Transportation Research Part A 78 (2015) 506–518

Contents lists available at ScienceDirect

Transportation Research Part A

journal homepage: www.elsevier .com/locate / t ra

Modeling travel behavior by the structural relationshipsbetween lifestyle, built environment and non-working trips

http://dx.doi.org/10.1016/j.tra.2015.06.0160965-8564/� 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +98 09177113447.E-mail addresses: roya.etminani@gmail.com (R. Etminani-Ghasrodashti), mahyarardeshiri@yahoo.com (M. Ardeshiri).

Roya Etminani-Ghasrodashti a,⇑, Mahyar Ardeshiri b

a Department of Art & Architecture, Science & Research Branch, Islamic Azad University, Tehran, Iranb Department of Architecture & Urban Planning, Beyza Branch, Islamic Azad University, Beyza, Iran

a r t i c l e i n f o

Article history:Received 5 April 2015Received in revised form 18 June 2015Accepted 29 June 2015

Keywords:Travel behaviorNon-working tripsLifestyleBuilt environmentStructural equation model

a b s t r a c t

In the context of sustainable urban transport in developing countries, individuals’ travelbehavior faces multiple factors which influence their mobility patterns. Recognizing thesefactors could be a favorable method to organize more regular and sustainable trip patterns.This study aims to identify the less well-known lifestyle along with more popular builtenvironment as the main factors which shape travel behaviors. Employing data from 900respondents of 22 urban areas in city of Shiraz, Iran, this paper explores travel behaviorsas non-working trip frequencies by different modes. Results of structural equation modelindicate a strong significant effect of individual’s lifestyle patterns on their non-workingtrips. However, built environment impact on travel behavior is small compared to lifestyle.Besides, other variables such as travel attitudes and socio-economic factors stay crucial inthe mode choice selection. These findings indicate the necessity of regarding lifestyle ori-entations in travel studies as well as objective factors such as land use attributes.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction

In recent decades, Iran like many developing countries, have encountered the growth of usage private vehicles as a dom-inant pattern of trips in urban areas. Reports suggest that, Iran is the second-largest oil-consuming country in the MiddleEast and Iranian domestic oil consumption is mainly diesel, gasoline, and fuel oil. In 2013, FGE estimates that Iran importedalmost 17,000 bbl/d of petroleum products, of which roughly 85% was gasoline (EIA, 2014). The daily trips by cars are amongimportant sources of energy consumption and air pollution. Among the world’s top ten most polluted cities, four are in Iran,according to data based on a 2013 World Health Organization index. Automobile use in large cities of Iran has extended inthe past years according to transportation ministry estimates (The guardian, 2014). To manage this increasing energy con-sumption, we should first recognize the factors which shape individuals travel patterns and the reasons behind vast car use.

There is a huge body of researches concerned with the impact of built environment indicators (e.g. high density,mixed-land use, design and accessibility) on the individual’s travel behavior (e.g., Crane, 2000; Ewing and Cervero, 2001;Handy et al., 2005; Stead and Marshall, 2001; Cervero and Murakami, 2010). Accordingly, multiple studies investigate thefactors which influence urban trip patterns in Iranian cities by considering demographic and land use characteristics (e.g.Etminani Ghasrodashti, 2009; Soltani and Esmaeili, 2011; Masoumi, 2013). Cited studies which have been conducted in citieswith high traffic jam, indicate a strong impact of demographic characteristics, in contrast to urban form and built environ-ment, in determining individual mobility patterns (Soltani and Etminani Ghasrodashti, 2010: 151). The results manifest that

R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518 507

apart from urban form characteristics, some hidden factors play a crucial role in shaping the mobility patterns in Iraniancities; these factors can be derived from culture, lifestyle, individuals’ attitudes and beliefs toward mobility modes.

In recent years attention to the impacts of subjective factors such as attitudes, beliefs, personality and also lifestyle ontravel patterns has significantly increased (Mokhtarian and Salomon, 2001; Bagley and Mokhtarian, 2002; Anable andGatersleben, 2005; Ory and Mokhtarian, 2009; Van Acker et al., 2014). These studies have tried to answer the question thathow the subjective factors such as lifestyle and attitudes along with land use attributes could influence travel behavior.Although they argue the lifestyle as the patterns of leisure and consumption, in most of studies it has limited to how peopleexpress their social position through behavioral patterns. In contrast, in present study the lifestyle is explored as the patternof consumption and leisure activities by different modes of mobility. Besides, built environment factors along withsocio-economic characteristics and attitudes toward travel and residential neighborhood have been taken into account asother factors affect travel behavior.

The paper is organized as follows. Section 2 reviews the literature on the interaction of travel behavior, built environmentand lifestyle in brief. Section 3 expresses the data used in the study. Section 4 indicates the empirical results. The final sectionconcludes the paper.

2. Literature review

Since 1970s, activity-based modeling led to offer more realistic representation in individuals travel behaviors comparedto traditional trip-based travel demand models. Activity-based modeling predicts travel behaviors through differences inhousehold factors (such as socio demographics and life stage) and non-household factors (built environment, socio environ-ment, travel policies).

In this regard, vast studies argue that travel behaviors has notably influenced by built environment via different variables.Some research have explored the impacts of land use attributes on motorized and non-motorized trip frequencies (e.g.,Handy, 1993; Boarnet and Sarmiento, 1998; Boarnet and Crane, 2001; Chatman, 2008); while others have been conductedon vehicle mile traveled (Chatman, 2008). Moreover, to decrease methodological limitations, some studies have tried tosearch the impact of built environment on individual mode choice (e.g., Cervero, 2002; Chatman, 2003; Ewing et al.,2004; Frank et al., 2008; Lee et al., 2014). On the other hand, built environment factors are undoubtedly one of the mostheavily research subject in travel studies and the most cited factors of land use are named as Ds. The original ‘‘3Ds’’ createdby Cervero and Kockelman (1997), are density, diversity, and design, followed later by destination accessibility and distanceto transit (Ewing and Cervero, 2001, 2010; Ewing et al., 2009). However, most of research have simplified travel studies andsuffer from ignoring the subjective factors which influence individual trip patterns.

Travel behavior can also explain by social expectations about behavior such as norms, values, beliefs, attitudes and finallylifestyle. In this regard, two leading theories consist of value-belief-norm theory (Stern et al., 1999) and the theory of plannedbehavior (Ajzen, 1991; Fishbein and Ajzen, 1972) support this fact that perceived social norms are considered to be anotherpossible determinant of behavior. Therefore, a hierarchy of decisions is made by travelers where decisions at a higher level(such as lifestyle) determine the scope of actions at lower levels (such as travel behavior).

The notion of lifestyle in transport studies was introduced by 1970s. At the early stage, lifestyle concept was defined asbehavioral responses in terms of socio-economic differences, personal and social actions (Reichman, 1977). Using thisdefinition, some travel studies tried to investigate the impact of lifestyle on travel patterns, but in fact they just refer thelifestyle to some objective characteristics such as stage of life or household composition (e.g., Salomon and Ben-Akiva,1983; Cooper et al., 2001; Hildebrand, 2003). Moreover, some studies indicate that households with similarsocio-economic attributes do not travel in similar patterns (van Wee, 2002; Mokhtarian and Cao, 2008). This disparitiesin travel patterns among individuals originate from various lifestyles. So, considering the lifestyle just as observable behav-iors result from socio-economic differences between groups may not clearly explain the trip behaviors.

In another definition, lifestyle is behavioral patterns which derived from underlying opinions and orientations, includingbeliefs, interests and attitudes (Kitamura, 2009). Therefore, in some travel studies lifestyle has referred to individual’s atti-tudes toward work, family, money, status, and the value of time. For example, in order to determine the subjective factorswhich influence travel demand, Collantes and Mokhtarian (2007) introduced lifestyle groups such as status seeker, worka-holic, family/community-oriented and frustrated. They found that, Individuals with a family-oriented lifestyle as well asindividuals with a frustrated lifestyle frequently used their car for short-distance trips, family-oriented lifestyle was relatedwith fewer long-distance leisure trips and also, workaholics travel significantly fewer short distance as well as long-distancetrips for leisure purposes.

So, individuals indicate their social situation via specific patterns in consumption and leisure. Therefore, some studiesfocus on lifestyle expressions which are observable patterns of behaviors and reflect someone’s lifestyle. Using this definitionof lifestyle, in an empirical study conducted by Scheiner (2010) in order to modeling trip distance traveled, lifestyle data wascollected as leisure preferences, values and life aims, esthetic taste and frequency of social contacts. The results of the studyindicated that lifestyle has the strongest impact on leisure trip distances.

Similarly, Van Acker et al. (2014) in a travel study utilized holiday aspects, literary interests and leisure activities as thelong-term lifestyle decisions to develop a modal choice model. She found that, car availability tends to be higher amongrespondents with a more active lifestyle.

Lifestyle

Travel behavior

Socio-economic attributes

Individual attitudes

Built environment

Direct effects of key variables on travel behavior

Interrelations between lifestyle and other variables

Feedback procedures not modeled

Fig. 1. Conceptual model of travel behavior.

508 R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518

This paper aims to contribute to both lifestyle and built environment debates by studying travel behavior as non-workingtravel amounts in three different modes (car, public transit and walking/cycling trips). It also takes into considerationsocio-economic characteristics as well as individual attitudes toward the travel and residential neighborhoods. Fig. 1 visu-alize how travel behavior can be explained by key variables and other factors. According to the special context of Iran, weconsider the socio-economic attributes as the factor which influences travel behavior from the top of the hierarchy of deci-sions that are made by travelers. We assume that socio-economic attributes of individuals influence their attitudes towardtravel and residential neighborhoods, and individual attitudes have effect on their lifestyle. In our conceptual model,socio-economic attributes also influence travel behavior directly and indirectly through built environment. Furthermore,built environment which influence travel behavior, can also impact on lifestyle of individuals. The reverse relations can alsobe possible, but not modeled in this study.

3. Methodology

3.1. Data and study area

The data used in this study was based on authors’ field survey and questionnaire and conducted in 22 study areas in thecity of Shiraz, Iran in 2014 (see Fig. 2). Shiraz is the largest city in south of Iran with the population of 1.5 million, an area of178,891 km2 and 416,141 households (Shiraz Municipality, 2014). The sample was split between areas with various urbanforms and socio-economic characteristics. Selecting the case studies in both urban and suburban areas allow us to examinethe impact of residential neighborhood type on travel-related behaviors (Schwanen and Mokhtarian, 2005). The samplingmethod used in this study was Multi-Stage Cluster sampling. The priority of sampling was with areas lay within municipalregions. Then, according to the property prices, three ranges of price (high, medium, low) for land use were determined.Finally, with regards to the 160 traffic zones of Shiraz, the most important residential neighborhoods were selected as studyareas in three ranges of land prices. Therefor, 22 study areas were selected as one-quarter mile buffer area, becausebuffer-based built environment measures are more reasonable and stringent than those considering TAZ-based variablesin terms of explanatory power and statistical significance (Lee et al., 2014). Accordingly, to determine the sample size basedon the population size, 900 heads of household contributed in vast face to face interviews and completing questionnairesabout their socio-economic characteristics, travel behaviors, lifestyles, residential and travel attitudes. At the end of the sur-vey, we geocoded the residential location in ArcGIS 10 and added objectively measured built environment characteristics ofresidential locations. The compositions of collected data are unique in Iran in so far that they let us to draw out a large varietyof information in the level of individuals.

Table 1 displays some key socio-economic characteristics of 900 respondents. This table reveals that people between the41–64 years old and full time workers have the most observations. As the monthly income has changed from rail to dollar,the real value of income toward the cost of living in a developing country has not shown in this table.

Fig. 2. Location of study areas in Shiraz.

R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518 509

3.2. Travel behavior

As discussed in Section 2, in this paper travel behavior has explored as individuals’ non-working trips. The most commontravel measures are trip frequency, trip length, mode choice, and VMT (Ewing and Cervero, 2001). Hence, in this study bothof trip frequency and mode choice have been modeled as the travel outcome. Frequency of home-based non-working trips isthe measure of travel behavior in our model. Respondents were asked to report the number of their non-working trips todifferent destinations in a typical week by six travel modes; private car and alone, private car by others, bus, taxi, walkingand biking. Each of these six variables is somehow related to another one, so they can be combined into three generalnon-working trip factors (principal axis factoring, 73.4% variance explained). These extracted factors are characterized bythe following frequencies: 43% non-working trips by car, 27% by public transport and 30% walking/biking trips.

3.3. Lifestyle

The particular focus of this paper is the impact of lifestyle on travel behavior. Recent travel behavior studies mainly usethe behavioral patterns of activity and time use, especially in the field of leisure, as the lifestyle measurements (etc. VanAcker et al., 2014). However, the past travel behavior surveys generally lack information on the lifestyles which result frompatterns of activities by different means of mobility. This is the distinction between current study and others. Lifestyles arepresented in the data using two main fields: leisure and consumption activities by three modes of travel (private car, publictransport and walking/cycling). Since lifestyles influenced by other factors like economic, cultural and living status dimen-sions (Ganzeboom, 1988), we tried to choose activities which include various economic and cultural conditions. In general,the wealthier, knowledge- and education-oriented, democratic and technologically evolved a society is, the more lifestylesare created in it that synchronically coexist along each other. In less developed societies four basics social fields of

Table 1Socio-economic characteristics of sample (N = 900).

Individual’s characteristics Frequency (%)

Sex Female 198 (22)Male 702 (78)

Age 18–23 34 (3.8)24–40 349 (38.8)41–64 452 (50.2)>65 years 65 (7.2)

IncomeLess than 500 $ 580 (64.4)500–1000 $ 290 (32.2)>1000 $ 25 (2.8)Missed data 5 (0.6)

Current employment statusPart-time 45 (5)Unemployed 254 (28.2)Full-time 403 (44.8)Retired 173 (19.2)Student 25 (2.8)

Household characteristics Mean (std. dev.)

People under 18 in family 0.80 (0.87)People above 18 in family 2.71 (1.21)Number of employees in family 1.56 (0.88)Number of private cars in family 1.13 (0.78)

510 R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518

economics, politics, culture and religion are more tied to each other and closely connected. So, it follows that societies inwhich all four discourses are influencing each other directly produce on average fewer different lifestyles (Benedikteret al., 2011).

The survey included a list of 17 leisure interests and shopping-service activities. Respondents had to indicate how oftenthey monthly refer to different places for these activities in a six-point Likert-type answer scales from ‘never’ to ‘more thanfive times per week’. Inquiry about these activities performed three times separately according to the place of each activity;inside and outside the neighborhood; and based on the various modes of transport.

Overall trip frequency to each destination was calculated by summing the individuals’ trip frequencies by three differenttravel modes through an indicator (from 0 to 20) across 17 leisure and shopping-services destinations and then the overallfrequencies have changed to the six-point Likert-type answer scales. Responses on these survey items were factor analyzedin SPSS 22. In the factor analysis, the number of factors was determined based on the interpretability of factors, combinedwith interpretation of the scree plot and all eigenvalues larger than one.

These 17 types of leisure interests and shopping-service activities were factor analyzed (maximum likelihood, promax rota-tion, 53.3% variance explained, KMO = 0.879) into four underlying dimensions as: modern-oriented, traditional-oriented,consumer-oriented and education-oriented lifestyles. As it has shown in Table 2, in this study modern-oriented lifestyle refersto leisure activities such as going to a club for exercise, going to cinema and theater, and going to restaurant and coffee shops.These leisure activities are mostly shaped on modern life patterns. Modern-oriented lifestyle which is linked to contemporarytechnology, often associated with more modernity and less religious-traditional approaches. Traditional-oriented lifestyle hereis related to the leisure activities that are based on simple and routine patterns of individuals’ life and religious terms such asgoing to park and green spaces, visiting relatives and attend in religious ceremonies. Consumer-oriented lifestyle stem fromshopping-service activities which are primarily and weekly needed by individuals such as going to service providers or foodand cloth stores. Education-oriented lifestyle is count for training activities which individuals have performed to promote theirknowledge or professional expertise.

3.4. Built environment

Using information from various land use databases (ArcGIS 10), we calculated several spatial characteristics of the builtenvironment in 1=4 miles buffer areas. The built environment characteristics include diversity measure (entropy index),design measures (street density, internal connectivity), density measures (population density, residential density, job den-sity) and accessibility measures (distance from home to closest bus station and closest intersection). All of these variousmeasures were computed at trip origins by utilizing the parcel-based land use data and GIS technique. Hence, it is possibleto evaluate the impacts of built environment of buffer areas on home-based non-working trips.

However, these land use characteristics can be related to each other (a place with high job density is a diverse place too). Inorder to display the structure among built environment measures, we conducted a factor analysis (principal axis factoring, 64.4%variance explained, see Table 3) which revealed three factors: spatial distribution, density and local accessibility to transit.

Table 2Pattern matrix for lifestyles.

How often do you go to each of following places? Lifestyle factors

MOL TOL COL EOL

Service provider (e.g., banks, ATM, post office) 0.268Children school 0.626Educational centers of yours or your spouse 0.516Health centers (e.g., family doctor, clinic) 0.347Food stores 0.593Cloth stores 0.541Health care stores 0.936A club for exercise (e.g., gym, swim pool, tennis court, basketball, football, and etc.) 0.538A place for walking, cycling, mountaineering, jugging 0.515Cinema, theater, music concert, art exhibition 0.605Park and green spaces 0.563Visiting relatives 0.461Restaurant and coffee shops 0.671Natural gardens (e.g., garden cities) 0.437Strolling in malls and shopping centers for fun 0.438Strolling in streets without any purpose 0.281Attend in religious ceremonies 0.518

Note: MOL = modern-oriented lifestyle, TOL = traditional-oriented lifestyle, COL = consumer-oriented lifestyle, EOL = education-oriented lifestyle.

Table 3Pattern matrix for built environment factors.

Land use measures Built environment factors

Spatial distribution Density Local accessibility to transit

Entropy index 0.786Street density 0.711Internal connectivity �0.636Population density 0.884Residential density 0.786Job density �0.563Distance from home to closest bus station (minute) 0.791Distance from home to closest intersection (minute) 0.772

R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518 511

3.5. Other key variables

This paper investigates subjective as well as objective influences on travel behavior, and it particularly emphasizes on theeffects of lifestyle and built environment. Other subjective influences include individual attitudes toward the travel and res-idential neighborhood. Individual attitudes can influence decision-making indirectly and irrationally (Ajzen, 2001). Somestudies have argued the importance of individual attitudes impacts on travel behavior by investigating people’s travel likesand dislikes, their opinions about environmental issues, their views about commute benefits, and their sense about travelfreedom and travel stress (Mokhtarian et al., 2001; Ory and Mokhtarian, 2009). The survey contained 13 statements aboutindividual’s attitudes toward travel. Respondents were asked to state their opinions about each statement on a five-pointordinal scale ranging from ‘strongly disagree’ to ‘strongly agree’ according to variables that influence their beliefs about tra-vel. These statements were then factor analyzed (principal axis factoring, promax rotation, 58.4% variance explained,KMO = 0.720) into five main factors: physical-oriented, time management, pro-environmental, individualist and eco-nomic–political attitudes (Table 4).

Since previous travel studies on lifestyle examine individuals attitudes and preferences toward their residential neighbor-hood (see, e.g., Bagely and Mokhtarian, 2001; Van Acker et al., 2014), the survey also includes 14 statements on attitudestoward residential neighborhoods. Individuals were asked to indicate their preferences about residential areas if they werelooking for a new place to live (even if they were not intending about moving) on a five-point ordinal scale ranging from‘unimportant’ to ‘very important’. By using factor analysis (principal axis factoring, promax rotation, 51.4% varianceexplained, KMO = 0.814), three main factors were extracted for residential attitudes: comfort, social cohesion and trafficresistance (Table 5).

Our survey also supported data on key objective variables such as socio-demographic characteristics. A factor analysis ofsocio-demographic characteristics (principal axis factoring, 77.7% variance explained, see Table 6) resulted in three factorsconcludes: household composition (People above 18 and Number of employers in the family), financial position (Income andNumber of private car in the family), profile (age of respondent). These three factors have been addressed in models associo-economic factors.

Table 4Pattern matrix for travel attitudes.

How much do you agree with the followingstatements?

Travel attitudes factors

Physical-oriented Timemanagement

Pro-environment Individualist Economic–political

Increasing the fuel price was needed in order tolimit car use by people in Iran

0.711

We need more public transport, even if we have to paymore tax for it

0.747

I am ready to limit my car use in order to decreasetraffic congestion and air pollution

0.595

I would definitely use bicycle for my trips, if a suitablebike lane be designed in Shiraz

0.786

We need more highways in Shiraz 0.88It is needed to expand the roads and make more

multilevel junctions in Shiraz0.882

Widening the roads leads to flow the traffic 0.6I feel more safety when I travel by my private car 0.585I do not feel comfortable when I travel with others by

bus0.709

When I travel by my private car, I can travel in anypath, any road

0.672

Getting stuck in traffic makes me nervous 0.544Nowadays, it is difficult to find a place in streets to

park my car0.784

When I park my car in a street for a long time, I amworried about it being stolen

0.719

Table 5Pattern matrix for residential attitudes.

If you are looking for a new place, how important is each offollowing attributes of neighborhood?

Residential attitudes factors

Comfort Social cohesion Traffic resistance

Existing suitable sidewalks 0.603Low traffic 0.543Close to bus stop 0.382Green and open spaces 0.55Calmness 0.811Social safety, lack of crime 0.836Cleanliness 0.329Good views and scenery 0.66Close to grocery stores 0.734Close to job location 0.3Close to neighborhood park 0.45Close to family and friends 0.627Existing neighbors with special ethnicity 0.75Friendly relationship with neighbors 0.778

Table 6Pattern matrix for socio-economic factors.

Socio-demographic characteristics Socio-economic factors

Household composition Financial position Profile

People above 18 in family 0.851Number of employees in family 0.81Income 0.885Number of private cars in family 0.637Age 0.964

512 R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518

4. Structural equation model

In this part, the variables witch illustrated in Section 3 has been entered as input for the evaluating of structural equationmodel by the software package Amos 22. Using SEM in transport studies has more advantages toward single equation mod-els. It allows the researcher to evaluate the effect of variables simultaneously (Hoyle, 2012). The conceptual model of the

R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518 513

paper implies multiple relationships between lifestyle, built environment and travel behavior variables. So, in SEM modelsone variable that is an endogenous variable in one set of the relationships, can be an exogenous variable in another equationat the same time.

4.1. Estimation methods

The discrepancy between the population moments and the moments implied by a model depends not only on the modelbut also on the estimation method. To estimate the models, a maximum likelihood (ML) method has been used. This methodreveals that to what extent the observed variances and covariance between measures can be explained by the model. UsingML estimation of an SEM needs the normal distribution of endogenous variables (Harrington, 2009). Since some key vari-ables in this study (e.g., lifestyles and attitudes) are categorical, great consideration has been taken in the model estimation.Amos 22.0 performs Bayesian model fitting for ordered-categorical data. The other three alternative methods have been usedto estimate SEM models include: Asymptotically distribution-free (ADF), generalized least squares (GLS), and un-weightedleast squares (ULS). ADF method works better in estimation of standard errors and does not desire normal data. Each of thesemethods which have their own merits and demerits, demonstrate relatively constant results and lead to increasing our con-fidence to model estimation. Moreover, to indicate the quality of model specification (compare the covariance matrix ofobserved variables of population with sample) some fitting indices have been used. The v2-statistics test is an appropriatefitting index, although it is almost significant in large sample sizes (Iacobucci, 2009). Hence, other indices such as v2 testvalue divided by the model degrees of freedom, normed fit index (NFI), comparative fit index (CFI) and root-mean squareerror of approximation (RMSEA) have been employed in this study. Table 7 indicates the brief summary ofgoodness-of-fit based on estimation methods for SEM model on travel behavior. These values compare well to standard val-ues which have been showed in the table, where the v2/df value is about 2, the NFI value is 0.99, the of CFI is 0.97 and theRMSEA average value is 0.067.

4.2. Model results

After specifying the key variables and estimation methods, in this part the modeling results have been discussed. We firstconsider some interrelations examined in the conceptual model of the study, afterward the total effects of key variables onnon-working trips will be explained. Other interrelations examined in the model framework, such as effects ofsocio-economic factors on built environment and attitudes, are excluded from interpretation due to lack of space. In orderto realize direct and indirect relationships between all variables, both significant and insignificant total effects will bereported in SEM tables. Tables 8 and 9 summarize standardized total effects of SEM model and report significant effectsbetween variables.

4.2.1. Interrelations between lifestyle and other variablesIn this study it has assumed that lifestyle which impact travel behavior, influence by other variables like socio-economic

attributes, attitudes and built environment. Accordingly, SEM model on travel behavior, consider the impacts of interrela-tions on lifestyle too. In the early theoretical discussions on the definition of lifestyle, socio-economic and demographic attri-butes considered as the main determinants of lifestyle (Ganzeboom, 1988; Kitamura, 1988). In this paper socio-economicfactors are still a crucial determinant of individuals’ lifestyle. It can be observed from Table 8 that one of the most importantinfluences of socio-economic factors is the negative impact of profile, on MOL. It seems that elderly people were lessmodern-oriented. In contrast, employees in the family, number of peoples above 18, car ownership and monthly incomehave a positive association with MOL. Also, it is clear that age of respondents were negatively associated with EOL.

Through travel attitudes, pro-environment attitude (such as individuals who were willing to reduce their car usage) haveindicated a direct relationship with both MOL and TOL. Nonetheless, people with traditional lifestyle were more likely to beenvironment friendly. Individualist attitude (such as people who liked to travel alone by their private car), has a significantnegative impact on TOL and a positive impact on COL. Moreover, economic–political attitude (such as people who agree withincreasing the fuel price) has showed a strong positive impact on TOL. So it reveals that individuals’ travel attitudes haveaffected on their lifestyle.

Among the residential attitudes, social cohesion positively influence on TOL. This result can be rises from the fact thatrespondents who sought social relationships in a neighborhood were more willing to have traditional leisure activities suchas going to parks or religious ceremonies. This finding could be the same but weaker for traffic resistance attitude toward aresidential neighborhood. Respondents who were preferred to live in a neighborhood with suitable sidewalks and low traffic,

Table 7Goodness-of-fit measures.

Model standard value v2, (df), q v2/df < 2 NFI > 0.95 CFI > 0.95 RMSEA < 0.1

Model-based value 273.24, 138, 0.000 1.97 0.99 0.97 0.067

Table 8The total standardized interrelation effects on lifestyles.

Effect of . . . ; on . . . ? Lifestyles

MOL TOL COL EOL

Socio-economic factorsHousehold composition 0.151*** 0.053 0.068 �0.021Financial position 0.150*** �0.032 0.045 0.039Profile �0.340*** 0.001 0.041 �0.134***

Travel attitudesPhysical-oriented 0.039 0.034 0.084* �0.012Time management 0.021 0.013 0.044 0.006Pro-environment 0.079** 0.128*** 0.039 0.067Individualist 0.056 �0.075* 0.097* 0.014Economic–political �0.05 0.124*** 0.012 0.061

Residential attitudesComfort �0.090** 0.01 �0.085* �0.105**

Social cohesion 0.067 0.153*** �0.012 �0.005Traffic resistance 0.006 0.080* 0.002 �0.018

Built environmentSpatial distribution 0.103** 0.068 0.012 �0.031Density �0.058 0.032 �0.095** �0.043Local accessibility to transit 0.012 0.015 0.1 �0.009

Note:* Significant a = 0.1.

** Significant a = 0.05.*** Significant a = 0.01.

Table 9Total standardized effects on non-working trips.

Effect of . . . ; on . . . ? Non-working trips by car Non-working trips by public transport Non-working trips by walking/cycling

LifestyleMOL 0.106*** 0.009 �0.054TOL 0.058 0.146*** 0.205***

COL 0.102*** 0.022 0.100***

EOL 0.015 0.160*** 0.045

Travel attitudesPhysical-oriented 0.067* 0.049 0.057Time management �0.025 0.058 0.029Pro-environment �0.161*** 0.226*** 0.117***

Individualist 0.083* �0.025 �0.037Economic–political 0.001 0.181*** 0.160***

Residential attitudesComfort 0.064* �0.071* �0.052Social cohesion �0.001 �0.021 0.067*

Traffic resistance 0.041 0.028 0.059

Built environmentSpatial distribution �0.176*** 0.013 0.024Density �0.065* 0.014 0.062*

Local accessibility to transit 0.024 0 �0.021

Socio-economic factorsHousehold composition 0.97* 0.076* 0.002Financial position 0.191*** �0.174*** 0.003Profile �0.017 �0.007 0.109**

Note:* Significant a = 0.1.

** Significant a = 0.05.*** Significant a = 0.01.

514 R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518

had traditional-oriented lifestyle. Besides, comfort attitude have indicated negative association with MOL, COL and EOL. Itseems that in the case of this residential attitude, respondents’ preferences do not necessarily follow their lifestyle.

Some of the built environment factors are also associated to lifestyle. Spatial distribution positively impacts on MOL. Itmeans that Mixed-land uses, street design and connectivity in residential neighborhood were the factors which lead peopleto have modern-oriented lifestyle. Interestingly, density has negatively influenced on COL. Respondents who reside in dense

R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518 515

neighborhoods were less willing to have a consumer-oriented lifestyle. On the other hand, these respondents have indicatedmore traditional-oriented lifestyle.

4.2.2. Effects of key variables on non-working tripsTable 9 indicates SEM model of non-working trips by three modes of transport based on our conceptual model. As men-

tioned in previous section, the model explores the impacts of key variables on non-working trips by controlling all interre-lations on lifestyle. At the first row of Table 9, the impacts of lifestyle on non-working trip frequencies by different modeshave been investigated. Lifestyle is among the most important determinants of non-working trips. It is clear from Table 9that MOL and COL are positively related to number of non-working trips by car per week. It reveals respondents who wereconcerned about modern entertainments and people with more frequency of tours to shopping-service destinations, pre-ferred to use their private car to do their activities. Non-working trips by transit were significantly related to TOL andEOL. Among these kinds of lifestyles, education-oriented respondents were more likely to use public transport in their dailytrips. Moreover, non-working trips by walking/cycling were influenced by TOL and COL. It seems that individuals with tra-ditional leisure activities were strongly eager about walking and cycling trips. Since COL in this study refers to consumptionof goods or services which are primary and weekly needs of individuals such as shopping foods, cloths or using service pro-viders like banks, this lifestyle was performed inside the borders of individuals’ neighborhood by walking/cycling and out-side the neighborhood by private car. So, it can be concluded that COL refers to shopping-services activities with both shortand long distances from home and it was associated with car and walking trips.

The association between travel attitudes and non-working trips by different modes has been noted in second row ofTable 9. Non-working trips by car influenced by pro-environment attitudes. This travel attitude shows a significant negativeimpact on car trips. Respondents who were ready to decrease and alter their car use to sustainable modes were less likely toutilize private car in daily non-working trips. In spite of pro-environment attitude, physical-oriented and individualist atti-tudes were positively associated with car trips. It shows that respondents who expected more facilities in roads andextended highways and people who felt more comfortable in using their private car, were eventually more car users. Inconcern with non-working trips by public transport and walking–cycling, again pro-environmental and economic–politicalattitudes has known significantly. Both of these attitudes were positively related to using public transport and walking/cycling. It is clear that pro-environment attitudes were the most determinants of non-working trips.

Residential attitudes are slightly related to frequencies of non-working trips. Respondents who were prefer to live in acomfort neighborhood (e.g., green and open spaces, calm and clean residential area) used more private car and less publictransport in their daily trips. Attitudes toward the residential neighborhood might change once the residential location ischosen (Van Acker et al., 2014), but in our model residential attitudes does not influenced by residential environment.We assumed residential attitudes are the preferences of individuals about a residential neighborhood which could impacton their travel behavior. This assumption only has been truly proved through social cohesion attitude; respondents whotended to have social cohesion in their neighborhood were more willing to do walking/cycling trips.

The built environment factors seem to have less significant impacts on non-working trips. The most significant factor isspatial distribution. Respondents, who reside in diverse neighborhoods with well-connected and designed streets, were lesslikely to use private car for their weekly trips. This finding is compatible with the impact of diversity on travel behavior inprevious studies results. Moreover, density in neighborhoods led to decrease of car use and increase in walking/cycling.

As we found from socio-economic factors, financial position have a positive effect on non-working trips by car and neg-ative impact on public transport. It is clear that higher income and car ownership are responsible for more non-working tripsby car and less tendency to use public transport.

Household composition which point to number of employees and people above 18 in the family, were related positivelyto both car and public transport. It is obvious that large families logically produce more non-working trips. In the case ofwalking–cycling trips, it is interested that elderly people were more willing to do their activities by this sustainable travelmode.

5. Conclusion

This paper has represented structural equation model of travel behavior emphasizing on subjective as well as objectivedeterminants of non-working trips by three modes of transport. Using data collected from individual survey, structural rela-tionships between lifestyle, built environment and non-working trips were studied. The results demonstrate that, lifestyle asa subjective factor which influenced by other variables such as attitudes, land use and socio-economic factors, could be agreat determinant of non-working trips.

In this study, modern-oriented lifestyle (MOL) has significantly influence on individual’s car use. It implies thatnon-working trips by car markedly stem from respondent’s demand for leisure and entertainment activities which root inevolution of society to a new community with tendency to contribute in modern activities (Ecorys, 2008). So, respondentswith modern lifestyle are expected to make frequently more car trips. In our study, this lifestyle which includes leisure activ-ities such as going to a club for exercise, strolling in malls and shopping centers for fun or going to natural gardens, due to thelong distances, people prefer to do them by private car. Although, unlike the study of Collantes and Mokhtarian (2007) whichshows family-oriented lifestyle is related to more short distance trips by car, in our study modern-oriented lifestyle is

516 R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518

associated with long distance trips by car. Based on our studies, Consumption-oriented lifestyle is another determinant ofnon-working trips by cars. Although COL was associated with walking/cycling trips too, it should be considered that con-sumption pattern is an important factor of car usage. This finding could be due to the fact that Persian lifestyle is unsustain-able in many ways and are based on overproduction and overconsumption; so it counts for repetitive car trips to shoppingand service providers.

According to our findings, other two lifestyles (TOL and EOL) result in more non-working trips by sustainable modes suchas public transport and walking/cycling. Traditional-oriented lifestyle which sharply was influenced by residential and travelattitudes, is among the most crucial determinants of non-working trips by walking/cycling. A possible justification behindthis fact is that people with traditional-oriented lifestyle prefer to do leisure activities which require destinations close totheir residential areas such as spending time in parks and green spaces or attending in religious ceremonies. So, it can beconcluded that respondents’ willingness of walking changes inline with the distance of an activity from home; the furtherthey travel, the less they walk (Susilo et al., 2012).

The contribution of built environment as an objective factor to travel behavior was founded significant but less importantthan lifestyle. Non-working trips by car tend to be lower in mixed-land use neighborhoods with well-designed streets whichbenefit from enough internal connectivity. This result is almost similar to findings of previous studies about the impact ofdesign and diversity on mode choices (e.g. Handy, 1996; Cervero and Kockelman, 1997; Ewing and Cervero, 2001 andetc.). Compare to finding of past studies (Ewing and Cervero, 2001; Ewing et al., 2009; Zegras, 2007; Greenwald andBoarnet, 2001), in this paper density has a small but significant impact on car and walking trips, So, moderate density mightlead to making trips by sustainable travel modes. Although, in contrast with lifestyle, the impacts of built environment ontravel behavior is mainly weak.

Apart from lifestyle and built environment, other key variables like attitudes toward travel are recognized as travelbehavior determinants. As we found for travel attitudes, pro-environment attitude is related to fewer car trips and morepublic transport and walking trips. In this study, individuals with more pro-environmental attitudes are more eager touse alternative modes of travel like public transport due to the better understanding of moral norms (Anable andGatersleben, 2005).

Economic–political attitude which is referred to the individuals who were prepare to pay the tax for reducing negativeconsequences of car using and benefit from alternative modes, is directly related to public transport and walking trips. Itreveals that pro-environmental solutions could decrease using of personal vehicle (Mokhtarian et al., 2001; Ory andMokhtarian, 2009). Although, it should be pointed that the way individuals response to a policy which aimed to reducecar travel will depend on the value they allocate to travel utility and the desirability of other travel alternatives. On the otherhand, it is found by the results that people gives positive and negative points to their car travels and alternative modes andthey choose private vehicle when they found that the merits of car use for their travel is more than it’s demerits. So, respon-dents with physical-oriented and individualist attitudes feel that the advantageous of car use is more than related disadvan-tageous. This result has a clear policy implication toward car reduction: advanced and credible public transport systemsencourage people to choose more sustainable travel modes. This approach operates on subjective aspect of individuals’ travelbehaviors.

Contrary to past studies (Van Acker et al., 2014), in this research residential attitudes are slightly associated to travelbehavior. Hence, it can be suggested that in our case study residential self-selection cannot be taken into account.Furthermore, travel behavior can also be regarded as expression of socio-economic factors. The results indicate that financialposition which refers to income and number of cars in the household is associated with more non-working trips by car andless using of public transport.

The present paper seeks to achieve a behavioral approach in travel demand studies; the issue which has been less studiedin Iran. Although this study is largely theoretical in what we are examining individual behavior beyond the prevalent traveldemand models; the concept of our model can bring about practical suggestions. The notable influence of lifestyle on travelbehavior is a new idea which can be regarded to promote travel demand strategies. Even though some studies have partic-ularly mentioned the impact of ‘subjective’ lifestyle on travel behavior (e.g., Scheiner, 2010; Van Acker et al., 2014); more ofthem are conceptualized it mainly sociological’ terms with no presumptive connection to mobility. In this research lifestyleis measured in a way which is closely related to mobility and travel modes.

Since, the lifestyles are based on past and current consumption and production patterns and are tangled with people’severyday choices and practices (Mont, 2007), shaping a sustainable lifestyle could be a complicated but favorable approachto achieve sustainable development. The findings of this study reveal this fact that recognizing the factors that influenceindividuals’ travel behavior in sustainable patterns requires a vast understanding of individuals’ lifestyle contexts and thesystems within which different individuals’ lifestyles perform.

For instance, Consumer-oriented lifestyle is putting too much pressure on natural resources and imposing negative envi-ronmental, economic, social and health impacts. Besides, we have showed that this lifestyle is associated to private car use.So, recognizing consumption patterns and their resulting environmental and social impacts should be a major focus of policyplanners. It is important to understand factors which drive from current individuals’ consumption and the methods that canalter their behaviors. Besides, modern-oriented lifestyle which is usually encountered with leisure trips by roads could becompatible with local options when it comes to traveling for pleasure and entertainments.

Although, it should be noted that car-oriented urbanization make it more difficult for people to have appropriate choicesin their lifestyle patterns. It could be exampled by exorbitance founding to large scale infrastructure systems, expanding

R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518 517

roads and large shopping malls in outskirts of cities that compel individuals to choose car-oriented and consumer-orientedlifestyles. Accordingly, the relationship between lifestyles and individuals trip patterns could be interfered through inappro-priate urban planning. Mutual relation of lifestyle and transport planning is a major issue which should be regarded in futurestudies.

Moreover, it should keep in mind that this analysis is based on cross-section data of lifestyle which has its limitation.Further studies will need longitudinal data collection on lifestyle.

Acknowledgements

The authors gratefully acknowledge Prof. Patricia L. Mokhtarian at Georgia Institute of Technology and Dr. Veronique VanAcker at University of Amsterdam for their helpful guidelines and worthy researches on travel behaviors which lead to paveour study way.

References

Ajzen, I., 1991. The theory of planned behavior. Organ. Behav. Hum. Decis. Process. 50, 179–211.Ajzen, I., 2001. Attitudes: In the Corsini Encyclopedia of Psychology and Behavioral Science. In: Craighead, W.E., Nemeroff, C.B. (Eds.), John Wiley & Sons,

New York.Anable, J., Gatersleben, B., 2005. All work and no play? The role of instrumental and affective factors in work and leisure journeys by different travel modes.

Transp. Res. Part A 39 (2–3), 163–181.Bagley, M.N., Mokhtarian, P.L., 2002. The impact of residential neighborhood type on travel behavior: a structural equations modeling approach. Ann. Reg.

Sci. 36 (2), 279–297.Benedikter, R., Juergensmeyer, M., Anheier, H., 2011. Lifestyles. SAGE Encyclopedia of Global Studies.Boarnet, M., Crane, R., 2001. The influence of land use on travel behavior: specification and estimation strategies. Transp. Res. Part A 35 (9), 823–845.Boarnet, M., Sarmiento, S., 1998. Can land-use policy really affect travel behavior? A study of the link between non-work travel. Urban Stud. 35 (7), 1155–

1169.Cervero, R., 2002. Built environments and mode choice: toward a normative framework. Transp. Res. Part D 7 (4), 265–284.Cervero, R., Kockelman, K., 1997. Travel demand and the 3Ds: density, diversity, and design. Transp. Res. Part D 2 (3), 199–219.Cervero, R., Murakami, J., 2010. Effects of built environments on vehicle miles traveled: evidence from 370 U.S. metropolitan areas. Environ. Plan. Part A 42

(2), 400–418.Chatman, D.G., 2003. How density and mixed uses at the workplace affect personal commercial travel and commute mode choice. Transp. Res. Rec. 1831,

193–201.Chatman, D.G., 2008. Deconstructing development density: quality, quantity and price effects on household non-work travel. Transp. Res. Part A 42 (7),

1008–1030.Collantes, G.O., Mokhtarian, P.L., 2007. Subjective assessments of personal mobility: what makes the difference between a little and a lot? Transp. Policy 14,

181–192.Cooper, J., Ryley, T., Smith, A., 2001. Contemporary lifestyles and the implications for sustainable development policy: lessons from the UK’s most car

dependent city; Belfast Cities 18, 103–113.Crane, R., 2000. The influence of urban form on travel: an interpretive review. J. Plan. Lit. 15 (1), 3–23.Ecorys, 2008. Progress on EU Sustainable Development Strategy. Brussels/Rotterdam.EIA, 2014. U.S. Energy Information Administration, Iran, International Energy Data and Analysis <http://www.eia.gov/beta/international/analysis.cfm?iso=

IRN> (accessed 22.07.14).Etminani Ghasrodashti, R., 2009. The investigation of urban density on car dependency: a case study of three urban districts in Shiraz (Master Thesis) Shiraz

University, Iran.Ewing, R., Cervero, R., 2001. Travel and the built environment: a synthesis. Transp. Res. Rec. 1780, 87–114.Ewing, R., Cervero, R., 2010. Travel and the built environment: a meta-analysis. J. Am. Plan. Assoc. 76 (3), 265–294.Ewing, R., Schroeer, W., Greene, W., 2004. School location and student travel: analysis of factors affecting mode choice. Transp. Res. Rec. 1895, 55–63.Ewing, R., Greenwald, M.J., Zhang, M., Walters, J., Feldman, M., Cervero, R., Thomas, J., 2009. Measuring the Impact of Urban form and Transit Access on

Mixed Use Site Trip Generation Rates. Portland Pilot Study. U.S. Environmental Protection Agency, Washington, DC.Fishbein, M., Ajzen, I., 1972. Attitudes and opinions. Annu. Rev. Psychol., 487–544Frank, L.D., Bradley, M., Kavage, S., Chapman, J., Lawton, K., 2008. Urban form, travel time, and cost relationships with tour complexity and mode choice.

Transportation 35 (1), 37–54Ganzeboom, H., 1988. Leefstijlen in Nederland: Een verkennende studie. Sociaal Cultureel Planbureau, Rijswijk.Greenwald, M.J., Boarnet, M.G., 2001. The built environment as a determinant of walking behavior: analyzing non-work pedestrian travel in Portland,

Oregon. Transp. Res. Rec. 1780, 33–43.Handy, S., 1993. Regional versus local accessibility: implications for nonworking travel. Transp. Res. Rec. 1400, 58–66.Handy, S., 1996. Methodologies for exploring the link between urban form and travel behavior. Transp. Res. Part D 1 (2), 151–165.Handy, S.L., Cao, X., Mokhtarian, P.L., 2005. Correlation or causality between the built environment and travel behavior? Evidence from Northern California.

Transp. Res. D 10 (6), 427–444.Harrington, D., 2009. Confirmatory Factor Analysis, first ed. Oxford University Press, New York.Hildebrand, E.D., 2003. Dimensions in elderly travel behaviour: a simplified activity-based model using lifestyle clusters. Transportation 30, 285–306.Hoyle, R.H., 2012. Handbook of Structural Equation Modeling, first ed. The Guilford press, New York.Iacobucci, D., 2009. Structural equations modeling: fit indices, sample size, and advanced topics. J. Consum. Psychol. 20 (1), 90–98.Kitamura, R., 1988. Formulation of trip generation models using panel data. Transp. Res. Rec. 1203, 60–68.Kitamura, R., 2009. Life-style and travel demand. Transportation 36, 679–710.Lee, J.S., Nam, J., Lee, S.S., 2014. Built environment impacts on individual mode choice: an empirical study of the Houston-Galveston metropolitan area. Int. J.

Sustain. Transp. 8, 447–470.Masoumi, H., 2013. Modeling the travel behavior impacts of micro scale land-use and socio-economic factors. J. Land Use, Mobility Environ. 2, 235–250.Mokhtarian, P.L., Cao, X.Y., 2008. Examining the impacts of residential self selection on travel behaviour: a focus on methodologies. Transp. Res. Part B 42,

204–228.Mokhtarian, P.L., Salomon, I., 2001. How derived is the demand for travel? Some conceptual and measurement considerations. Transp. Res. Part A 35,

695–719.Mokhtarian, P.L., Salomon, I., Redmond, L.S., 2001. Understanding the demand for travel: it’s not purely ‘derived’. Innov.: Eur. J. Soc. Sci. Res. 14, 355–380.Mont, O., 2007. Concept paper for the International Task Force on Sustainable Lifestyles. Paper Presented At Third International Expert Meeting on

Sustainable Consumption and Production, Stockholm.

518 R. Etminani-Ghasrodashti, M. Ardeshiri / Transportation Research Part A 78 (2015) 506–518

Ory, David., Mokhtarian, P., 2009. Modeling the structural relationships among short-distance travel amounts, perceptions, affections, and desires. Transp.Res. Part A 43, 26–43.

Reichman, S., 1977. Instrumental and life style aspects of urban travel behavior. Transp. Res. Rec. 649, 38–42.Salomon, I., Ben-Akiva, M., 1983. The use of the life-style concept in travel-demand models. Environ. Plan. Part A 15, 623–638.Scheiner, J., 2010. Social inequalities in travel behaviour: trip distances in the context of residential self-selection and lifestyles. Transp. Geogr. 18, 679–690.Schwanen, T., Mokhtarian, P.L., 2005. What if you live in the wrong neighborhood? The impact of residential neighborhood dissonance on distance traveled.

Transp. Res. Part D 10, 127–151.Shiraz Municipality Official Website, 2014. <http://www.eshiraz.ir/main/en/index> (accessed 15.10.14).Soltani, A., Esmaeili, Y., 2011. The influence of urban physical form on trip generation, evidence from metropolitan Shiraz, Iran. Indian J. Sci. Technol. 4 (9),

1168–1174.Soltani, A., Etminani Ghasrodashti, R., 2010. The impact of urban density on car dependency, a case study of three residential districts of region 1, Shiraz.

Urb. Reg. Res. J. 5, 139–154.Stead, D., Marshall, S., 2001. The relationships between urban form and travel patterns. An international review and evaluation. Eur. J. Transp. Infrastructure

Res. 1 (2), 113–141.Stern, P.C., Dietz, T., Abel, T., Guagnano, G.A., Kalof, L., 1999. A value-belief-norm theory of support for social movements: the case of environmentalism.

Human Ecol. Rev. 6, 81–97.Susilo, Y., Williams, K., Lindsay, M., Dair, K., 2012. The influence of individuals’ environmental attitudes and urban design features on their travel patterns in

sustainable neighborhoods in the UK. Transp. Res. Part D 17, 190–200.The guardian, 2014. Iran’s Government Steps Up Efforts to Tackle Pollution <http://www.theguardian.com/world/iran-blog/2014/mar/10/irans-

government-steps-up-efforts-to-tackle-pollution> (accessed 03.10.14).Van Acker, V., Mokhtarian, P., Witlox, F., 2014. Car availability explained by the structural relationships between lifestyles, residential location, and

underlying residential and travel attitudes. Transp. Policy 35, 88–99.Van Wee, B., 2002. Land use and transport: research and policy challenges. J. Transp. Geogr. 10, 259–271.Zegras, P.C., 2007. The Built Environment and Motor Vehicle Ownership and Use: Evidence from Santiago de Chile. Paper Presented at the 86th Annual

Meeting of the Transportation Research Board, Washington, DC.