purely “ derived”. Innovation: The European Journal of
Social Science Research 14( 4), 355- 380.
Mokhtarian, Patricia L. and Ilan Salomon ( 2001) How derived is the demand for travel?
Some conceptual and measurement considerations. Transportation Research A 35( 8),
695- 719.
v
Ory, David T. and Patricia L. Mokhtarian ( 2004) When is getting there half the fun?
Modeling the liking for travel. Manuscript under review for publication.
Ory, David T., Patricia L. Mokhtarian, Ilan Salomon, Lothlorien S. Redmond, Gustavo O.
Collantes, and Sangho Choo ( 2004) When is commuting desirable to the individual?
Growth and Change 35( 3) ( Summer), special issue on Advances in Commuting Studies,
Peter Nijkamp and Jan Rouwendal, eds., 334- 359.
Redmond, Lothlorien S. and Patricia L. Mokhtarian ( 2001) The positive utility of the
commute: Modeling ideal commute time and relative desired commute amount.
Transportation 28( 2) ( May), 179- 205.
Salomon, Ilan and Patricia L. Mokhtarian ( 1999) Travel for the fun of it. Access ( a
publication of the University of California Transportation Center) 15 ( Fall), 26- 31. Available
at www. uctc. net/ access/ access15. pdf or ( without graphics) .../ access15lite. pdf.
Salomon, Ilan and Patricia L. Mokhtarian ( 1998) What happens when mobility- inclined
market segments face accessibility- enhancing policies? Transportation Research D
3( 3), 129- 140.
Salomon, Ilan and Patricia L. Mokhtarian ( 2002) Driven to travel: The identification of
mobility- inclined market segments. Chapter 22 in William R. Black and Peter Nijkamp,
eds., Social Change and Sustainable Transport. Bloomington, IN: Indiana University
Press, pp. 173- 179. Included in the Regional Futures Compendium of the Capital Region
Institute ( Valley Vision), Sacramento, California.
Schwanen, Tim and Patricia L. Mokhtarian ( 2004) The extent and determinants of disson-ance
between actual and preferred residential neighborhood type. Environment and
Planning B.
Schwanen, Tim and Patricia L. Mokhtarian ( 2003a) Does dissonance between desired
and current neighborhood type affect individual travel behaviour? An empirical
assessment from the San Francisco Bay Area. Proceedings of the European Transport
Conference ( ETC), October 8- 10, 2003, Strasbourg, France.
Schwanen, Tim and Patricia L. Mokhtarian ( 2003b) The role of attitudes toward travel
and land use in residential location behavior: Some empirical evidence from the San
Francisco Bay Area. Under review for publication.
Schwanen, Tim and Patricia L. Mokhtarian ( forthcoming) What affects commute mode
choice: Neighborhood physical structure or preferences toward neighborhoods? Journal
of Transport Geography.
Schwanen, Tim and Patricia L. Mokhtarian ( 2004) What if you live in the wrong
neighborhood? The impact of residential neighborhood type dissonance on distance
traveled. Under review for publication.
vi
Reports Produced by this Project to Date
Cao, Xinyu and Patricia L. Mokhtarian ( 2003) Modeling the Individual Consideration of
Travel- Related Strategies. Research Report UCD- ITS- RR- 03- 3, Institute of Transpor-tation
Studies, University of California, Davis, June.
Available at www. its. ucdavis. edu/ publications/ 2003/ RR- 03- 3. pdf.
Choo, Sangho and Patricia L. Mokhtarian ( 2002) The Relationship of Vehicle Type
Choice to Personality, Lifestyle, Attitudinal, and Demographic Variables. Research
Report, Institute of Transportation Studies, University of California, Davis, October.
Available at www. its. ucdavis. edu/ publications/ 2002/ RR- 02- 06. pdf.
Choo, Sangho and Patricia L. Mokhtarian ( 2004) Modeling the Consideration of Travel-
Related Strategy Bundles. Research Report, Institute of Transportation Studies,
University of California, Davis, April.
Available at http:// www. its. ucdavis. edu/ publications/ 2004/ UCD- ITS- RR- 04- 07. pdf.
Choo, Sangho, Gustavo O. Collantes, and Patricia L. Mokhtarian ( 2001) Modeling
Individuals' Relative Desired Travel Amounts. Research Report UCD- ITS- RR- 01- 13,
Institute of Transportation Studies, University of California, Davis, November.
Available at www. its. ucdavis. edu/ publications/ 2001/ RR- 01- 13. pdf.
Clay, Michael J. and Patricia L. Mokhtarian ( 2002) The Adoption and Consideration of
Commute- Oriented Travel Alternatives. Research Report UCD- ITS- RR- 02- 04, Institute
of Transportation Studies, University of California, Davis, September.
Available at www. its. ucdavis. edu/ publications/ 2002/ RR- 02- 04. pdf.
Collantes, Gustavo O. and Patricia L. Mokhtarian ( 2002) Determinants of Subjective
Assessments of Personal Mobility. Research Report, Institute of Transportation Studies,
University of California, Davis, August.
Available at www. its. ucdavis. edu/ publications/ 2002/ RR- 02- 11. pdf.
Curry, Richard W. ( 2000) Attitudes toward Travel: The Relationships among Perceived
Mobility, Travel Liking, and Relative Desired Mobility. Master’s Thesis, Department of
Civil and Environmental Engineering, University of California, Davis, June. Research
Report UCD- ITS- RR- 00- 06, Institute of Transportation Studies, University of California,
Davis. Available at www. its. ucdavis. edu/ publications/ 2000/ RR- 00- 06. pdf.
Ory, David T. and Patricia L. Mokhtarian ( 2004) Who Likes Traveling? Models of the
Individual’s Affinity for Various Kinds of Travel. Research Report UCD- ITS- RR- 04- xx,
Institute of Transportation Studies, University of California, Davis, June.
Redmond, Lothlorien S. and Patricia L. Mokhtarian ( 2001) Modeling Objective Mobility:
The Impact of Travel- Related Attitudes, Personality, and Lifestyle on Distance Traveled.
Research Report UCD- ITS- RR- 01- 09, Institute of Transportation Studies, University of
California, Davis, June. Available at http:// repositories. cdlib. org/ itsdavis/ UCD- ITS- RR-
01- 09/
vii
Redmond, Lothlorien S. ( 2000) Identifying and Analyzing Travel- related Attitudinal,
Personality, and Lifestyle Clusters in the San Francisco Bay Area. Master’s Thesis,
Transportation Technology and Policy Graduate Group, Institute of Transportation Studies,
University of California, Davis, September. Research Report UCD- ITS- RR- 00- 08.
Available at www. its. ucdavis. edu/ publications/ 2000/ RR- 00- 08. pdf
viii
LIST OF TABLES AND FIGURES
Table ES. 1: Short- Distance Travel Liking Dependent Variables ( N= 1,358) .................... xii
Table ES. 2: Long- Distance Travel Liking Dependent Variables ( N= 1,358) .................... xiii
Table ES. 3: Summary of Short- Distance Travel Liking Models ...................................... xvi
Table ES. 4: Summary of Long- Distance Travel Liking Models...................................... xvii
Table ES. 5: Comparison of Hypotheses and Travel Liking Model Results................... xviii
Table 1.1: Socio- Demographic Characteristics of Sample ( N= 1,358) .............................. 7
Table 2.1: Short- Distance Travel Liking Dependent Variables ( N= 1,358) ........................ 8
Table 2.2: Long- Distance Travel Liking Dependent Variables ( N= 1,358)......................... 9
Table 2.3: Pattern Matrix for Attitude Factors ( commuters only, N= 1,427)..................... 14
Table 2.4: Pattern Matrix for Personality Factors ( N= 1,904)........................................... 17
Table 2.5: Pattern Matrix for Lifestyle Factors ( N= 1,904) ............................................... 18
Table 3.1: Summary of Short- Distance Travel Liking Models......................................... 25
Table 3.2: Model of Liking for All Short- Distance Travel ( N= 1,321)................................ 29
Table 3.3: Model of Liking for Work/ School Commute Travel ( N= 1,338)........................ 30
Table 3.4: Model of Liking for Short- Distance Work/ School- Related Travel ( N= 1,351).. 32
Table 3.5: Model of Liking for Short- Distance Entertainment/ Recreation/ Social Travel ( N= 1,327)
............................................................................................................................... . 34
Table 3.6: Model of Liking for All Short- Distance Travel by Personal Vehicle ( N= 1,344) 35
Table 3.7: Model of Liking for All Short- Distance Travel by Bus ( N= 1,319).................... 37
Table 3.8: Model of Liking for All Short- Distance Travel by Rail ( N= 1,295).................... 39
ix
Table 3.9: Model of Liking for All Short- Distance Travel by Walking, Jogging, Bicycling
( N= 1,299)................................................................................................................. 40
Table 3.10: Summary of Long- Distance Travel Liking Models ....................................... 43
Table 3.11: Model of Liking for All Long- Distance Travel ( N= 1,345) .............................. 44
Table 3.12: Model of Liking for Long- Distance Work/ School- Related Travel ( N= 1,356) 45
Table 3.13: Model of Liking for Long- Distance Entertainment/ Recreation/ Social Travel
( N= 1,351)................................................................................................................. 46
Table 3.14: Model of Liking for All Long- Distance Travel by Personal Vehicle ( N= 1,318) 47
Table 3.15: Model of Liking for All Long- Distance Travel by Airplane ( N= 1,350) ........... 48
Table 4.1: Comparison of Hypotheses and Travel Liking Model Results ....................... 52
Table A. 1: Distributions for Short- Distance Travel Liking Variables ............................... 57
Table A. 2: Distributions for Long- Distance Travel Liking Variables ................................ 59
Table A. 3: Descriptive Statistics for Continuous Explanatory Variables ......................... 60
Table A. 4: Distributions for Ordinal Explanatory Variables ............................................. 61
Figure 1.1: Conceptual Model of an Individual's Affinity for Travel ................................... 3
Figure 3.1: Hypothesized Relationship between Certain Explanatory Variables and Dependent
Variables.................................................................................................................. 21
Figure 3.2: Average Shares of Mileage by Purpose ....................................................... 26
Figure 3.3: Average Shares of Mileage by Travel Mode................................................. 27
x
ACKNOWLEDGEMENTS
This research is funded by the DaimlerChrysler Corporation and the National Science Founda-tion
Integrative Graduate Education and Research Traineeships ( IGERT) program. The original
survey design and data collection were funded by the University of California Transportation
Center. We gratefully acknowledge the numerous colleagues who have previously worked on
this ongoing project, many of whose contributions have fed into the current report. In particular,
portions of Sections 1 and 2 heavily relied on previous reports in this series.
xi
EXECUTIVE SUMMARY
Do people actually like traveling? According to conventional wisdom, the answer is no: travel is
simply a means to the desired end of participating in spatially- separated activities. However,
substantial evidence ( Albertson, 1977; Beroldo, 2002; Edmonson, 1998; Larson, 1998;
Mokhtarian and Salomon, 1997; Richter, 1990; Higano and Orishimo, 1990; Shamir, 1991)
suggests that travel does more than play this purely utilitarian role. Rather, travel has some
qualities, such as an opportunity to mentally switch from the work realm to the personal realm of
daily life or to move quickly through space, that are desirable in themselves.
This report is part of an ongoing research program investigating the mobility and attitudes
towards travel of individuals. The portion of the research presented here derives relationships
between measures of Travel Liking ( how much an individual likes travel, differentiated by trip
purpose, mode, and length) and other variables in the data. The data set contains 1,358
residents of three neighborhoods in the San Francisco Bay Area, who work part- or full- time and
commute with some regularity.
A key premise of the entire research effort is that while individuals travel primarily to participate
in spatially- separated activities, there is an additional component driving some travel. We
believe individuals have a positive utility both for travel itself ( e. g. the sensation of motion and
movement through space that travel provides) and for activities that can be conducted while
traveling ( e. g. listening to music, talking on the telephone). A primary goal of the research effort
as a whole is to better understand the causes and motivations of this affinity for travel. The
modeling of Travel Liking is a key piece in this effort.
The types of variables in the data set can be segmented into ten general categories, namely:
Objective Mobility, Subjective Mobility, Relative Desired Mobility, Travel Liking, Attitudes,
Personality, Lifestyle, Excess Travel, Mobility Constraints, and Socio- demographics. Ultimately,
the broader research project will develop structural equations models to account for the many
interrelationships present among these variable groups. To more fully explore these
relationships, we first use single equation models for the major endogenous variable categories,
namely: Objective Mobility, Subjective Mobility, Travel Liking, and Relative Desired Mobility.
This report focuses on the single equation models for Travel Liking; previously published
companion reports and papers discuss the single equation models for the other three key
xii
endogenous variables ( see the list of Documents Produced by the Attitudes towards Mobility
Project at the front of this report).
Before examining the estimated single equation models of the Travel Liking variables, it is
interesting to look directly at the dependent variables. The Travel Liking measures ( varied by
distance, purpose, and mode) were captured in the survey by the following question: “ How do
you feel about traveling in each of the following categories? We are not asking how you feel
about the activity at the destination, but about the travel required to get there. Even if you
seldom or never travel in a certain category, you may still have a feeling about it.” Table ES. 1
summarizes the short- distance ( one- way trips of less than 100 miles) Travel Liking responses
for those ( so- called working- commuters, as defined previously) analyzed in this study. The
results ( at least in the absence of comparable data for other societies) support the stereotype of
car- loving Americans, in that only the non- motorized category of travel ( walking/ jogging/ bicy-cling)
received more “ like” and “ strongly like” responses than the personal vehicle category. Also
of interest is the relative contentment of most people ( even in this sample of residents of the
highly urbanized San Francisco metropolitan area) with commute travel, where only about 40
percent indicated any level of dislike.
Table ES. 1: Short- Distance Travel Liking Dependent Variables ( N= 1,358)
Travel Liking Variable Strongly
dislike Dislike Neutral Like Strongly
like
N 15 178 762 360 43
Overall
% 1.1 13.1 56.1 26.5 3.2
N 123 424 520 254 37
Commute
% 9.1 31.2 38.3 18.7 2.7
Work/ School- N 64 292 749 227 26
Related % 4.7 21.5 55.2 16.7 1.9
Entertain./ Social/ N 6 66 543 605 138
Recreation % 0.4 4.9 40.0 44.6 10.2
N 34 125 410 647 142
Personal Vehicle
% 2.5 9.2 30.2 47.6 10.5
N 389 473 384 103 9
Bus
% 28.6 34.8 28.3 7.6 0.7
N 161 231 540 384 42
Rail
% 11.9 17.0 39.8 28.3 3.1
N 54 66 332 663 243
Walk/ Jog/ Bicycle
% 4.0 4.9 24.4 48.8 17.9
xiii
The Travel Liking responses for long- distance travel are presented in Table ES. 2. Overall travel
is viewed favorably by most, as is travel by airplane and for entertainment/ social/ recreation
purposes. While these results may seem intuitive to some, they bring forth myriad questions:
What is generating a liking for personal vehicle travel? The sense of freedom it brings? A need
to be alone? A desire for status? What is driving dislike for work/ school- related long- distance
travel? Too much travel? Attempting to answer these types of questions is precisely the
motivation for estimating models of Travel Liking.
Table ES. 2: Long- Distance Travel Liking Dependent Variables ( N= 1,358)
Travel Liking Variable Strongly
dislike Dislike Neutral Like Strongly
like
N 19 119 368 671 181
Overall
% 1.4 8.8 27.7 49.4 13.3
Work/ School- N 153 331 576 267 31
Related % 11.3 24.4 42.4 19.7 2.3
Entertain./ Social/ N 23 83 320 597 335
Recreation % 1.7 6.1 23.6 44.0 24.7
N 48 211 420 563 116
Personal Vehicle
% 3.5 15.5 30.9 41.5 8.5
N 54 130 272 632 270
Airplane
% 4.0 9.6 20.0 46.5 19.9
The Travel Liking measures are potentially a function of all the variable categories in the dataset
save Relative Desired Mobility ( which, as an indicator of desired change, we take to be the final
outcome of influences such as Lifestyle, Mobility, and Travel Liking). The general hypothesis
underlying the models is that Travel Liking measures will primarily be a function of the Attitude,
Personality, and Lifestyle variables. In essence, we hypothesize that through intrinsic human
nature and life experiences, individuals develop various degrees of a desire for, and liking of,
travel. Once these attitudes and personalities are developed, they will dominate the liking for
travel. While we expect that the amount of travel an individual engages in from day to day will
play a role in either reinforcing or contradicting existing beliefs, it will not be the key determinant
of Travel Liking. For example, if an individual is forced ( taking work travel to be mandatory) to
commute long distances as a result of her residential and workplace location choices, she may
begin to dislike travel in an automobile. However, we do not expect that this factor will be a
more powerful explanatory variable than the measures of Attitude, Personality, and Lifestyle in
the data set. Similarly, while we certainly expect Socio- Demographic variables to play a role in
xiv
the explanation of Travel Liking, Socio- Demographics are not anticipated to be strong
explanatory variables. For example, while it is expected that those with high incomes may not
like to travel long distances in automobiles, we do not expect this variable to be more important
than Attitude measures. In fact, unlike travel behavior itself ( which is strongly related to socio-demographic
traits), we expect the Liking for at least some types of travel to be relatively
independent of such variables.
A summary of the Travel Liking models is presented in Table ES. 3 and Table ES. 4. A total of 13
models are estimated using ordinary least- squares regression – eight for short- distance travel
and five for long- distance travel. The short- distance models include the following categories of
travel: overall, commute to work/ school, work/ school- related, entertainment/ recreation/ social,
personal vehicle, bus, rail, and walk/ jog/ bicycle; the long- distance models include: overall,
work/ school- related, entertainment/ recreation/ social, personal vehicle, and airplane. The
adjusted R2 values range from 0.346 to 0.106, which, while not low for disaggregate travel
models, indicates the difficulty in explaining a variable that measures individuals’ liking.
Interestingly, the model with the highest R2 value is the model of commute Travel Liking and,
importantly, little of this model’s explanatory power is found in objective measures of commute
distance and time ( meaning, commute dislike is not simply due to having a long commute, or
conversely).
For the most part, the model results confirm our primary hypothesis in that the majority of
explanatory power in the models is provided by the Attitude, Personality, and Lifestyle variables
( see the tables in Section 3 for detailed model estimates). In fact, all six of the Attitude factor
score variables were significant in at least one of the models, as were all four of the Lifestyle
factor score variables and three of four Personality factor scores.
This report directly addresses the positive utility of travel recently articulated by Salomon and
Mokhtarian ( 1998) and Mokhtarian and Salomon ( 2001), among others. Salomon and
Mokhtarian ( 1998, p. 136) hypothesized that in “ some people and in some contexts, travel for its
own sake is valued due to one or more … character traits or desires”; they went on to list a
number of specific traits/ desires. In Table ES. 5 we compare these hypothesized traits/ desires,
along with two other traits ( curiosity and escape/ therapy) not included in the 1998 paper, with
the results from the Travel Liking models presented in this report. The table indicates generally
strong support for all originally hypothesized traits ( note that several variables in the models
relate to more than one trait). Although, after all, the survey was designed specifically to capture
xv
a number of these traits, it is noteworthy that Travel Liking arises from such a variety of sources.
The most important positively associated factors appear to be status, independence, curiosity
and variety- seeking, and the escape/ therapeutic benefits of traveling, as well as a craving for
transition time between work and home and the synergy effects of trip chaining. The most
important negatively associated variables were travel dislike and travel stress. These factors
represent reasons why travel is generally expected to be a disutility, but viewed in the opposite
way, it can be said that a positive Travel Liking is partly defined by a person’s refusal to see
travel as boring, stressful, unsafe, and so on.
The general conclusion from the modeling is that attitudes and personality are important factors
in describing travel behavior. The previous single- equation models of Objective Mobility,
Subjective Mobility and Relative Desired Mobility indicated Travel Liking measures to be key
explanatory variables. Here, Travel Liking is shown to be primarily a function of Attitude,
Lifestyle, and Personality variables. Just as previous research suggested that attitudes are an
important factor in mode choice ( those who do not like public transit, for whatever reason,
probably will not choose a transit mode even if it offers better service than an automobile), the
research presented here, along with the companion Objective Mobility, Subjective Mobility, and
Relative Desired Mobility reports, suggests that attitudes towards travel itself, along with
inherent differences in personality and lifestyle, can influence the amount of travel in which an
individual engages, or wishes to engage. Such results have important policy implications as
they offer increased insight into why not all individuals will react similarly when faced with travel-reducing
policies, such as the promotion of telecommuting alternatives.
xvi
Table ES. 3: Summary of Short- Distance Travel Liking Models
Explanatory variables Dependent variable ( adjusted R- squared): Travel Liking for…
Category Variable
Overall
( 0.214)
N= 1321
Cmt.
( 0.346)
N= 1339
Work/
Sch- rel
( 0.143)
N= 1351
Ent/ Rec
( 0.118)
N= 1327
Pers
veh
( 0.182)
N= 1344
Bus
( 0.170)
N= 1319
Rail
( 0.182)
N= 1295
Walk,
etc.
( 0.196)
N= 1299
Weekly commuting distance ( miles) [ 0,800] - -
Weekly total SD travel ( miles) [ 5,1500] -
Commute mode dummy – bus or ferry [ 0,1] -
Commute mode dummy – rail [ 0,1] +
One- way commute time ( minutes) [ 2,130] -
One- way commute distance ( miles) [ 0,…, 108] + +
Weekly travel by other means ( miles) [ 0,600] -
Objective
Mobility
Past year ( log) total long distance miles [ 0,12.8]* -
Subj. Mob. Overall short distance travel [ 1,…, 5] -
Travel dislike factor score [- 1.8,3.7] - - - -
Travel stress factor score [- 1.9,2.9] -
Commute benefit factor score [- 2.9,2.6] + + + + + +
Travel freedom factor score [- 3.0,2.3] + + + +
Pro- environ. solutions factor score [- 2.3,2.4] - + + +
Attitude
Pro- high density factor score [- 2.5,2.3] - - + +
Family/ com- related factor score [- 3.9,2.1] - - + + +
Status seeker factor score [- 1.7,2.7] + + + +
Workaholic factor score [- 2.1,2.7] +
Lifestyle
Frustrated factor score [- 2.0,2.7] -
Organizer factor score [- 2.90,2.6] - -
Personality Calm factor score [- 2.9,2.4] + + +
How often do you travel … just to relax +
… to clear your head +
… to explore new places + +
… when you need time to think +
… by a longer route to exp. more of your srndgs. + +
Excess
Travel
[ 1,2,3]
... mainly to be alone + +
Conditions which prevent or limit air travel +
Conditions w Mobility hich prevent or limit public transit - -
Limit. [ 1,2,3]
Conditions which prevent or limit bicycle + +
Luxury vehicle type dummy [ 0,1] -
Minivan vehicle type dummy [ 0,1] -
Suburban dummy [ 0,1] +
Concord dummy [ 0,1] -
Sales occupation dummy [ 0,1] +
Professional occupation dummy [ 0,1] -
Personal income category [ 1,…, 6] - - -
Number of persons age 6- 15 in HH [ 0,…, 3] +
Number of persons age 24- 40 in HH [ 0,…, 7] - +
Number of persons age 41- 64 in HH [ 0,…, 3] +
Number of persons age 65- 74 in HH [ 0,1,2] -
Number of persons in HH [ 1,…, 8] + +
Single adult with children family status dmy [ 0,1] +
Female [ 0,1] + -
Socio-
Demo-graphic
Educational background [ 1,…, 6] + +
Notes: [ ] represents variable range; HH = household; SD = short distance; * Logarithm ( miles + 1) to avoid taking the log of zero
xvii
Table ES. 4: Summary of Long- Distance Travel Liking Models
Explanatory variables Dependent variable [ adjusted R- squared]: Travel Liking
for…
Category Variable Overalli
[ 0.206]
Work
relatedj
[ 0.106]
Ent. /
soc. /
rec. k
[ 0.183]
Personal
vehicle
[ 0.178] l
Airplanem
[ 0.149]
Obj. Mobility Past year work- related long- distance trips [ 0,230] -
Subjective Long- distance work/ school- related travel [ 1,…, 5] -
Mobility Long- distance airplane travel [ 1,…, 5] -
Travel dislike factor score [- 1.8,3.7] - - - - -
Travel stress factor score [- 1.9,2.9] - - - - -
Commute benefit factor score [- 2.9,2.6] +
Attitude
Pro- high density factor score [- 2.5,2.3] + -
Family/ community- related factor score [- 3.9,2.1] + +
Lifestyle Status seeker factor score [- 1.7,2.7] + + +
Workaholic factor score [- 2.1,2.7] +
Personality Adventure seeker factor score [- 2.6,2.7] +
… to explore new places + +
Excess Travel … when you need time to think +
[ 1,2,3]
… out of your way to see beautiful scenery +
Mobility Limit. Conditions which prevent or limit air travel [ 1,2,3] -
Number of full- time workers in HH [ 0,…, 6] -
Management/ administrator occupation dummy [ 0,1] +
Production- construction- crafts occupation dummy [ 0,1] -
Personal income category [ 1,…, 6] -
Number of persons age 24- 40 in HH [ 0,…, 7] +
Number of persons age 41- 64 in HH [ 0,…, 3] +
Two or more adults with children family status dummy [ 0,1] +
Single adult without children family status dummy [ 0,1] +
Socio-
Demographic
Educational background [ 1,…, 6] +
Notes: [ ] represents variable range; N = 1345i, 1356j, 1351k, 1318l, 1354m; HH = household
xviii
Table ES. 5: Comparison of Hypotheses and Travel Liking Model Results
Hypothesized trait or
desire
Evidence
in TL
Models?
Travel Liking Model( s)
Explanatory
Variable
Category
Explanatory Variable
LD Airplane Personality Adventure- seeking
Adventure- or variety- factor score
seeking Yes
SD Entertainment, SD Walk, LD
Overall, LD Entertainment Excess Travel
How often do you
travel to explore new
places?
Independence Yes
SD Overall, SD Commute, SD
Work/ School- Related, SD
Personal vehicle
Attitude Travel freedom factor
score
Control Somewhat SD Bus, SD Rail Personality Organizer factor score
( negative direction)
SD Overall, SD Work/ School-
Related, SD Entertainment, SD
Personal vehicle, LD Overall, LD
Work- related, LD Personal
vehicle
Lifestyle Status seeker factor
score
Status Yes
SD Rail Socio-
Demographics
Luxury vehicle type
( negative direction)
Buffer Yes
SD Overall, SD Commute, SD
Work/ School- Related, SD
Personal vehicle, SD Bus, SD
Rail, LD Personal vehicle
Attitude Commute benefit
factor score
Exposure to the
environment Yes SD Personal vehicle, SD Walk Excess Travel
How often do you
travel by a longer
route to experience
more of your
surroundings?
SD Personal vehicle, SD Walk Excess Travel
… by a longer route to
experience more of
Scenery or other your surroundings?
amenities Yes
LD Personal vehicle Excess Travel
… out of your way to
see beautiful
scenery?
Synergy
( multiple activities) Yes
SD Overall, SD Commute, SD
Work/ School- Related, SD
Personal vehicle, SD Bus, SD
Rail, LD Personal vehicle
Attitude Commute benefit
factor score
SD Entertainment, SD Walk, LD
Overall, LD Entertainment Excess Travel
How often do you
travel to explore new
Curiosity Yes places?
SD Personal vehicle, SD Walk Excess Travel
… by a longer route to
experience more of
your surroundings?
SD Entertainment, LD
Entertainment Excess Travel … when you need
time to think?
SD Commute, SD Work/ School-
Related Excess Travel … mainly to be
alone?
SD Overall Excess Travel … just to relax?
SD Walk Excess Travel … to clear your head?
SD Work/ School- Related Socio-
Demographic
Number of persons
age 6- 15 in household
Escape/ Therapy Yes
LD Work- related Socio-
Demographic
Two or more adults
with children family
status
Notes: SD = Short- distance, LD = Long- distance, TL = Travel Liking, Walk = walk/ jog/ bicycle, Entertainment =
entertainment/ recreation/ social
1
1. INTRODUCTION
1.1 Background
This report is part of a larger research program investigating the relationships among attitudes,
personality, and travel. A key premise of the entire research effort is that while individuals travel
primarily to participate in spatially- separated activities, there is an additional component driving
some travel. We believe individuals have a positive utility both for travel itself ( e. g. the sensation
of motion and movement through space which travel provides) and for activities that can be
conducted while traveling ( e. g. listening to music, talking on the telephone). The primary goals
of the research effort as a whole are to better understand both the causes/ motivations of this
affinity for travel, and its effects on travel and related indicators. Prior reports and papers
produced by this study ( see the list provided in the front of this report) have investigated effects
of an affinity for travel by including explicit measures of Travel Liking ( among other variables) in
models of Objective Mobility ( the amount people actually travel), Subjective Mobility ( peoples’
qualitative perception of the amount they travel), and Relative Desired Mobility ( qualitative
measures of how much people want to travel relative to their current amounts). The Liking for
travel has been an important influence in most of those models. Given that importance, it
becomes critical to better understand this affinity for travel: What kinds of people have it, under
what circumstances? This report directly examines the causes of individuals’ liking for travel by
using ordinary least- squares regression to model the relationship between Travel Liking and
other variables in our data set.
Thus, while an exploration of individuals’ pure affinity for travel is interesting in its own right, this
investigation fits into a broader context. Figure 1.1 presents a conceptual model of an
individual’s affinity for travel as modified by a collection of exogenous variables and four key
endogenous variables ( shown in bold type face). Each box denotes a category of variables,
which is operationalized through a number of different specific measures. One end goal for the
( larger) research program is to develop a structural equations model, which will represent the
directional relationships between the endogenous variables identified. At this point in the study,
each key endogenous variable in Figure 1.1 ( namely Objective Mobility, Subjective Mobility,
Travel Liking, and Relative Desired Mobility) is being examined individually.
Examining the conceptual model, it is noticed that the Mobility Constraints variables ( mental or
physical limitations on individuals’ ability to fly, walk, bicycle, ride public transit, drive, drive at
2
night, or drive on a freeway) are the only completely exogenous set of variables in the model.
While some Socio- Demographic variables are clearly exogenous ( i. e. age, gender), others could
be influenced by any number of variables ( e. g. residential location may be a function of
attitudes). Similarly, Attitudes may be intrinsic or influenced by life- stage or other Socio-
Demographics; Personality may also be intrinsic, but ( at least as we have operationalized it in
this study) could be related to income or gender, for example, as well.
The four key endogenous variables – Travel Liking, Subjective Mobility, Objective Mobility, and
Relative Desired Mobility – have been identified as interesting and important measures of travel
behavior. Objective Mobility comprises common measures used by regional planning
organizations in modeling exercises that have the typical end goal of predicting daily travel
amounts. Subjective Mobility is of great interest because two individuals who travel the same
objective distance may not consider their amounts of travel to be the same – as such, they may
respond differently to travel- reducing policies. Relative Desired Mobility is a sort of end- outcome
to all the other variables, resulting in a desire to travel more or less than the current amount. All
these variables are related to the measure discussed in this report – Travel Liking – whose
investigation aims to answer such questions as: What type of person enjoys traveling? Do some
people actually enjoy their daily commute? If so, what kinds of people are they? Are they more
likely to be driving a car? Are they more likely to be wealthy? Understanding what types of
individuals enjoy or don’t enjoy travel could have substantial policy implications. Any policy
aimed at reducing the use of a good or service that a significant segment of the population
“ likes” ( especially if that liking were largely independent of travel amounts) would certainly be
more difficult, and probably less successful, than reducing the use of a uniformly despised good
or service.
In developing the current set of single- equation models of Travel Liking, we do not limit
ourselves to relationships shown in the conceptual model of Figure 1.1 ( which in any case is not
necessarily considered to be final). For example, while we hypothesize that the impact of
Objective Mobility on Travel Liking occurs only through the impact of Objective Mobility on
Subjective Mobility and then Subjective Mobility on Travel Liking, at this stage of the study we
allow Objective Mobility to enter the models directly as well. Such variables, if significant, may
be capturing residual effects due to error in our measurement of Subjective Mobility and/ or in
our specification of the functional form of the relationship, as well as indirect effects of other
variables ( unobserved as well as observed). It should be noted that the single- equation
approach is subject to simultaneity bias due to the inclusion of variables endogenous to the
3
conceptual model as explanatory variables. As such, the model results may be viewed as
suggestive rather than definitive. However, the single- equation models do offer insight into the
individual measures of travel behavior and greatly aid in the end goal of structural equations
modeling.
Figure 1.1: Conceptual Model of an Individual's Affinity for Travel
1.2 What are the Sources for a Liking of Travel?
Why would anyone like to travel? After all, conventional engineering and economic wisdom
holds that the purpose of urban travel is purely to participate in spatially- separated activities. As
General Travel
Attitudes
Personality
& Lifestyle
Objective Mobility
Relative Desired Mobility
Travel Liking
Subjective Mobility
Mobility
Constraints
Socio-
Demographics
Link between key endogenous variables
Link between key endogenous variable and background variable
Link between background variables
4
such, models of travel demand treat time spent in a car or aboard a transit vehicle strictly as a
cost to be minimized – an assumption that ignores the possibility that any portion of urban travel
could provide positive utility. However, the concept of liking travel for its own sake is not entirely
foreign to the profession. For example, there is a sizable literature relating to the so- called “ love
affair with the automobile” ( e. g., Wachs and Crawford, 1992; Marsh and Collett, 1986; Sachs,
1992), which, although perhaps stereotypically associated with Americans, is by no means
unique to them, as attested by studies in Denmark ( Jensen, 1999), the Netherlands ( Steg, et al.,
2001), Scotland ( Hiscock, et al., 2002), and elsewhere, as well as by rising rates of auto
ownership and vehicle- miles traveled throughout the world. Recent psychological studies have
examined the relationship between the opposing desires for personal car use and pro-environmental
behavior, which is increasingly associated with conforming to social norms ( for
recent examples, see Tertoolen, et al., 1998; Nordlund and Garvill, 2003; Tanner, 1999).
Beyond the obvious utilitarian benefits of the automobile ( its often unmatchable convenience
and comfort), these and other studies point out the psychological benefits of automobile use
( e. g. it satisfies the need for self expression and helps demonstrate one’s social position) and
also state that driving a car is simply pleasurable ( e. g. the sensation of movement and control)
( Steg, et al., 2001). The research presented here addresses automobile use as well, but more
broadly investigates all types of travel, including purpose- specific travel, walking and the use of
public transportation, and long- distance travel.
A number of transportation scholars have also commented in a general way on the intrinsic
benefits of travel ( see Mokhtarian, et al., 2001 for citations). With those sources as background,
Salomon and Mokhtarian ( 1998, pp. 136- 137) suggest a number of reasons why travel
( including, but not limited to automobile travel) might have a positive utility:
􀂃 “ adventure- seeking: the quest for novel, exciting, or unusual experiences will in some
cases involve travel as part or all of the experience itself, not just as a means to the
end (‘ getting there is half the fun’);
􀂃 variety- seeking: a more mundane version of the adventure seeking trait, the desire to
vary from a monotonous routine may lead one, for example, to occasionally take a
longer route to work or visit a more distant grocery store;
􀂃 independence: the ability to get around on one’s own is one common manifestation
of this trait;
􀂃 control: this trait is likely to partially explain travel by car when reasonable transit
service is available;
􀂃 status: traveling a lot, traveling to interesting destinations, and traveling ‘ in style’ ( e. g.
in a luxury car) can be symbols of a desired socio- economic class or lifestyle;
5
􀂃 buffer: … a certain amount of travel can provide a valued transition between
activities such as home and work;
􀂃 exposure to the environment: ‘ cabin fever’ is one manifestation of this desire, to
leave an enclosed building and ‘ go somewhere’, just to experience something of the
outdoors;
􀂃 scenery and other amenities: may lead someone, for example, to take a longer route
than necessary to a destination;
􀂃 synergy: the ability to conduct multiple activities at or on the way to a more distant
destination, or the ability to be productive while traveling, may result in apparently
excess travel.”
To this list, we would add:
􀂃 escape: using travel to, for example, temporarily escape family obligations and/ or
domestic tensions;
􀂃 curiosity: certainly curiosity drives, to a certain extent, the adventure- seeking and
variety- seeking mentioned above, but may not be limited to these two traits ( see,
e. g., Stagl, 1995); individuals may be curious about who may be taking the bus with
them on a given day;
􀂃 physical exercise: although most naturally associated with non- motorized forms of
travel such as walking, jogging, or bicycling, even the use of motorized modes
requires a modicum of physical effort, beyond, e. g., sitting and watching television
( walking to/ from, getting into/ out of the vehicle; see, e. g., Mackett, et al., 2004). A
desire for exercise may lead one to engage in “ undirected” ( recreational) trips by
non- motorized means, to choose a slower non- motorized mode over a faster
motorized one, to park ( or alight from transit) farther from the destination than
necessary, or to make a trip when it could be foregone ( e. g. substituted by
telecommunications technology, as in telecommuting versus commuting); and,
closely related,
􀂃 the therapeutic value of movement/ travel: this dimension contains a number of
aspects, including some already touched upon: the sensation of movement can have
a soothing or ( e. g., at high speeds) stimulative quality; fields such as yoga and dance
therapy ( Stanton- Jones, 1992) attest to both the physical and psychological benefits
of movement; movement on a larger scale, i. e. travel, has been advised as mental
therapy at least since Burton’s ( 1621) Anatomy of Melancholy ( see his Part II,
Section II, Movement IV). The need to escape can also fall under this category when
it represents a healthy response to stress, but we leave it separate since it can also
constitute an unhealthy abdication of responsibility.
The exploration undertaken here attempts to identify which factors ( if any at all) among those
available to us most strongly explain the stated Travel Liking, as captured by our survey
instrument. The report concludes with a comparison between the above hypothesized factors
and the model estimation results.
6
1.3 Data
The data analyzed in this study are collected from a fourteen- page self- administered survey of
approximately 2,000 individuals in the San Francisco Bay Area. A total of 8,000 surveys were
mailed ( leading to a response rate of about 25%) to randomly- selected households in three
neighborhoods, namely: North San Francisco ( half of the surveys), Concord ( one- quarter) and
Pleasant Hill ( one- quarter). North San Francisco is an urban neighborhood, located close to the
regional central business district ( CBD) and well- served by transit. Concord and Pleasant Hill, in
contrast, are both suburban cities, located across the San Francisco Bay from the regional
CBD. This report focuses on a subset of the 2,000 respondents – those who work either part-time
or full- time and commute at least once a month. This subset contains 1,358 respondents
with relatively complete data on most variables of interest; some key Socio- Demographic
characteristics of the sample are shown in Table 1.1. The decision to only consider commuters
was based on the assumption ( supported by a few tests) that relationships among Attitudes,
Personality, and Mobility variables could be rather different for commuters than for non-commuters.
Table 1.1 indicates that our sample is relatively balanced in terms of gender and neighborhood
location. The youngest and oldest age categories have few observations, but as the sample
comprises full- and part- time workers, this is not surprising. Higher incomes are over-represented
compared to the Census ( see Curry, 2000 for further discussion). However, as the
focus of the work is to model the impact of income and other variables on Travel Liking
measures, rather than purely to ascertain the population distribution of such measures, it is
more important simply to have a reasonable spread of incomes than that they be exactly
representative ( Babbie, 1998).
The organization of the rest of this report is as follows. The following section describes in more
detail each of the variable categories considered in the modeling. Section 3 presents models of
Travel Liking for different types of short- and long- distance travel. The final section summarizes
the results and puts forth suggestions for further research.
7
Table 1.1: Socio- Demographic Characteristics of Sample ( N= 1,358)
Characteristic Number ( percent)
Concord 318 ( 23.4)
Pleasant Hill 369 ( 27.2)
North San Francisco 671 ( 49.4)
Femalea 692 ( 51.1)
Have a driver’s licenseb 1,338 ( 98.7)
Work full- time 1,141 ( 84.0)
Personal incomec < $ 15,000 31 ( 2.3)
$ 15,000 – 34,999 141 ( 10.6)
$ 35,000 – 54,999 269 ( 20.3)
$ 55,000 – 74,999 250 ( 18.9)
$ 75,000 – 94,999 220 ( 16.6)
> $ 95,000 411 ( 31.1)
Aged 18 – 23 44 ( 3.2)
24 – 40 584 ( 43.0)
41 – 64 686 ( 50.5)
> 65 43 ( 3.2)
Characteristic Mean ( std. dev.)
Total people in household 2.39 ( 1.22)
Total children under 18 in HHe 0.45 ( 0.84)
Total workers in HH ( full/ part- time) f 1.77 ( 0.80)
Number of personal vehicles in HHg 1.87 ( 1.08)
Total short distance travel ( miles/ week) d 219.46 ( 188.67)
a N= 1,352; b N= 1,356; c N= 1,322; d N= 1,357; e N= 1,351; f N= 1,354; g N= 1,353
2. VARIABLES
2.1 The Dependent Variables: Travel Liking
The Travel Liking dependent variables were drawn directly from the survey via the question:
“ How do you feel about traveling in each of the following categories? We are not asking how
you feel about the activity at the destination, but about the travel required to get there. Even if
you seldom or never travel in a certain category, you may still have a feeling about it.”
Respondents then rated their liking for travel in various categories on a five- point ordinal scale
anchored by “ strongly dislike” and “ strongly like”. In addition to distinguishing Travel Liking by
trip purpose and mode, these measures were further disaggregated into short- distance and
long- distance. In keeping with the definition formerly used by the American Travel Survey, long-distance
travel includes trips with a one- way distance of 100 miles or more. A summary of the
8
responses to the short- distance Travel Liking questions is presented in Table 2.1 and the long-distance
responses are presented in Table 2.2
Turning first to the short- distance Travel Liking measures, the raw distributions in Table 2.1
certainly seem to support the contention that a subset of individuals has an affinity for travel.
Even the stereotypically loathed daily commute is liked or strongly liked by more than a fifth of
the sample ( 21.4 percent), with a similar proportion ( 18.6 percent) liking or strongly liking
work/ school- related travel. In fact, only three of the eight categories ( those two plus bus) have a
smaller share of “ likers” ( those in the strongly like and like categories) than “ dislikers” ( those in
the strongly dislike and dislike categories).
Table 2.1: Short- Distance Travel Liking Dependent Variables ( N= 1,358)
Travel Liking Variable Strongly
dislike Dislike Neutral Like Strongly
like
N 15 178 762 360 43
Overall
% 1.1 13.1 56.1 26.5 3.2
N 123 424 520 254 37
Commute
% 9.1 31.2 38.3 18.7 2.7
N 64 292 749 227 26
Work/ School- Related
% 4.7 21.5 55.2 16.7 1.9
N 6 66 543 605 138
Purpose
Entertain./ Social/
Recreation % 0.4 4.9 40.0 44.6 10.2
N 34 125 410 647 142
Personal Vehicle
% 2.5 9.2 30.2 47.6 10.5
N 389 473 384 103 9
Bus
% 28.6 34.8 28.3 7.6 0.7
N 161 231 540 384 42
Rail
% 11.9 17.0 39.8 28.3 3.1
N 54 66 332 663 243
Mode
Walk/ Jog/ Bicycle
% 4.0 4.9 24.4 48.8 17.9
Looking more closely at the purpose- specific categories, by far the most liked category of travel
is entertainment/ recreation/ social – viewed favorably by more than half ( 54.8 percent) of the
respondents. Of course, individuals liking leisure travel is not surprising; in addition to being
influenced by the anticipated enjoyment at the destination, this type of travel often occurs with
family or friends and is probably done with fewer time constraints ( and less stress) than
mandatory travel.
9
With respect to the mode- specific measures, surprisingly, travel by personal vehicle has an
even higher share of “ likers” ( 58.1 percent) than entertainment/ recreation/ social travel. In fact,
among the short- distance categories, only travel by non- motorized modes ( walking, jogging, and
bicycling) is more beloved by survey respondents ( 66.7 percent). In line with stereotype, rail
modes are viewed much more fondly than bus modes. Rail likers and dislikers each comprise
about 30 percent of the sample, whereas bus dislikers outnumber likers nearly 8 to 1 ( 63.4
percent to 8.3 percent).
The responses to the long- distance Travel Liking questions are summarized in Table 2.2. Here,
entertainment/ recreation/ social travel is enjoyed by a substantial majority of the sample ( 68.7%),
as are overall ( 62.7%) and airplane travel ( 66.4%). Exactly half of the sample reports liking long-distance
personal vehicle travel, though nearly a third ( 30.9%) feel neutral about it. The sizable
amount of neutrality ( 42.4%) with respect to work/ school- related long- distance travel may reflect
both a balancing of pros and cons for this category and ( for some) a relative lack of engagement
in it.
Table 2.2: Long- Distance Travel Liking Dependent Variables ( N= 1,358)
Travel Liking Variable Strongly
dislike Dislike Neutral Like Strongly
like
N 19 119 368 671 181
Overall
% 1.4 8.8 27.7 49.4 13.3
N 153 331 576 267 31
Work/ School- Related
% 11.3 24.4 42.4 19.7 2.3
N 23 83 320 597 335
Purpose
Entertain./ Social/
Recreation % 1.7 6.1 23.6 44.0 24.7
N 48 211 420 563 116
Personal Vehicle
% 3.5 15.5 30.9 41.5 8.5
N 54 130 272 632 270
Mode
Airplane
% 4.0 9.6 20.0 46.5 19.9
Since, for the most part, these responses vary in expected ways, a first reaction to the results
may be that the respondents, even with the explicit survey instructions that emphasized
consideration of the trip or travel rather than the activity at the end of the trip, confounded, to
some degree, their liking for the activity with their liking for travel. As discussed in Mokhtarian
and Salomon ( 2001), someone who reports a love for recreation travel may not be referring to
the hours spent in the airport, on the airplane, and in a rental car. One may wonder how
accurately the survey measured a liking for the actual travel.
10
In response to this justifiable concern, a number of considerations are relevant. We first discuss
the potential for confusion between travel and the destination activities. Next, we present less
obvious interactions between trip characteristics and travel, and explain how each may
influence the Travel Liking results.
First, suppose that in the worst case the responses were entirely about the destination activity
and not at all about the travel. They still have travel implications. Although the activities ( work,
entertainment, etc.) captured by these variables have in- home alternatives, it is well understood
that those alternatives are often inferior to their out- of- home counterparts on a number of
dimensions. To the extent that that is the case, the simple descriptive data shown in Table 2.1
and Table 2.2 point to a substantial level of current and potential demand for out- of- home
activities and, as follows, the travel required to engage in out- of- home activities.
However, the argument that people confound destination activities with the travel required to
reach them is most compelling for the five categories that relate to travel purposes: short-distance
commute, work/ school- related and entertainment/ recreation/ social; long- distance
work/ school- related and entertainment/ recreation/ social. It is less persuasive ( although not
entirely baseless) to suggest that the six mode- based ratings of travel ( short- distance personal
vehicle, bus, rail, and non- motorized; long- distance personal vehicle and airplane), or the two
overall ratings of travel ( short- and long- distance, each placed first in their respective sections
so that the respondent was reacting first to the “ abstract concept” of travel rather than travel tied
to a particular type of activity or mode), have the same problem. The fact that respondents could
like “ generic” travel is telling.
Further, the variation in the purpose- specific Travel Liking responses may indicate interactions
between travel and purpose, independent of destination. For example, an individual traveling
from Chicago to Miami for business may enjoy the trip itself less than another individual
traveling on the same flight to visit family. The businessman may have anxiety over his
performance at the destination; may be burdened by traveling with ( and needing to work using)
his laptop and cellular phone; or may feel stress due to pre- trip preparations. Without such
preoccupations, the vacationer may be able to enjoy the in- flight movie or do some pleasure
reading. Thus, two individuals traveling on the same flight may experience the travel differently
due to their differences in trip purpose. In these types of interactions, the survey appropriately
captures purpose- specific variation in the Liking for travel.
11
Interactions also exist between travel and the route or destination, rather than the activity at the
destination per se. One may dislike congested travel, and local commute trips are often
congested, so one expresses a dislike for commute travel. Or, an individual traveling to work via
a bus route that overlooks the San Francisco Bay may express a liking for commute travel,
when the motivation for the liking is really the scenic beauty. In either case, individuals are again
responding to differences in the travel itself, that happen to be associated with certain trip
purposes more than others. This is consistent with the findings of Anable and Gatersleben
( 2004), that both car and public transport trips were viewed with more positive emotion when
they were undertaken for leisure purposes than for commuting purposes.
The latter two types of interactions constitute legitimate variations in the quality of the travel
experience ( leading to legitimate variations in the Travel Liking measure); only the first form of
response ( complete mental “ substitution” of the travel for the activity, and responding to the
activity instead of the travel) constitutes the spurious confounding that we are concerned about.
Of course, the conceptual considerations presented in the Introduction and at greater length in
the references cited there provide a number of reasons why travel itself could have positive
utility. Thus, the concept is not prima facie untenable; the question is not whether people can
possibly like travel for its own sake, but only the degree to which they do. Overall then, we
believe that, although imperfect, these responses are telling us something valid about the Liking
for travel itself. Nevertheless, as we discuss further in the Summary and Discussion section, it is
important to refine these measures in future work.
2.2 The Potential Explanatory Variables
The potential explanatory variables used in the models can be placed into nine general
categories, namely: Objective Mobility, Subjective Mobility, Relative Desired Mobility, Attitudes,
Personality, Lifestyle, Excess Travel, Mobility Constraints, and Socio- Demographics. Each
category is described very generally in this section. Variables included in the models will be
given more discussion in Section 3 and descriptive statistics ( for only those variables that are
significant in at least one of the models) are included in the Appendix.
The survey questions capturing Objective Mobility, Subjective Mobility, and Relative Desired
Mobility had structures similar to those for Travel Liking. In each section, the measures were
obtained for overall travel, travel segmented by purpose, and travel segmented by mode for
both short- and long- distance ( greater than 100 miles one way) trips. The short- distance trip
12
purposes selected for inclusion in the survey are as follows: commute, work/ school- related,
grocery shopping, to eat a meal, for entertainment/ recreation/ social activities, and chauffeuring
( taking others where they need to go). The short- distance travel modes are the following:
personal vehicle, bus, commuter train/ heavy rail/ light rail, and walking/ jogging/ bicycling. The
long- distance trip purposes are work/ school- related and entertainment/ recreation/ social
activities; the modes are personal vehicle and airplane.
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”. Respondents were also asked to specify how many miles they traveled
each week, in total and by mode and purpose.
The long- distance Objective Mobility variables come from a section of the survey in which
respondents were asked how often they traveled to various parts of the globe “ last year”, by
purpose ( for entertainment and work/ school- related activities) and mode ( personal vehicle,
airplane and other) combinations, with an “ other” category to catch any remaining travel. These
responses indicated number of trips directly, and were also converted into approximate
distances by measuring from a central position in the Bay Area to a central location within the
destination region.
Trips were combined across world regions to obtain three different measures of distance:
􀂃 Total miles, the simple sum of the estimated miles for each reported trip;
􀂃 Log of miles, the natural logarithm of one plus the total number of miles. One mile was
added to each total so that when zero miles were actually traveled in a given category,
the log transformation would return the value zero (= ln( 1)) rather than -∞ (= ln( 0));
􀂃 Sum of the log- miles, obtained by taking the natural logarithm of one plus the number of
miles of each trip in the category separately, and summing across all trips in the
category.
13
The log transformations represent a hypothesized diminishing marginal influence of trip length
on another variable of interest. The third measure listed above differs from the second by
incorporating the number of trips as well as total distance traveled into the measure ( the same
number of total miles will have a larger sum of log- miles value if it is divided among several trips
than if it constitutes only a single trip).
Discriminating each of these variables by travel mode ( personal vehicle, airplane, and other
means), plus retaining the original “ total” variables, yielded a set of 12 measures of distance that
were used in the models.
Subjective Mobility
Here we ask respondents for a subjective assessment of their travel. Again segmenting travel
by mode, trip purpose, and trip length ( short and long), respondents rated their amount of travel
on a five- point semantic- differential scale anchored by “ none” and “ a lot”.
Relative Desired Mobility
These questions focused on how much travel individuals wish to undertake, compared to their
current levels. Again, a five- point scale, here anchored by “ much less” and “ much more”, was
used, and travel was segmented in a manner similar to Objective Mobility, Subjective Mobility,
and Travel Liking.
Attitudes
Attitudes towards travel, land use, and the environment were captured using responses on a
five- point Likert- type scale, to 32 statements. Through factor analysis ( see Redmond, 2000 or
Mokhtarian, et al., 2001 for details of the factor analyses on these as well as the Personality and
Lifestyle variables), the statements were distilled into six basic dimensions, namely: travel
dislike, pro- environmental solutions, commute benefit, travel freedom, travel stress, and pro-high
density. Table 2.3 presents a pattern matrix indicating the strength of the association of
each of the survey statements with each of the Attitude factors. The closer in magnitude a
pattern matrix loading is to 1.0, the more strongly a given statement is associated with the
corresponding factor. A score for each individual on each factor can be computed from these
14
Table 2.3: Pattern Matrix for Attitude Factors ( commuters only, N= 1,427)
Factor label
Variable Travel
dislike
Pro-environment
Commute
benefit
Travel
freedom
Pro- high
density
Travel
stress
Traveling is boring. 0.621
I like exploring new places. - 0.537
The only good thing about traveling is arriving at your destination. 0.525
Getting there is half the fun. - 0.465
To improve air quality, I am willing to pay a little more to use an electric or other
clean- fuel vehicle. 0.641
We should raise the price of gasoline to reduce congestion and air pollution. 0.617
We need more public transportation, even if taxes have to pay for a lot of the costs. 0.612
We can find cost- effective technological solutions to the problem of air pollution. 0.353
I limit my auto travel to help improve congestion and air quality. 0.372
We need more highways, even if taxes have to pay for a lot of the costs. - 0.194
My commute is a real hassle. - 0.695
My commute trip is a useful transition between home and work. 0.583
The traveling that I need to do interferes with doing other things I like. - 0.530
I use my commute time productively. 0.467
Travel time is generally wasted time. 0.379 - 0.461
Getting stuck in traffic doesn’t bother me too much. 0.419
In terms of local travel – I have the freedom to go anywhere I want to. 0.511
In terms of long- distance travel – I have the freedom to go anywhere I want to . 0.422
The vehicles I travel in are comfortable. 0.295
15
Factor label
Variable Travel
dislike
Pro-environment
Commute
benefit
Travel
freedom
Pro- high
density
Travel
stress
It is nice to be able to do errands on the way to and from work. 0.269
I am willing to pay a toll to travel on an uncongested road. 0.212
Living in a multiple family unit wouldn’t give me enough privacy. - 0.617
I like living in a neighborhood where there is a lot going on. 0.486
Having shops and services within walking distance of my home is important to me. 0.243 0.401
I like having a large yard at my home. - 0.323
I worry about my safety when I travel. 0.544
Traveling makes me nervous. 0.201 0.537
Traveling is generally tiring for me. 0.266 - 0.225 0.410
I’d rather have someone else do the driving. 0.227 0.329
I tend to get sick when traveling. 0.318
I am uncomfortable being around people I don’t know when I travel. 0.297
I like traveling alone. - 0.194
Source: Redmond ( 2000). Note: For ease of interpretation, only loadings higher than about 0.200 in magnitude are shown.
16
loadings; it is those factor scores that were included as potential explanatory variables in the
models.
Personality
Respondents rated 17 attributes on a five- point scale ( anchored by “ hardly at all” to “ almost
completely”) in terms of how well the attributes described them. Here, the factor analysis
revealed four personality types: adventure- seeker, organizer, loner, and the calm personality.
Three of these personality types proved significant in the Travel Liking models – calm,
adventure- seeker, and organizer. The pattern matrix is presented in Table 2.4.
Lifestyle
The survey contained 18 statements related to work, family, money, status, and the value of
time. Respondents agreed or disagreed with the statements using a five- point Likert- type scale.
Four lifestyle factors emerged: status seeker, workaholic, family/ community related, and a
frustrated factor. Each of these factors is significant in at least one of the Travel Liking models;
the associated pattern matrix is presented in Table 2.5.
Excess Travel
To qualitatively measure excess travel, participants indicated how often ( on a three- point scale:
“ never/ seldom”, “ sometimes”, “ often”) they engaged in each of 13 activities involving seemingly
unnecessary travel. Questions included, “ how often do you travel…”: “ with no destination in
mind?”, “ just for the fun of it?”, and “ mainly to be alone?”
Mobility Constraints
Here, participants selected, on a three point scale (“ No limitation”, “ Limits how often or how
long”, “ Absolutely prevents”), the degree to which physical conditions or anxieties prevented
them from engaging in a variety of travel forms, including: “ driving on the freeway”, “ driving at
night”, and “ flying in an airplane”. The percentage of time an automobile is available to the
participant is also considered to be a Mobility Constraint ( oriented in the reverse direction).
17
Table 2.4: Pattern Matrix for Personality Factors ( N= 1,904)
Factor label
Variable
Adventure seeking Organizer Loner Calm
Adventurous 0.776
Variety seeking 0.685
Spontaneous 0.574
Risk taking 0.557 - 0.192
Like to stay close to home - 0.435 0.168
Ambitious 0.422 0.330 - 0.217
Like moving at high speeds 0.398 - 0.345
Like being outdoors 0.385
Efficient 0.624
On time 0.371
Like a routine - 0.355 0.364
Like being alone 0.935
Like being independent 0.250 0.301 0.314
Aggressive 0.162 0.312 - 0.599
Patient 0.163 0.532
Restless - 0.389
Like being in charge 0.199 0.363 - 0.380
Source: Redmond ( 2000)
18
Table 2.5: Pattern Matrix for Lifestyle Factors ( N= 1,904)
Factor label
Variable
Frustrated
Family /
community
oriented
Status seeking Workaholic
I often feel I don’t have much control over my life. 0.720
I am generally satisfied with my life. - 0.618
Work and family do not leave me enough time for myself. 0.357 0.262 0.203
I wouldn’t necessarily have to like my work that much, as long as I made enough money. 0.214 - 0.037
I feel that I am wasting time when I have to wait. 0.160 0.156
I’d like to spend more time with my family and friends. 0.585
My family and friends are more important to me than my work. 0.472 - 0.233
I’d like to spend more time on social, environmental, or religious causes. 0.418
Occasionally, I’d be willing to give up a day’s pay to get a day off work. 0.273
To me, the car is a status symbol. 0.698
A lot of the fun of having something nice is showing it off. 0.518
To me, the car is nothing more than a convenient way to get around. - 0.411
The one who dies with the most toys win. 0.410
I’m pretty much a workaholic. 0.652
I’d like to spend more time on work. - 0.164 0.373
I generally try to spend some time each week just on myself. - 0.178
I don’t like to stay in one place for long. 0.171
Source: Redmond ( 2000)
19
Socio- Demographics
The survey captured an extensive amount of typical Socio- Demographic data to allow for
comparison of our sample with more general populations. The data included measures of age,
income, household size, employment type, number of household workers, education level,
gender, and make/ model of the vehicle driven most often by the respondent. The latter variable
was allocated to one of nine major vehicle categories: small, compact, mid- sized, large, luxury,
sport utility vehicle, minivan/ van, pick- up truck, and sports ( for more details, see Curry, 2000).
3. MODELS
3.1 General Specification Issues
A total of 13 linear regression models are developed from the Travel Liking survey responses –
eight models for short- distance travel, specifically: overall, work/ school commute, work/ school-related,
entertainment/ recreation/ social, personal vehicle, bus, rail, and non- motorized ( walk,
jog, and bicycle); and five models for long- distance travel: overall, work/ school- related,
entertainment/ recreation/ social, personal vehicle, and airplane. The ordinal Travel Liking
dependent variables are treated as continuous in this application and the sample includes only
working commuters ( those who work full- or part- time and commute at least once a month).
Though an ordered probit model would be more theoretically appropriate in this context, the
number of models estimated along with the number of potential explanatory variables made the
use of regression, primarily due to the availability of higher quality commercial software
packages ( with automated stepwise specification capabilities), the preferred approach ( for an
ordered probit version of the commute Travel Liking model, please see Ory, et al., 2004).
Due to the variety of variables in the data set, certain a priori decisions as to which variables
could reasonably be expected to influence a Liking for travel had to be made. The variables in
the Relative Desired Mobility category were completely excluded from consideration: we
assume that wanting to travel more than currently is an effect rather than a cause of Travel
Liking. Further, it was assumed that travel itself could cause an individual to dislike travel, that is
that Subjective or Objective Mobility could have a negative impact on Travel Liking, but we
excluded such variables when they appeared with a positive coefficient. Although it is possible
20
that greater Mobility in a certain category could lead to greater Travel Liking ( riding the bus a lot
could generate a fondness for the bus), we consider it more likely that a positive relationship is
indicative of the opposite direction of causality – that is, that higher Travel Liking leads to higher
mobility. Thus, we excluded Mobility variables that initially appeared in the Travel Liking models
with positive signs.
We hypothesize that Travel Liking will be most heavily influenced by the various Personality,
Lifestyle and Attitude variables included in the data set. We believe Travel Liking to be an
intrinsic human characteristic, which is shaped by one’s experiences, and most readily revealed
by the attitudes individuals hold toward travel- related issues. Travel demand researchers have
demonstrated the powerful impacts of attitudes on traveler decisions for more than two decades
( e. g. Dobson et al., 1978; Dumas and Dobson, 1979; Tischer and Phillips, 1979; Kitamura et al.,
1997). The modeling of Travel Liking, already demonstrated to be an important determinant of
objective, perceived and desired travel in the single- equation models of Objective Mobility
( Mokhtarian, et al., 2001), Subjective Mobility ( Collantes and Mokhtarian, 2002) and Relative
Desired Mobility ( Choo et al., forthcoming), allows for a more complete picture of how these
attitudes impact travel.
While the data used to estimate the Travel Liking models included myriad Attitude, Personality,
and Lifestyle variables, these variables do not perfectly capture the relevant intrinsic
characteristics of all individuals. For this reason, a handful of variables included in the models
are intended to represent human characteristics not otherwise captured, as illustrated in Figure
3.1. For example, certain models include the Excess Travel variable “ How often do you travel …
to explore new places.” This question probably better captures a sense of curiosity than any of
the other variables in the data set. As such, it serves as a proxy for the influence of curiosity on
Travel Liking – a very plausible relationship.
It should be noted that certain Excess Travel measures, specifically “ How often do you travel
… just for the fun of it”, and “… to a more distant destination than necessary, partly for the fun of
traveling there”, were not considered as potential explanatory variables. Due to the use of the
word “ fun”, it seems more likely that those who enjoy traveling will engage in this type of Excess
Travel. Again referring to Figure 3.1, it seems the underlying human characteristic these
variables are representing is, in fact, Travel Liking, and that including them in the models would
therefore be conceptually tautological.
21
Figure 3.1: Hypothesized Relationship between Certain Explanatory Variables and Dependent
Variables
In the following sections, each of the models is presented and discussed in detail – first the
short- distance models and then the long- distance models. As many of the models have
estimated coefficients with the same sign for the same variable, to streamline the presentation a
section discussing variables common to several models precedes the detailed discussion of the
individual models.
3.2 Discussion of Variables Appearing in Multiple Short- Distance Models
A summary of all the short- distance models is presented in Table 3.1. The adjusted R2 values
for these models range from 0.118 ( for entertainment/ recreation/ social) to 0.346 ( for
commuting), which are typical- to- high for disaggregate models of travel behavior.
The first interesting result is the expected negative influence of amounts of travel on the Liking
for travel. Those who commute long distances or durations tend to enjoy travel less than those
with shorter commutes. As commute travel constitutes a large portion of total travel, the weekly
commute distance variable, as expected, also influences overall Travel Liking. These results fit
the conventional stereotype of travel as a cost and, for those with large travel amounts, these
costs manifest themselves in stated negative feelings toward travel.
Human Characteristic
( e. g., curiosity, or need to
escape)
Explanatory Variable
( e. g., frequency of travel to
explore new places, or to
be alone)
Travel Liking Dependent
Variable
Modeled Relationship
Hypothesized Actual Relationships
22
Next, we examine those variables that are common to the models of Liking for bus and rail
( commuter rail, light rail, and BART – the Bay Area’s Rapid Transit regional rail system) travel.
Both of these models contain the one- way commute distance measure, which indicates that, in
the San Francisco Bay Area, those with longer commutes are more likely to enjoy transit modes
than those with shorter commutes. It may be that those who spend a substantial amount of time
on transit vehicles are less troubled by initially waiting for the arrival of the vehicle, or may enjoy
avoiding the potentially longer automobile commute, or may simply have more time on the
vehicle to read and/ or relax. Further, the Bay Area has many commuter buses, similar to tour or
Greyhound buses, that offer more comfort than typical city buses for longer trips. Those who
have long commutes but are not able to take transit may be reflecting an expectation that their
commute would be more enjoyable if only they didn’t have to drive in congestion.
Other variables significant in the models of Liking for bus and rail travel are Mobility Limitations
on taking public transit and riding bicycles. Those who are unable to use or are limited in taking
public transit, not surprisingly, have a lower Liking for the modes in question than those with no
physical or psychological limitations. Similarly, those who have difficulty riding or are unable to
ride a bicycle have a higher tendency to enjoy transit ( this can generally be extended to those
who have difficulty with non- motorized modes, as there is a strong correlation – coefficient of
0.503 – between limitations on bicycle use and on walking). This may indicate not only a greater
familiarity with transit on the part of those for whom bike is not an option, but perhaps also that
the unattractiveness of non- motorized modes for these individuals produces a compensating
affection for the alternative modes that are available.
Another variable included in the Liking for bus and rail travel is the organizer Personality factor
score, with a negative coefficient. This is logical since ( based on the variables in our survey that
loaded heavily on this factor, as shown in Table 2.4) organizers are those who like to be
efficient, in charge and on time – traits not traditionally associated with riding transit in the
United States.
One of the most significant variables in many of the models is the commute benefit Attitude
factor score. This variable appears in all but two ( entertainment/ recreation/ social and walk) of
the short- distance Travel Liking models and is often ( based on the beta coefficient) among the
most powerful variables. This result suggests that those who view their commute time as
productive and do not find it to be very stressful ( whether because the commute is, in fact,
objectively not stressful, or because their personality is on the calm side, or because they
23
actively adopt coping mechanisms to improve their productivity and reduce the stress of the
commute) have a higher Liking for different types of travel ( by extension, it could be inferred that
these individuals find not only the commute time, but other kinds of travel time to be productive).
The travel freedom Attitude factor score entered into four of the models. Those who feel as
though they have the ability to go wherever they choose, whenever they choose, tend to like
various types of travel more than those who have less travel freedom. This result is important in
that it reinforces the joy individuals find in mobility and the potential for mobility. Although the
travel freedom factor is not mode- specific, the Attitude it represents is certainly one reason for
the nearly- universal popular appeal of automobiles ( as discussed in the Introduction).
Perhaps the most expected result is the common negative sign on the coefficient for the travel
dislike Attitude factor score variable, which appears in four of the eight short- distance models.
Though measured independently ( see Table 2.3), it is certainly expected that, for example,
those who agree that “ traveling is boring” would also dislike certain types of travel. It is
surprising that the travel dislike variable does not enter more of the models, and is, in fact, often
of less significance than other Attitude, Personality, and Lifestyle measures. For example, in the
model of overall Travel Liking, the travel freedom and commute benefit factor scores also enter
into the model ( with the expected positive signs) and both have more explanatory power ( from
the beta coefficients) than the travel dislike factor score. This result indicates that a general
distaste for travel is not as powerful a determinant of overall short- distance Travel Liking as
finding the commute to be a productive time ( commute benefit) or, to a lesser extent, enjoying
the freedom travel provides ( travel freedom). As we will see in Sections 3.11 to 3.16, this
variable is substantially more influential with respect to long- distance travel.
Also entering four of the models is the status seeker Lifestyle factor score. Daily travel may be
the best opportunity for these individuals to proudly display a key symbol of conspicuous
consumption – a nice automobile. This result is consistent with other studies that have found
that the desire to display one’s status, or social standing, influences car use ( see, e. g. Steg, et
al., 2001; Steg, 2004), as it does here, operating through the Travel Liking variable.
Entering both the rail and walk/ jog/ bicycle mode- specific models is the educational background
variable. Both fit the stereotype of the affluent, well- educated commuter well- served by rail and
favoring it over bus, and using non- motorized travel as a means of exercise. Also fitting with
stereotype ( and the literature referenced in the Introduction) is the positive coefficient on the
24
pro- environmental solutions and pro- high density Attitude variables entering the bus, rail, and
non- motorized Travel Liking models, along with the reverse sign on the same variables’
coefficients in the personal vehicle model.
The calm Personality factor variable also enters multiple models – Liking for work/ school-related,
entertainment/ recreation/ social, and bus travel. Individuals with high scores on this trait
may be more relaxed when they encounter the inevitable stresses of travel, and hence more
inclined to enjoy it.
Finally, a variety of variables in the Excess Travel category enter into many models. Those who
often travel “ mainly to be alone”, and also those having children under 15 years old, tend to
enjoy commuting and work- related travel. These results support the notion, as mentioned in the
Introduction of this paper, that travel offers an opportunity to be alone – to temporarily escape
the stresses of family or work obligations ( Edmonson, 1998; Zitnik, 2004).
Those who engage in Excess Travel “ to explore new places”, following intuition, like to travel for
entertainment/ recreation/ social purposes and also enjoy non- motorized modes – both types of
travel are typically associated with exploration. Interestingly, the Excess Travel variable “ by a
longer route to experience more of your surroundings” appears in both the walk/ jog/ bicycle
model and the personal vehicle model. Although the experience may be more participatory and
up- close for walking, and more observational and arms- length for the personal vehicle mode,
the desire for more information about one’s environment may be similar in both cases ( see, e. g.,
Arentze and Timmermans, 2004).
25
Table 3.1: Summary of Short- Distance Travel Liking Models
Explanatory variables Dependent variable ( adjusted R- squared): Travel Liking for…
Category Variable
Overall
( 0.214)
N= 1321
Cmt.
( 0.346)
N= 1339
Work/
Sch- rel
( 0.143)
N= 1351
Ent/ Rec
( 0.118)
N= 1327
Pers
veh
( 0.182)
N= 1344
Bus
( 0.170)
N= 1319
Rail
( 0.182)
N= 1295
Walk,
etc.
( 0.196)
N= 1299
Weekly commuting distance ( miles) [ 0,800] - -
Weekly total SD travel ( miles) [ 5,1500] -
Commute mode dummy – bus or ferry [ 0,1] -
Commute mode dummy – rail [ 0,1] +
One- way commute time ( minutes) [ 2,130] -
One- way commute distance ( miles) [ 0,…, 108] + +
Weekly travel by other means ( miles) [ 0,600] -
Objective
Mobility
Past year ( log) total long distance miles [ 0,12.8]* -
Subj. Mob. Overall short distance travel [ 1,…, 5] -
Travel dislike factor score [- 1.8,3.7] - - - -
Travel stress factor score [- 1.9,2.9] -
Commute benefit factor score [- 2.9,2.6] + + + + + +
Travel freedom factor score [- 3.0,2.3] + + + +
Pro- environ. solutions factor score [- 2.3,2.4] - + + +
Attitude
Pro- high density factor score [- 2.5,2.3] - - + +
Family/ com- related factor score [- 3.9,2.1] - - + + +
Status seeker factor score [- 1.7,2.7] + + + +
Workaholic factor score [- 2.1,2.7] +
Lifestyle
Frustrated factor score [- 2.0,2.7] -
Organizer factor score [- 2.9,2.6] - -
Personality Calm factor score [- 2.9,2.4] + + +
How often do you travel … just to relax +
… to clear your head +
… to explore new places + +
… when you need time to think +
… by a longer route to exp. more of your srndgs. + +
Excess
Travel
[ 1,2,3]
... mainly to be alone + +
Conditions which prevent or limit air travel +
Conditions w Mobility hich prevent or limit public transit - -
Limit. [ 1,2,3]
Conditions which prevent or limit bicycle + +
Luxury vehicle type dummy [ 0,1] -
Minivan vehicle type dummy [ 0,1] -
Suburban dummy [ 0,1] +
Concord dummy [ 0,1] -
Sales occupation dummy [ 0,1] +
Professional occupation dummy [ 0,1] -
Personal income category [ 1,…, 6] - - -
Number of persons age 6- 15 in HH [ 0,…, 3] +
Number of persons age 24- 40 in HH [ 0,…, 7] - +
Number of persons age 41- 64 in HH [ 0,…, 3] +
Number of persons age 65- 74 in HH [ 0,1,2] -
Number of persons in HH [ 1,…, 8] + +
Single adult with children family status dmy [ 0,1] +
Female [ 0,1] + -
Socio-demo-graphic
Educational background [ 1,…, 6] + +
Notes: [ ] represents variable range; HH = household; SD = short distance; * Logarithm ( miles + 1) to avoid taking the log of zero
26
3.3 Overall Short- Distance Travel
This section discusses the model of Travel Liking for all short- distance travel ( by all modes, for
all trip purposes). As a majority of the short- distance travel for commuters in our sample is
commuting to and from work or school ( Figure 3.21 averages the individual purpose shares, by
mileage, across the sample; Figure 3.32 shows the average individual mode shares, by mileage,
across the sample), it is expected that this model will be highly similar to the model of commute
travel presented in Section 3.4.
Commute
58%
Work/ school- related
10%
Grocery shopping
6%
Eat a meal
6%
Enter./ social/ recr.
14%
Chauffeuring
5%
Other
1%
Figure 3.2: Average Shares of Mileage by Purpose
1 Respondents reported miles traveled in each category “ in a typical week”. The “ other” category was not
explicitly provided, but distance traveled in that category is taken to be the difference between the total
distance traveled in a typical week ( explicitly obtained) and the sum of distances in each of the other cate-gories.
As such, these measures are only approximations, and probably not comparable to shares ob-tained
from a more rigorous diary- based measurement instrument. For example, shares for the provided
purposes are probably overestimated and for “ other” purposes probably underestimated. MTC ( 2001a)
estimates the following shares of mileage by purpose: home- based work, 41.2%; home- based school,
5.3%; home- based social/ recreation, 10.8%; home- based shop/ other, 20.1%; non- home- based, 22.6%.
2 For modes, respondents were provided all five categories ( including “ other”), and asked to ensure that
distance traveled by each mode summed to their total distance traveled “ in a typical week”. Thus, we
expect these responses to be somewhat less biased than the purpose- specific ones, but still dependent
on respondents’ abilities to accurately estimate distances by each mode and aggregate across multiple
trips in a week. MTC ( 2001b) estimates the following shares of trips ( not mileage) as follows: personal
vehicle, 83.9%; transit, 5.6%; non- motorized, 10.5%.
27
Personal vehicle
75%
Bus
8%
Train/ BART/ light rail
7%
Non- motorized
9%
Other means
1%
Figure 3.3: Average Shares of Mileage by Travel Mode
Our a priori expectations entering into the modeling are that, consistent with our primary
hypothesis, measures of Attitudes, Personality, and Lifestyle will be the dominant explanatory
variables. However, due to the influence of commute travel on overall travel, it is expected that
certain Objective Mobility measures will negatively impact Travel Liking. Specifically, it is
expected that those with long commutes will enjoy this type of travel less, all else equal, than
those with shorter commutes.
Table 3.2 summarizes the overall short- distance Travel Liking model estimation results. As
expected, those who are forced ( viewing commute travel as mandatory, in the typical tripartite
segmentation of mandatory, maintenance and discretionary travel) to commute long distances
are less likely to enjoy traveling overall. This result is shown through the negative coefficient on
the weekly miles commuting Objective Mobility variable.
In addition to the Objective Mobility measure, a variety of Socio- Demographic measures, for
which we had no strong expectations, also enter into the model. Those with higher incomes
enjoy traveling less, as do those in professional occupations. It is plausible in both cases that
this relative dislike reflects a higher value of time ( i. e. a greater opportunity cost for, and hence
greater resentment of, time spent traveling). Given that higher incomes are generally associated
with more travel in this sample as elsewhere ( see e. g., Ory, et al, 2004), this result may also
28
partly represent a further Objective Mobility effect. Those living outside of San Francisco in the
Pleasant Hill and Concord neighborhoods ( both considered “ suburbs”) enjoy short- distance
travel more, on average, than those living in San Francisco. Such a result could certainly be
attributed to the greater ease of automobile usage and faster speeds present in the suburbs.
There could also be an endogeneity effect, in that those who like traveling less may be more
inclined to choose a central urban residential location that will reduce the need to travel. As the
number of individuals 41 to 64 in the household increases, the Travel Liking also increases.
Taking this variable to be relatively representative of the respondent’s age, it is likely that
individuals in this age category have less pressing needs at home ( such as young children) and,
over time, have been able to either adopt a more preferential commute, or adapt to the one they
have.
In addition to the Objective Mobility and Socio- Demographic measures, impacts on Travel Liking
are found among the Attitude, Lifestyle and Excess Travel measures. In fact, the Attitude and
Lifestyle variables ( commute benefit, travel freedom, travel dislike, and status seeker) are the
most powerful explanatory variables in the model, as shown by their beta ( standardized
coefficient) values, which supports our primary hypothesis ( see Section 3.2 for further
discussion of these variables).
The Excess Travel variable “ how often do you travel … just to relax” also enters the model with
a positive sign. This measure is probably capturing the relaxing sensation many individuals
obtain from the movement or sense of control found in traveling, representing one reason
individuals may have a positive utility ( and hence a Liking) for travel. As shown in Figure 3.1,
this variable may be serving as a proxy for this difficult- to- define human characteristic.
3.4 Commute to Work/ School
As alluded to in the previous section, the model here considers Liking specifically for commute
( to work or school) travel ( this model is also discussed in Ory et al., 2004). Expectations for this
model mirror those discussed previously for overall short- distance travel – certain measures of
Objective Mobility will enter the model along with the dominant Attitude, Lifestyle, and
Personality measures.
29
Table 3.2: Model of Liking for All Short- Distance Travel ( N= 1,321)
Dependent Variable : Overall liking for short- distance travel [ 1, …, 5]
Explanatory Variables Coefficient t- statistic Beta
Constant 3.278 41.42
Objective Mobility
Weekly commute miles [ 0,800] - 0.000536 - 3.69 - 0.100
Socio- Demographics
Suburban dummy [ 0,1] 0.122 3.11 0.0833
Personal income category [ 1,…, 6] - 0.0427 - 3.11 - 0.0838
Professional occupation dummy [ 0,1] - 0.0664 - 2.26 - 0.0562
Number of persons age 41- 64 in household [ 0,…, 3] 0.0450 1.99 0.0516
Attitudes
Commute benefit factor score [- 2.9,2.6] 0.235 9.89 0.278
Travel freedom factor score [- 3.0,2.3] 0.108 4.12 0.109
Travel dislike factor score [- 1.8,3.7] - 0.0904 - 3.64 - 0.106
Lifestyle
Status seeker factor score [- 1.7,2.7] 0.0918 4.13 0.102
Excess Travel [ 1,2,3]
How often do you travel … just to relax 0.0984 3.15 0.0813
[ ] = range of possible or observed responses
Adjusted R2 = 0.214 ( R2 = 0.220) F- statistic = 36.87 ( p = 0.000)
The results of the commute Travel Liking model estimation are presented in Table 3.3. The
adjusted R2 for this model is 0.346, the highest among the short- distance models of Travel
Liking. While some of the significant variables are similar to those presented in the previous
section, important differences do emerge. Examining first the measures of Objective Mobility, a
more detailed decomposition of the effects of actual commute travel emerges. Rather than
simply the weekly commute distance ( which was the lone Objective Mobility measure in the
overall short- distance Travel Liking model), here multiple commute descriptors are significant,
including weekly commute distance, commute time, and primary commute mode3. Those with
3 To ease the burden on the respondent, we collected data on the distance traveled for each specific
mode and purpose separately, rather than for each mode- purpose combination. The primary commute
mode variable was derived from a set of rules based on those reported travel distances. By comparing
reported weekly miles traveled by each mode to the fraction of weekly miles traveled for commuting, one
of five modes ( personal vehicle/ motorcycle, bus/ ferry, train/ BART/ light rail, walking/ jogging/ bicycling, and
other) was assigned to each individual as a primary commute mode. The assignment was made with
100% confidence for 13.5% ( single- mode users) of the sample of 1,358 commuting workers, with a high
degree of confidence for an additional 55.6% ( those whose miles of travel by a single mode exceeded
30
long commutes, both in terms of distance and time, are more likely to disdain travel, as are
those who commute primarily by bus or ferry – perhaps not their desired mode. Further, how
much short- distance travel individuals perceive themselves to be engaged in overall, operating
through the Subjective Mobility measure, is also ( negatively) important to Travel Liking. This
result follows intuition: individuals who feel as though they are always traveling may be less able
to enjoy their commute than those who travel little beyond the commute, and who, as a result,
may relish such travel.
Table 3.3: Model of Liking for Work/ School Commute Travel ( N= 1,338)
Dependent Variable : Liking for work/ school commute travel [ 1, …, 5]
Explanatory Variables Coefficient t- statistic Beta
Constant 2.936 28.50
Objective Mobility
Weekly commute miles [ 0,800] - 0.000786 - 3.57 - 0.112
One- way commute time ( minutes) [ 2,130] - 0.00412 - 2.83 - 0.0885
Commute mode dummy – bus or ferry [ 0,1] - 0.129 - 2.26 - 0.0524
Subjective Mobility
Overall short distance [ 1,…, 5] - 0.0731 - 3.21 - 0.0763
Socio- Demographics
Number of people in the household [ 1, …, 8] 0.0911 4.93 0.116
Number of persons age 24- 40 in household [ 0,…, 7] - 0.0663 - 2.92 - 0.0678
Attitudes
Commute benefit factor score [- 2.9,2.6] 0.449 16.82 0.409
Travel freedom factor score [- 2.9,2.3] 0.120 4.08 0.0922
Lifestyle
Family/ community related [- 3.9,2.1] - 0.168 - 5.70 - 0.132
Excess Travel [ 1,2,3]
How often do you travel … mainly to be alone 0.122 2.99 0.0678
[ ] = range of possible or observed responses
Adjusted R2 = 0.346 ( R2 = 0.350) F- statistic = 71.63 ( p = 0.000)
half their commute miles traveled, with travel by all other modes for all purposes summing to less than
half the commute miles), and with moderate confidence for the remaining 30.9% ( by identifying the mode
used for the greatest proportion of total weekly distance traveled). We have no way of distinguishing
driving alone from carpooling, so the personal vehicle category includes both cases. For the 1,358
commuting workers analyzed in this study, the shares of the five primary commute modes listed above
are 79.4%, 9.7%, 8.2%, 2.4%, and 0.1%, respectively.
31
Moving to the measures of Attitude, the commute benefit factor score is, not surprisingly, by far
the dominant explanatory variable in the model, with a beta coefficient more than three times
larger ( in magnitude) than that of the next most important variable. The travel freedom factor
score is also important and supports the notion that the Liking for travel is partly based on the
independence it offers. These variables are discussed further in Section 3.2.
The family/ community related Lifestyle measure, which corresponds to positive responses to
such statements as “ My family and friends are more important to me than work” ( see Table 2.5),
is the second- strongest variable in the model and has a negative impact on commute Travel
Liking. This result seems intuitive – the more individuals value time with their families, the less
they enjoy being apart from them while commuting. This result is supported by the inclusion of
the number of persons age 24 to 40 Socio- Demographic variable. Respondents having people
in this age group in the household are likely to be in that age group themselves, and may be
more anxious to arrive home to young families and/ or active social lives.
Seemingly contradictory to these results, the Socio- Demographic measure of overall household
size is positively related to Travel Liking. However, this result is illuminated by the Excess Travel
measure, which shows that commute travel can provide a means of escape – a chance to be
alone. As the household size increases, one’s liking for the solitude offered by commute travel
may also increase.
The Lifestyle, Excess Travel and Socio- Demographic variables together offer a finely nuanced
view of a paradox that is probably experienced by many. Although one’s primary focus may be
family and social activities, many also crave time for themselves – which, in modern society,
may be most readily available in the automobile during the daily commute ( Edmonson, 1998).
Even if commuting by public transportation, one can be alone with one’s thoughts, or otherwise
engaged in solitary activities such as reading or listening to music through headphones. This
result illustrates the valuable role that Attitude and Lifestyle measures play in describing Travel
Liking, which, in turn, impacts Subjective Mobility and Relative Desired Mobility. Such measures
greatly enhance our ability to distinguish the often conflicting behaviors and perceptions relating
to travel in general, and commute travel in particular.
3.5 Short- Distance Work/ School- Related
The model of work/ school- related travel, shown in Table 3.4, begins to solidify the importance of
certain measures ( such as the travel freedom, commute benefit, status seeker, and
32
family/ community- related factor scores) in the Liking for mandatory travel, as they again are
relevant in this model.
The Objective Mobility measure of travel by other means, which enters the model with a
negative coefficient, refers to travel by non- traditional modes, such as airplane, taxi,
rollerblades, and boat. Those who often travel by such unusual modes may not enjoy being
confined to the mundane modes usually associated with work/ school- related travel.
Table 3.4: Model of Liking for Short- Distance Work/ School- Related Travel ( N= 1,351)
Dependent Variable : Liking for short- distance work/ school related travel [ 1, …, 5]
Explanatory Variables Coefficient t- statistic Beta
Constant 2.66 46.79
Objective Mobility
Weekly travel by other means ( miles) [ 0,600] - 0.00271 - 3.04 - 0.0768
Socio- Demographics
Number of persons 6- 15 in household [ 0,…, 3] 0.125 3.64 0.0926
Attitudes
Commute benefit factor score [- 2.9,2.6] 0.255 10.56 0.278
Travel freedom factor score [- 2.9,2.3] 0.101 3.60 0.0929
Lifestyle
Status seeker factor score [- 1.7,2.7] 0.0789 3.00 0.0805
Family/ community- related factor score [- 3.9,2.1] - 0.112 - 3.94 - 0.105
Personality
Calm factor score [- 2.9,2.4] 0.0703 2.63 0.0714
Excess Travel [ 1,2,3]
How often do you travel… mainly to be alone 0.165 4.27 0.110
[ ] = range of possible or observed responses
Adjusted R2 = 0.143 ( R2 = 0.148) F- statistic = 29.14 ( p = 0.000)
As with the commute model, we again see the paradoxically positive as well as negative impact
of family on Travel Liking, in the negative coefficient of the family/ community- related Lifestyle
factor, and the positive coefficient for the number of older ( age 6 to 15) children in the
household and the frequency of traveling mainly to be alone.
The remaining “ new” variable in this model is the calm Personality factor score, with a positive
impact on Travel Liking. It is natural that those who are less ruffled by the stresses ( last- minute
33
preparations, unexpected delays or difficulties) of traveling to a business meeting would have a
greater enjoyment of that travel.
3.6 Short- Distance Entertainment/ Recreation/ Social
Moving from mandatory to discretionary travel, Table 3.5 presents the model of Liking for short-distance
entertainment/ recreation/ social travel. This model contains many variables ( with the
same signs) as previous models, including: personal income category, travel dislike factor
score, status seeker factor score, family/ community- related factor score, calm personality score,
and “ how often do you travel … to explore new places”. As similar interpretations could be
applied here, a detailed discussion of these variables is not presented.
An interesting variable is the commute mode dummy variable for the rail mode. This variable
enters with a positive coefficient, indicating that those who commute to work via rail modes
( heavy, light, BART) enjoy traveling for social purposes more, on average, than those with a
different primary commute mode. It may be that those who commute on rail have a strong
desire to participate in automobile travel, but are precluded from doing so during the work trip
because of congestion and/ or parking costs. Social travel during non- peak times allows for
congestion- free driving, which may be enjoyed more by those who are “ forced” to commute via
transit. Another interpretation is that traveling on common carrier modes such as rail may be an
indicator of a more socially- oriented personality, which would therefore be more likely to enjoy
travel for social purposes
The travel stress factor score appears in this model, and is similar to the travel dislike variable in
that it represents the negative side of travel, in this case due to factors such as unsafe or
nervous feelings when traveling ( unsurprisingly, travel stress and travel dislike are strongly
correlated, with a correlation coefficient of 0.428 – significant at the 99 percent confidence
level). As expected, this variable enters the model with a negative sign, and is more prevalent in
the long- distance Travel Liking models presented later in this Section. The negative impact of
the frustrated Lifestyle variable ( a measure associated with such statements as “ I often feel like
I don’t have much control over my life”) on the Liking for entertainment travel may represent a
type of person who has a dour outlook on life in general, someone for whom life does not hold a
great deal of fun.
An interesting result is the appearance of the pro- high density factor score with a negative
coefficient. Such a result indicates that those with favorable high- density attitudes are less likely
34
to enjoy traveling for social purposes. This result could be interpreted to mean that these
individuals prefer to live in an exciting neighborhood, where travel by automobile is not required
when dining out or seeing a movie. These individuals probably do not associate trips on foot to
these locations with travel, and their negative feelings for this type of travel may be related to
experiences of having to drive long distances, or in heavy congestion, to engage in social
activities.
Table 3.5: Model of Liking for Short- Distance Entertainment/ Recreation/ Social Travel ( N= 1,327)
Dependent Variable : Liking for short- distance entertainment/ recreation/ social travel [ 1,…, 5]
Explanatory Variables Coefficient t- statistic Beta
Constant 3.35 31.38
Socio- Demographics
Personal income category [ 1,…, 6] - 0.0387 - 2.66 - 0.0740
Commute mode dummy – rail [ 0,1] 0.115 2.17 0.0565
Attitudes
Travel dislike factor score [- 1.8,3.7] - 0.108 - 3.79 - 0.123
Pro- high density factor score [- 2.5,2.3] - 0.103 - 4.11 - 0.111
Travel stress factor score [- 1.9,2.9] - 0.106 - 3.74 - 0.116
Lifestyle
Status seeker factor score [- 1.7,2.7] 0.153 5.80 0.166
Frustrated factor score [- 2.0,2.7] - 0.0870 - 3.35 - 0.0958
Family/ community- related factor score [- 3.9,2.1] 0.0876 3.17 0.0864
Personality
Calm factor score [- 2.9,2.4] 0.0597 2.27 0.0643
Excess Travel [ 1,2,3]
How often do you travel … when you need time to think 0.0946 2.73 0.0757
… to explore new places 0.108 2.69 0.0772
[ ] = range of possible or observed responses
Adjusted R2 = 0.118 ( R2 = 0.125) F- statistic = 17.09 ( p = 0.000)
The final new variable in the model is the Excess Travel measure of traveling when needing
time to think. Again referring to Figure 3.1, it is possible that this variable best captures a need
for solitude and reflection. Although this may seem at odds with the social element of travel in
this category, it can be quite consistent with travel for recreation, where one purpose of the
recreation may be to “ recharge one’s mental batteries”, so to speak. Given that travel to
recreational activities and/ or for recreational purposes ( e. g. a walk or a jog) often occurs at less
35
congested times and places, the travel itself may contribute to this mental recharging role, and
thus increase its enjoyment.
3.7 Short- Distance Personal Vehicle
In addition to inquiring about travel by purpose, as summarized in the previous sections, the
survey instrument also collected data on travel by mode. Table 3.6 presents the model of Liking
for all short- distance travel in a personal vehicle – with no distinction made between travel as a
passenger and travel as a driver.
Table 3.6: Model of Liking for All Short- Distance Travel by Personal Vehicle ( N= 1,344)
Dependent Variable : Liking for all short- distance travel by personal vehicle [ 1,…, 5]
Explanatory Variables Coefficient t- statistic Beta
Constant 3.697 29.27
Objective Mobility
Past year ( log) total long distance miles [ 0,12.8]* - 0.0328 - 2.78 - 0.0729
Weekly total short- distance travel ( miles) [ 5,1500] - 0.000524 - 3.91 - 0.105
Socio- Demographics
Female [ 0,1] 0.106 2.32 0.0595
Attitudes
Travel dislike factor score [- 1.8,3.7] - 0.128 - 4.07 - 0.121
Pro- high density factor score [- 2.5,2.3] - 0.133 - 3.91 - 0.122
Commute benefit factor score [- 2.9,2.6] 0.0889 3.07 0.0871
Travel freedom factor score [- 3.0,2.3] 0.200 6.30 - 0.197
Pro- environmental solutions factor score [- 2.3,2.4] - 0.201 - 6.25 - 0.195
Lifestyle
Status seeker factor score [- 1.7,2.7] 0.105 3.68 0.0959
Excess Travel [ 1,2,3]
How often do you travel … by a longer route to experience
more of your surroundings
0.108 2.69 0.0720
[ ] = range of possible or observed responses; * Logarithm ( miles + 1) to avoid taking log of zero
Adjusted R2 = 0. 182 ( R2 = 0.187) F- statistic = 30.72 ( p = 0.000)
The Objective Mobility measures present in this model indicate that those who travel a lot ( for all
modes and purposes), either for long- distances or short- distances, tend to dislike travel by
automobile. This result is expected.
36
Perhaps contrary to popular belief, our data indicates that females, all else equal, enjoy
traveling in a personal vehicle more than males. It is possible that this result is partly capturing
an Objective Mobility effect, since, in our sample, men engage in substantially more short-distance
travel by personal vehicle than women do ( men travel an average of 207 miles per
week in a personal vehicle; females an average of 149 miles per week). Also, men in our
sample ( 91%) are more likely than women ( 77%) to be working full- time rather than part- time,
and hence, probably are more often traveling in congested traffic, which would reduce their
enjoyment of travel. Other evidence ( see, e. g., Sarmiento, 1996; Bernard, et al., 1996) shows
that women are more likely then men to be auto passengers rather than drivers, so women may
experience less stress associated with auto travel. However, the possibility of a remaining
gender effect after these confounding factors are accounted for is an intriguing subject for
further research.
Again, the Attitude and Lifestyle measures played the strongest roles in the model. Several
now- familiar variables enter into the model, namely: travel dislike, commute benefit, and travel
freedom. In contrast to ( but in support of) the implicit interpretation given in the previous model,
the pro- high density variable in this model explicitly shows that those who enjoy a high- density
neighborhood tend not to enjoy traveling in an automobile. The interpretation ( including the
potential for both directions of causality to apply) is similar to that of the suburban dummy in the
model for short- distance overall travel, discussed in Section 3.3. The single “ new” Attitude
variable is the pro- environmental solutions factor score, which has a negative impact on Liking
for personal vehicle travel. This result is logical: those with strong feelings for the environment
probably feel traveling in a personal vehicle has negative impacts on the environment. But one
can like traveling in a personal vehicle even while recognizing its negative externalities, and so it
is possible that ratings of personal vehicle Travel Liking by environmentalists are subject to a
social desirability bias.
Again, a measure of Excess Travel appears in the model. Here, the variable “ how often do you
travel … by a longer route to experience more of your surroundings” enters the model with a
positive sign. The interpretation here is that those who exhibit such behavior have an underlying
need to maintain a familiarity with their environment, which is best captured by this question in
the survey.
37
3.8 Short- Distance Bus
Another mode- specific Travel Liking measure, for travel in a bus, is presented in this section
( see Table 3.7). Variables common to this and the next model ( Travel Liking for rail), were
discussed in Section 3.2.
A few variables enter the model with expected signs. Those with higher personal incomes do
not enjoy travel in a bus; those with pro- high density and pro- environmental Attitudes do enjoy
travel in a bus.
Table 3.7: Model of Liking for All Short- Distance Travel by Bus ( N= 1,319)
Dependent Variable : Liking for all short- distance travel by bus [ 1,…, 5]
Explanatory Variables Coefficient t- statistic Beta
Constant 2.426 15.64
Socio- Demographics
Personal income category [ 1,…, 6] - 0.0824 - 4.71 - 0.125
One- way commute distance ( miles) [ 0,…, 108] 0.00707 4.04 0.109
Number of persons in HH [ 1,…, 8] 0.0532 2.67 0.0689
Attitudes
Pro- high density factor score [- 2.5,2.3] 0.256 6.95 0.219
Commute benefit factor score [- 2.9,2.6] 0.161 5.41 0.148
Pro- environmental solutions factor score [- 2.3,2.4] 0.180 5.28 0.163
Lifestyle
Workaholic factor score [- 2.1,2.7] 0.104 3.14 0.0829
Personality
Calm factor score – specific to SF [- 2.9,2.4] 0.0995 2.29 0.0593
Organizer factor score [- 2.9,2.6] - 0.110 - 3.66 - 0.0941
Mobility Limitations [ 1,2,3]
Conditions which prevent or limit taking public transportation - 0.344 - 3.09 - 0.0800
Conditions which prevent or limit riding a bicycle 0.142 2.09 0.0543
[ ] = range of possible or observed responses
Adjusted R2 = 0. 170 ( R2 = 0.177) F- statistic = 25.56 ( p = 0.000)
One of the interesting results from this model is the inclusion of the number of persons in the
household Socio- Demographic variable, with a positive coefficient. Our interpretation of this
result is that those in large households may have more constraints on their automobile usage,
and may, for the good of the family, prefer to take the bus and allow other members of the
household to use the automobiles. The expressed Liking for bus then, while different from a
38
constrained preference, may represent a “ post- purchase rationalization” of that preference –
making a virtue out of necessity, so to speak.
Perhaps a surprising result is the presence of the workaholic Lifestyle variable entering with a
positive coefficient. Workaholics, who may stereotypically be associated with automobile travel,
may find they can get to work more efficiently on transit ( especially on the tour bus- style
commuter buses) or they may see the bus as a convenient and reliable means to enter the
regional central business district ( CBD). They may also value the opportunity to work while
commuting that is presented by bus.
In contrast to the other models presented in this report, the models of Liking for bus and
rail/ train include certain variables segmented by neighborhood location. This was done because
in the San Francisco Bay Area, the bus and rail service within San Francisco is starkly different
than service in the suburbs of Pleasant Hill and Concord. Bus service in the City is frequent and
quintessentially urban. In contrast, service in the suburbs is infrequent, though comfortable
commuter buses do serve the regional CBD. Further, rail service in the City is dominated by on-street
light rail service, which is more similar to local bus service; BART only services a small
portion of San Francisco proper. In contrast, BART is the dominant rail mode in the suburbs.
For these reasons, we first estimated individual models for North San Francisco residents and
suburban residents, and then combined them to form a single, joint model. In the bus Liking
model shown in Table 3.7, only the calm Personality factor score is neighborhood- specific,
associated with North San Francisco residents. This result makes sense as those easily rattled
or made uneasy may not enjoy an urban bus service.
3.9 Short- Distance Rail
This section discusses the mode- specific model of Liking for rail ( see Table 3.8), which includes
heavy commuter rail, light urban rail, and BART ( the regional rail system in the Bay Area). This
model is similar to the previous model of Liking for bus, which is expected, though important
differences do arise. Also, this model includes certain neighborhood- specific variables, as
alluded to in the previous section.
The first neighborhood- specific variable is the female variable, entering with a negative coeffi-cient
specific to North San Francisco. It may be that females are less comfortable in driverless
rail vehicles than in city buses, where the driver is always within easy contact. Those in sales
39
occupations, if living in the suburbs, enjoy rail, a result which could be attributed to the ease with
which BART delivers passengers to the regional CBD – a prime sales market. Those in the
suburbs with higher levels of education are also more likely to enjoy rail. Again, this result could
be indicative of those having higher education levels being more likely to be working in the
regional CBD, which has excellent ( and therefore more likely to be enjoyable) rail service.
Table 3.8: Model of Liking for All Short- Distance Travel by Rail ( N= 1,295)
Dependent Variable : Liking for all short- distance travel by rail [ 1,…, 5]
Explanatory Variables Coefficient t- statistic Beta
Constant 2.777 15.52
Socio- Demographics
One- way commute distance ( miles) [ 0,…, 108] 0.00425 2.21 0.0604
Luxury vehicle type dummy [ 0,1] - 0.460 - 2.76 - 0.0696
Female – specific to SF [ 0,1] - 0.156 - 2.16 - 0.0654
Sales occupation dummy – specific to suburbs [ 0,1] 0.296 2.42 0.0622
Educational background – specific to suburbs [ 1,…, 6] 0.0619 4.02 0.132
Attitudes
Pro- high density factor score [- 2.5,2.3] 0.279 6.51 0.221
Commute benefit factor score [- 2.9,2.6] 0.203 6.26 0.173
Pro- environmental solutions factor score [- 2.3,2.4] 0.261 7.01 0.218
Lifestyle
Family/ community- related factor score [- 3.9,2.1] 0.103 2.84 0.0753
Personality
Organizer factor score [- 2.9,2.6] - 0.142 - 4.42 - 0.113
Mobility Limitations [ 1,2,3]
Conditions which prevent or limit taking public transportation - 0.560 - 4.28 - 0.116
Conditions which prevent or limit air travel 0.264 3.64 0.0948
Conditions which prevent or limit riding a bicycle 0.291 2.07 0.0552
[ ] = range of possible or observed responses
Adjusted R2 = 0. 182 ( R2 = 0.190) F- statistic = 23.13 ( p = 0.000)
A variable unique to this model is the luxury vehicle type dummy. The variable enters with a
negative coefficient, indicating that those who drive luxury cars, not surprisingly, do not enjoy
rail travel.
Two other variables of interest are the inclusion of the Mobility Limitations on air travel variable
and the family/ community- related Lifestyle score. Those unable or limited in their ability to travel
by air may be drawn to short- distance rail travel due to their familiarity with long- distance rail
40
travel, used in place of air travel. The positive coefficient on the family/ community- related factor
score may reflect the many enjoyed family trips to tourist and shopping locations served by rail
within San Francisco, taken by Bay Area families.
3.10 Short- Distance Walk/ Jog/ Bicycle
The final short- distance mode- specific model is Liking for non- motorized travel, specifically
walking, jogging, and bicycling. This definition is a bit nebulous as it may, perhaps more so than
the other categories, include both directed and undirected travel ( i. e. walking, jogging, or
bicycling as a means of exercise). However, the investigation allows for a comparison between
the types of variables included in this model with those in more directed travel categories, such
as commute travel. A summary of the coefficients is presented in Table 3.9.
Table 3.9: Model of Liking for All Short- Distance Travel by Walking, Jogging, Bicycling ( N= 1,299)
Dependent Variable : Liking for all short- distance travel by walking, jogging, bicycling [ 1,…, 5]
Explanatory Variables Coefficient t- statistic Beta
Constant 2.811 19.70
Socio- Demographics
Educational background [ 1,…, 6] 0.0595 3.08 0.0796
Number of persons age 24- 40 in HH [ 0,…, 7] 0.0933 3.81 0.0986
Number of persons age 65- 74 in HH [ 0,1,2] - 0.180 - 2.12 - 0.0535
Minivan vehicle type dummy [ 0,1] - 0.263 - 2.45 - 0.0619
Single adult with children family status dummy [ 0,1] 0.429 2.70 0.0680
Concord neighborhood dummy [ 0,1