research article activity-trip chaining behavior of urban low...

Research ArticleActivity-Trip Chaining Behavior of Urban Low-IncomePopulations in Nanjing China A Structural Equations Analysis

Zhaoming Chu1 Hui Chen2 Lin Cheng1 Xuewu Chen1 and Senlai Zhu1

1 School of Transportation Southeast University Nanjing 210096 China2 Chengxian College Southeast University Nanjing 210088 China

Correspondence should be addressed to Zhaoming Chu chuzhaoming126com

Received 31 October 2013 Revised 22 March 2014 Accepted 27 March 2014 Published 22 April 2014

Academic Editor Huimin Niu

Copyright copy 2014 Zhaoming Chu et alThis is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

This paper analyzes the activity-trip chaining behavior of urban low-income populations in Nanjing China based on a specifictravel survey of low-income residents of Nanjing city (2010) and the database of residents travel survey of Nanjing city (2009)Individualrsquos information of activity participation and trip chains is extracted from the daily travel diary and matched withindividual and household characteristics On top of correlation analysis and normalization process using the software AMOS twostructural equation models are formulated to analyze the relationship among individualsrsquo sociodemographics activity durationand trip chains of low-income populations and non-low-income populations respectively Seven household characteristics and sixindividual characteristics are chosen as the exogenous variables while 4 indices of activity duration and 4 indices of trip chainsare sleeted as the endogenous variablesThe result shows that the activity-travel behavior of urban low-income populations is quiteunique which offers promising insights into activity-trip chaining behavior of the poor and extends the need to crafting effectivetransportation policies specifically for urban low-income populations in developing countries

1 Introduction

Because of the fast urbanization in developing countrieslarge amount of peasants swarm to the urban area and workthere These peasant-workers together with unemployedcity residents consist the main parts of urban low-incomepopulations in developing countries However in recentyears the soaring house price in big cities forces urban low-income populations move to the urban fringe in developingcountries meanwhile transportation becomes a big problemto these low-income residents How to satisfy the traveldemand of urban low-income residents and how to providecheap and convenient service for them are urgent problemsto be solved The research on travel behavior of low-incomeresidents can help to promote social fairness and justice easethe social conflicts and build a harmonious society

To solve the various transportation problems encoun-tered by the urban low-income populations in developingcountries it is required to capture the characteristics oftheir activity-travel behavior first An activity-based surveymethod was selected because it typically yields higher rates of

trip recall than other methods and is therefore relatively wellsuited to investigate travel behavior in its fuller complexityAs an extremely flexible linear-in-parameters multivariatestatistical modeling technique structural equation modeling(SEM) has been proved to have considerable potential inmodeling activity-based travel demand modeling

The primary intention of this paper is to explore thecharacteristics of activity-trip chaining behavior of urban-low-income populations in developing countries Specificallywe use the structural equations to analyze the correlation ofsociodemographic activity participation and trip chainingbehavior of urban low-income populations and capture themost critical factors that influence the travel behavior ofthe poor by comparing the urban low income populationsand non-low-income populations Our findings are expectedto further add to the rich body of knowledge on activity-based travel demand modeling by focusing on urban low-income populations meanwhile providing useful informa-tion for crafting effective policies to guarantee the cheap andconvenient travel of the poor

Hindawi Publishing CorporationDiscrete Dynamics in Nature and SocietyVolume 2014 Article ID 360269 11 pageshttpdxdoiorg1011552014360269

2 Discrete Dynamics in Nature and Society

The organization of the paper is as follows In thefollowing section we briefly review the relevant literature onthe topic of this study In Section 3 we introduce the principleof structural equation modeling Section 4 describes the dataused in this paper Section 5 presents the development andcalibration of themodels Section 6 presents selected findingsof model estimation results Finally the paper ends withconclusions and future research directions

2 Literature Review

In the literature travel behavior of urban low-income res-idents has not been studied much due to the limited dataSeveral representative studies are listed as follows

Giuliano et al examined the use of public transit by low-incomehouseholds and they declared that public transit is nota reasonable substitute for the private vehicle formost peoplepoor or not poor [1 2] Blumenberg and Haas claimed thatwelfare recipients with unlimited access to automobiles havehigher employment rates and report fewer transportationproblems [3] Clifton presented a few of the challenges facingthose interested in the intersection between poverty andtravel behavior and introduced opportunities to explore low-income travel using some new approaches [4] McDonaldet al found that the free-bus pass program increased low-income studentsrsquo bus ridership and after-school participationthey also found that the increases in bus use were greateramong free-bus pass holders in areas with high levels of busservice and among high school students [5]

The above studies are all based on the data of developedcountries Recently some researches started to focus on thelow-income populations in developing countries Using anactivity diary survey administered in Cape Town a city ofSouth Africa Behrens found that travel occurring by nonmo-torized modes for non-work purposes and during off-peakperiods is considerable They also argued that restricting thefocus of analysis to motorized work and peak period trip-making can create serious misconceptions of the true natureof travel behavior particularly of low-income households[6] Srinivasan and Rogers surveyed 70 households whichwere located in two different parts of Chennai (a city ofIndia) The results indicated that residents in the centrallylocated settlement were more likely to use nonmotorizedtravel modes than the peripherally located residents [7]

More recently researchers paid more attention to themobility of low-income populations Thakuriah et al pro-posed an index of perceived service importance (PSI) toevaluate low-income transit services [8] Taylor et al focusedon the role of the car of the travel choices and needs of low-income households and they concluded that the car clearlyplays an important role in the lives of low-income households[9] Gao and Johnston examined possible impacts of carownership promotion versus transit improvements on jobaccessibility work trips and traveler benefits for low-incomehouseholds [10]

However most of the travel behavior studies pertainingto low-income populations are trip-based and ignored theeffects of activity participation on travel behavior None of

these studies have examined trip chaining behavior of thepoor

After 30 years of development activity-based traveldemand modeling has been widely used in travel behavioranalysis but there are few activity-based travel demandmod-els specifically for low-income populations Meanwhile dueto the availability of improved software structural equationmodeling (SEM) has become an effective tool to modeltravel behavior especially in the field of activity-based traveldemand modeling [11]

Kitamura et al was the first to apply SEM in modelingjoint demand for activity duration and travel Based on theCalifornia time use survey data they confirmed a negativefeedback of commute time to non-work activities [12] Luand Pas described the development estimation and inter-pretation of a model relating sociodemographics activityparticipation and travel behavior Using the structural equa-tion modeling methodology with the endogenous variablesof travel behavior indices a complex set of interrelationshipsamong the variables of interest is estimated simultaneouslyThey found that travel behavior can be better explained byincluding activity participation endogenously in the modelthan through sociodemographics alone [13] Golob estimateda joint model of work and non-work activity duration fourtypes of trip chains and three measures of travel time expen-diture In this model maximum likelihood (ML) estimationwas applied to Portland data and the effects of in-home workand residential accessibility were also explored [14]

More recently using activity-based travel survey datacollected in the Washington DC metropolitan area Kup-pam and Pendyala carried out an exploratory analysis ofcommutersrsquo activity and travel patterns to investigate andestimate relationships among sociodemographics activityparticipation and travel behavior The model estimationresults show that significant trade-offs exist between in-homeand out-of-home activity participation [15] Chung and Ahnused structural equation models to analyze the day-to-dayactivity participation and travel behavior in a developingcountry They confirmed that activity patterns are signifi-cantly different on weekdays and weekends Furthermorethey found that during weekdays there are some day-to-dayvariations in the patterns of activity participation and travelbehavior [16]

Due to the modern and hectic life style travel behaviorof people is becoming complex day by day especially in fastdeveloping countries Therefore the better understandingof trip chain decision making is necessary to transporta-tion researchers and policy makers In the last few yearsresearchers began to study the trip-chaining behavior indeveloping countries [17 18] However to the best of ourknowledge there is still no research specifically for the low-income populations in developing countries who are themain focuses of this study

To sum up previous studies have confirmed that activityparticipation had a significant relationship with travel behav-ior particularly in timeuseHowever lesswork has beendoneto explore the impact of activity participation on trip chain-ing specifically for low-income populations Since previousstudies have primarily focused on households in developed

Discrete Dynamics in Nature and Society 3

courtiers there is a great need to formulate relationships ofsociodemographics activity participation and trip chainingof urban low-income populations in developing countriessuch as China

3 Methodology

In order to estimate a simultaneous model of the inter-relationship among sociodemographics activity durationand trip chaining behavior of urban low-income residentswe applied the methodology of structural equation model(SEM) In addition we are also interested in the directand indirect effects of one variable on another which canbe provided by the estimation result of structure equationmodel

Since all variables used in this research are observedvariables structural equationmodels without latent variablesare therefore reduced to the following form

y = By + Γx + 120577 (1)

where y is a column vector of 119901 endogenous variables x is acolumnvector of 119902 exogenous variablesB is amatrix (119901times119901) ofdirect effects between pairs of 119901 endogenous variables Γ is amatrix (119901times119902) of regression effects associated with exogenousvariables and 120577 is a column vector of the error terms with thestandard assumption that 120577 is uncorrelated with x Furtherwe denote Φ by the covariance matrix of x and Ψ by thecovariance matrix of 120577

Structural equations systems are estimated by covariance-based structural analysis in which the difference betweenthe sample covariance and the model implied covariancematrices is minimized The fundamental hypothesis for thecovariance-based estimation procedures is that the covari-ance matrix of the observed variable is a function of a set ofparameters as shown in the following equation Σ = Σ(120579)where Σ is the population covariance matrix of observedvariables 120579 is a vector that contains the model parametersand Σ(120579) is the covariance matrix written as a function of 120579

The matrix Σ(120579) has three components namely thecovariance matrix of y the covariance matrix of x with y andthe covariance matrix of x Then it can be shown that

Σ (120579) = [Σyy (120579) Σyx (120579)Σxy (120579) Σxx (120579)

]

= [

(Ι minus Β)minus1

(ΓΦΓ1015840

+Ψ) (Ι minus Β)minus1

1015840

(Ι minus Β)minus1

ΓΦ

ΦΓ1015840

(Ι minus Β)minus1

1015840

Φ

]

(2)

The unknown parameters inB ΓΦ andΨ are estimatedso that the implied covariance matrix Σ is as close as possibleto the sample covariance matrix S In order to achieve thisa fitting function F(SΣ(120579)) which is to be minimized isdefinedThefitting function has the properties of being scalargreater than or equal to zero if and only if Σ(120579) = S andcontinuous in S and Σ(120579) [19]

Several methods can be used to estimate the parameterin structural equation model including maximum likeli-hood (ML) unweighted least squares (ULS) generalized

least squares (GLS) and diagonally weighted least squares(DWLS) In this paper we primarily used the ML estimationapproach

4 Data Description

The city selected for this study is Nanjing the capital ofJiangsuProvince ChinaWith a total land area of 6589 squarekilometers and an urban population of over eight million(2013) Nanjing is the second largest commercial center inEast China after Shanghai

There are two sources of data the group of low-incomepopulation is from a specific travel survey of low-incomeresidents of Nanjing City (2010) and the group of non-low-income is from the database of residents travel surveyof Nanjing City (2009) In both surveys all respondentswere asked to record their activity and travel informationwithin one weekday on a travel diary In addition to theactivity and travel information each respondent is requiredto report the usual set of hisher household and personalsociodemographics

In the specific survey of low-income populations (2010)1000 questionnaires were delivered to low-income peopleresiding in three parts of Nanjing City including shanty areasin inner-city (300 copies) welfare-oriented public housingneighborhoods in the edge area of inner-city (200 copies)and the economically affordable housing neighborhoods inurban fringe (500 copies) Then totally 904 questionnairesreturned from all the surveyed areas

The non-low-income group consists of residents whoseannual per-capita income is higher than the minimumsalary threshold of Nanjing CityThus 8666 non-low-incomeresidents are selected from the database of residents travelsurvey of Nanjing City (2009) After eliminatingmissing dataandperforming logic checkingwe selected 846 individuals inthe low-income group and 7534 individuals in the non-low-income group

Based on previous research and single factor analysisof sociodemographic attributes and endogenous variables 7household attributes and 6 individual attributes are selectedas the exogenous variables while the indices of activityparticipation and trip chains are selected as the endogenousvariables In particular the descriptors of activity participa-tion are defined by the duration of four types of activityin-home subsistence maintenance and leisure The tripchaining characteristics are defined by descriptors of 4 itemsnamely number of work chains travel time of work chainsnumber of non-work chains and travel time of non-workchains (see Table 1)

Statistical characteristics of exogenous variables areshown in Tables 2 and 3 It can be found that 675 of low-income residents live in the urban fringe while 601 of thenon-low-income residents live in main urban area Annualhousehold income of low-income populations mainly con-centrates on low groups of 10000sim20000 and 20000sim50000 RMB which take up 326 and 450 respectivelyIn contrast their non-low-income counterparts concentrateon middle-to-high groups of 20000sim50000 and 50000sim


Table 1 Endogenous variables and exogenous variables

Variable Label Notes

Exogenous variables

Household characteristics

Residential location Big zone Main urban area = 1 urban fringe = 2Number of family members 119873 peopleNumber of preschool children 119873 kidAnnual household income IncomeNumber of vehicles 119873 carNumber of bikes 119873 bikeNumber of electric bicycles 119873 ebike

Individual characteristics

Gender Sex Male = 1 female = 2Job Job 9 categoriesTransit IC card holding IC Hold a bus IC card = 1 other = 0Age Age 8 categoriesDriving license holding Lic Hold a driverrsquos license = 1 other = 0Educational level Edu 4 categories

Endogenous variables

Activity duration

In-home activity 119863 119867 Sleeping dinner housework and so forthSubsistence 119863 119878 Work work-related and schoolMaintenance 119863 119872 Obligations and so forthLeisure 119863 119871 Amusement exercise relaxation and so forth

Trip chaining

Number of work chains 119873 119882 Number of work related chains per dayTravel time of work chains 119879 119882 Travel time of work chains per dayNumber of non-work chains 119873 119874 Number of non-work chains per dayTravel time of non-work chains 119879 119874 Travel time of non-work chains per day

Table 2 Statistical characteristics of household characteristics

Variable Low-income group Non-low-income groupCases Valid percent Cumulative percent Cases Valid percent Cumulative percent

Big Zone Inner city 275 325 325 4526 601 601Urban fringe 571 675 1000 3008 399 1000

119873 people

1 15 18 18 39 05 052 104 123 141 1489 198 2033 444 525 665 5076 674 877ge4 283 334 1000 921 123 1000

119873 Kid0 632 747 747 6613 878 8781 200 236 983 892 118 996ge2 14 17 1000 29 04 1000

Income

ltyen10000 38 45 45 0 0 0yen10000simyen20000 276 326 371 0 0 0yen20000simyen50000 381 45 822 3962 526 526yen50000simyen100000 151 178 1000 2605 346 872gtyen100000 0 0 1000 967 128 1000

119873 Car0 692 818 818 5790 769 7691 150 177 995 1640 218 986ge2 4 05 1000 104 14 1000

119873 bike

0 225 266 266 1654 220 221 507 599 865 3315 440 6602 104 123 988 1988 264 923ge3 10 12 1000 577 77 1000

119873 ebike

0 343 405 405 3330 442 4421 347 41 816 3314 44 8822 140 165 981 833 111 992ge3 16 19 1000 57 08 1000


Table 3 Statistical characteristics of individual characteristics

Variable Low-income group Non-low-income groupCases Valid percent Cumulative percent cases Valid percent Cumulative percent

Gender Male 416 492 492 1 497 497Female 430 508 100 2 503 1000

Job

School children 73 86 86 1 113 113College student 15 18 104 2 26 139Factory worker 124 147 251 3 150 288Service staff 114 135 385 4 89 377Civil servant 79 93 479 5 258 635Self-employed 43 51 530 6 58 693

Retired 246 291 820 7 177 870Peasant 51 60 881 8 13 883Others 101 119 1000 9 117 1000

Transit IC Card Yes 722 853 853 1 639 639No 124 147 147 2 361 1000

Age

6sim14 43 51 51 1 66 6615sim19 30 35 86 2 51 11720sim24 75 89 175 3 56 17225sim29 115 136 311 4 88 26130sim39 127 150 461 5 194 45540sim49 110 130 591 6 243 69850sim59 148 175 766 7 186 884ge60 198 234 100 8 116 1000

Driving license Yes 118 139 139 1 276 276No 728 861 1000 2 724 1000

Educational level

Middle school 429 507 507 1 265 265High School 323 382 889 2 385 648

Undergraduate 94 111 1000 3 339 987Graduate 4 13 1000

100000 RMBNote that 182of low-incomehousehold ownat least one car and 139 of the low-income individuals holda driving license which indicates that automobile begin toenter the Chinese urban families even the not so affluentones

Table 4 shows statistical characteristics of the 8 endoge-nous variables that consist of descriptors of activity and tripchaining Note that on average the duration of out-of-homeactivities are less in the low-income group than that of thenon-low-income group The number of trip chains indicatesthat low-income populations generally do less out-of-homeactivities

5 Model Specification

On the basis of activity-based travel demand theory andprevious researches on SEMs a possible structural equationmodeling framework was laid out as shown in Figure 1 whichcaptures the interrelationships among sociodemographicsactivity participation and trip chaining simultaneously

There are three basic assumptions in the initial SEMmod-els First sociodemographics characteristics affect both theactivity participation and travel behavior of travelers Secondthe increase of in-home activity participation will reduce

D S

D H

D M D L

NONW

T W T O

Trip chaining

Socio-demographics

Activity duration

Figure 1 Causal structure linking the exogenous variables andendogenous variables

the time spent on out-of-home activities the three typesof out-of-home activities affect each other mutually Third


Table 4 Statistical characteristics of the endogenous variables

Endogenous variablesLow-income group (846 individuals) Non-low-income group (7534 individuals)

Population Nonzero sample Population Nonzero sampleMean Variance Mean Variance Sample size Mean Variance Mean Variance Sample size

119863 119867 (hour) 1693 403 1693 403 846 1536 342 1534 342 7534119863 119878 (hour) 511 459 886 181 488 636 419 867 197 5531119863 119872 (hour) 031 076 117 102 225 036 097 125 148 2164119863 119871 (hour) 062 137 218 180 238 065 160 255 228 1906119873 119882 (chain) 062 057 108 027 488 085 061 116 037 5531119879 119882 (hour) 064 069 112 054 488 085 085 115 080 5531119873 119874 (chain) 054 066 122 041 373 055 080 142 064 2915119879 119874 (hour) 039 057 088 054 373 044 077 114 086 2915

Table 5 Goodness-of-fit of the two models

Models 120594

2 DF 119875 120594

2DF RMSEA1 GFI2 CN3

Model A 798 88 0722 0907 0000 0991 1175Model B 877 92 0607 0953 0000 0999 99121RMSEA is root mean square error of approximation2GFI is goodness-of-fit index3CN is Hoelterrsquos critical119873

household and individual characteristics not only influencetrip chaining behavior directly but also affect trip chainingindirectly through activity participation of individuals

The above initial SEM models were estimated by usingthe software of AMOS 70 The maximum likelihood (ML)method was selected as the estimation method because itconverges more rapidly and the results are also easier tointerpret compared with the ldquodistribution freerdquo approach(eg DWLS) [14] Generally the initial model does notperform well thus it needs some modification by adding ordeleting links according to both their significance which issuggested by themodel output and their interpretability Afterthe modification procedures we obtained two final modelsas shown in Figures 2 and 3

Table 5 listed goodness-of-fit of the two models ForModel A which represents the low-income group the 1205942 is798 with 88 degrees of freedom and 119875 value is 0722 (greaterthan 005) indicating that the null hypothesis (119867

0

Σ =

Σ(120579)) cannot be rejected Other measures of fit such as GFI= 0991 (that ranges from 0 to 1) and root mean squareerror of approximation (RMSEA = 0000) are also found tobe acceptable by model fit criteria for structural equationmodel Hoelterrsquos critical 119873 (CN) statistic is found to be 1175(greater than 200 is considered a goodfit) which is the samplesize at which value of the fitting function 119865ML would leadto the rejection of the null hypothesis 119867

0

(ie Σ = Σ(120579))at a chosen significance level Similarly Model B which ispertaining to the nonpoor is also quite satisfactory

6 Model Estimation Results

Tables 6ndash10 are the estimation results of Model A and ModelB There are three distinct types of relationships that canbe obtained from structural equations modeling procedures

Income

Sex

Job

Edu

Lic

Ic

1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

00

0

0

0

0

0

00

00

00

00

0

00

Big_zone

0

0

N_kid

00

0

0

N_people

00

1

D_H

N_W

N_O

D_S

D_M

D_L

T_W

T_O

e1

e2

e3

e4

e5

e6

e7

e8

Figure 2 SEM path diagram for low-income group

They are called direct effects indirect effects and total effectsrespectively Note that direct and indirect effects may be ofdifferent signs thus having an important implication for theoverall total effect For example it can be seen in Table 10(Model A) that the subsistence activity duration (D S) hasa negative direct effect (minus0121) and positive indirect effect(0149) on the travel time of work chains (T W) BecauseD S has negative direct effects on D M and D L (eg minus0090and minus0418 resp) both of which have negative direct effects


Table 6 Total direct and indirect effects of sociodemographics on activity duration and trip chaining in Model A

Effects Big Zone 119873 people 119873 kid Income 119873 car 119873 bike 119873 Ebike Sex Job IC Age Lic Edu

119863 119867

Total minus0833 0 0504 0 0 0 minus0372 1245 0257 0 0554 0 minus1342Direct minus0883 0 0504 0 0 0 minus0372 1245 0257 0 0554 0 minus1342Indirect 0 0 0 0 0 0 0 0 0 0 0 0 0

119863 119878

Total 0747 0 0030 minus0270 0620 0245 0731 minus1354 minus0217 0 minus0922 0 1136Direct 0 0 0457 minus0270 0620 0245 0416 minus0300 0 0 minus0453 0 0Indirect 0747 0 minus0427 0 0 0 0315 minus1054 minus0217 0 minus0469 0 1136

119863 119872

Total minus0049 0 0084 0024 minus0056 0055 minus0058 0305 0037 0 0072 0 minus0075Direct 0 0 0097 0 0 0077 0 0209 0023 0 0 0 0Indirect minus0049 0 minus0013 0024 minus0056 minus0022 minus0058 0096 0014 0 minus0072 0 minus0075

119863 119871

Total 0003 0 minus0143 0045 minus0216 minus0144 minus0144 minus0060 0033 0 0229 0 minus0197Direct 0 0 0093 minus0050 0 0 0 0 0051 0 0073 0 minus0202Indirect 0003 0 minus0236 0094 minus0216 minus0144 minus0144 minus0060 minus0018 0 0156 0 0005

119873 119882

Total 0133 0 minus0037 0021 0042 0005 0052 minus0131 minus0031 0 minus0119 minus0087 0099Direct 0077 0 0 0028 0 0 0 0 0 0 minus0037 minus0087 minus0032Indirect 0056 0 minus0037 minus0007 0042 0005 0052 minus0131 minus0031 0 minus0082 0 0131

119879 119882

Total 0197 0 minus0081 0065 minus0069 minus0046 0017 minus0153 minus0037 minus0068 minus0101 0013 0263Direct 0090 0 0 0045 minus0086 minus0041 minus0029 0 0008 minus0068 0 0042 0080Indirect 0107 0 minus0081 0020 0017 minus0005 0045 minus0153 minus0045 0 minus0101 minus0029 0184

119873 119874

Total minus0038 0020 0032 0031 minus0065 minus0007 minus0073 0136 0038 0008 0128 0023 minus0155Direct 0059 0020 0 0027 0 0 0 minus0038 0 0 0 0 0Indirect minus0097 0 0032 0004 minus0065 minus0007 minus0073 0174 0038 0008 0128 0023 minus0155

119879 119874

Total minus0049 minus0002 minus0023 0011 minus0051 minus0006 minus0059 0107 0023 0021 0117 0018 minus0112Direct 0018 minus0011 minus0022 0 0 0 0 0 0 0 0009 0 0Indirect minus0067 0009 minus0001 0011 minus0051 minus0006 minus0059 0107 0023 0021 0108 0018 minus0112

(minus0312 minus0281) on T W According to the effect analysistheory the indirect effects ofD S onT W can be computed as(minus0090)times (minus0312)+ (minus0418)times (minus0281) = 0149 Thereforethe total effect (0028) of D S on T W is the algebraic sum ofdirect effect (minus0121) and indirect effect (0149)

It is can be found that strong relationship exists amongthe sociodemographics activity participation and travelbehavior both for the poor and the nonpoor In the followingwe will examine the effects in detail from 4 aspects effectsof sociodemographics on activity duration and trip chainingeffects of activity durations on each other effects of trip-chaining on trip chaining and effects of activity duration ontrip chaining behavior

61 Effects of Sociodemographics on Activity Duration and TripChaining From Tables 6 and 7 we can see that in bothgroups some sociodemographics significantly affect all fourtypes of activities and four trip chaining variablesThe house-hold and individual characteristics that are systematicallyimportant in explaining variations in activity participationand travel behavior include house location income numberof preschool children age gender and educational level

Combining the path diagram in Figures 2 and 3 it can alsobe found that household characteristics have more influenceon the activity participation of low-income population (12routes from household characteristics to activity durationsand 11 routes from individual characteristics to activity dura-tions) while individual characteristics have more influence

on the activity participation of the nonpoor (7 routes fromhousehold characteristics to activity durations and 17 routesfrom individual characteristics to activity durations) Inaddition sociodemographics have more direct influence (22routes) on the trip-chaining in the low-income group thanthat of the nonpoor (17 routes)

Specifically the number of preschool children signifi-cantly affects the activity duration of the low-income groupbut it has no effects on that of the non-low-income groupOn the contrary the IC factor does not influence low-incomepopulationsrsquo activity duration at all but has significant effectson non-income populations

62 Effects of Activity Duration on Activity Duration FromTable 8 it can be found that interaction effects among 4activity durations follow the same framework both in ModelA and Model B D H has negative direct effects on theduration of out-of-home activities D M has negative effectson D M and D L and D M has negative effects on D L

However the values of effects are not quite similar in thetwomodels For example the absolute values of effects ofD Hon other activity durations in Model A are all smaller thanthose inmodel B while the effects ofD M onD L inModel Aare larger than their counterparts inmodel B which indicatesthat the trade-off among the 4 type activities is differently intwo groups It can be interpreted that low-income populationspend more time at home and have lower value of time dueto their inferior social status and limited social network


Table 7 Total direct and indirect effects of sociodemographics on activity duration and trip chains in Model B


119863 119867

Total 0 0 0 0103 0 0 0 minus1158 0 minus0285 019 0 0Direct 0 0 0 0103 0 0 0 minus1158 0 minus0285 019 0 0Indirect 0 0 0 0 0 0 0 0 0 0 0 0 0

119863 119878

Total 0424 minus0119 0 minus0273 0159 0 0 1365 0054 0547 minus0189 0 0Direct 0424 minus0119 0 minus0171 0159 0 0 0213 0054 0263 0 0 0Indirect 0 0 0 minus0102 0 0 0 1152 0 0283 minus0189 0 0

119863 119872

Total minus0009 0017 0 0093 minus0022 0017 0 minus0247 minus0029 minus0067 minus0003 0 0Direct 005 0 0 0058 0 0017 0 minus0097 minus0021 0 minus0023 0 0Indirect minus0059 0017 0 0035 minus0022 0 0 minus015 minus0008 minus0067 002 0 0

119863 119871

Total minus0201 0096 0 0149 minus0068 minus0007 0 minus0226 minus0053 minus0131 minus0016 0 minus013Direct 0 0045 0 0092 0 0 0 minus0097 minus0038 0 minus0038 0 minus013Indirect minus0201 0051 0 0057 minus0068 minus0007 0 minus0129 minus0015 minus0131 0022 0 0

119873 119882

Total 0127 minus0018 minus0031 0012 0026 minus0003 0 0138 minus0005 0078 minus0016 minus0031 minus0016Direct 0105 0 minus0031 0 0 0 0 0 minus0021 0038 minus0014 minus0031 0035Indirect 0022 minus0018 0 0012 0026 minus0003 0 0138 0016 004 minus0002 0 minus005

119879 119882

Total minus0044 minus0019 minus0009 0002 0049 0005 minus0008 0243 0014 0059 minus0014 minus0009 minus0037Direct minus0069 0 0 0 0 0009 minus0008 0 0 0 0 0 002Indirect 0025 minus0019 minus0009 0002 0049 minus0004 0 0243 0014 0059 minus0014 minus0009 minus0057

119873 119874

Total minus0052 0034 0005 minus0027 minus0039 0005 0007 minus0217 minus0024 minus0081 minus001 0005 0096Direct 0019 0 0 0 0 0 0006 0 0 0 minus0014 0 0017Indirect minus0071 0034 0005 minus0027 minus0039 0005 0 minus0217 minus0024 minus0081 0004 0005 0079

119879 119874

Total minus0099 0034 0007 minus003 minus0033 0002 0006 minus0259 minus002 minus0084 minus0005 0007 0077Direct minus0036 0 0 0 0 0 0 minus0069 0 0 0 0 0Indirect minus0063 0034 0007 minus003 minus0033 0002 0006 minus019 minus002 minus0084 minus0005 0007 0077

Table 8 Total direct and indirect effects of activity duration on activity duration

Effects Model A Model B119863 119867 119863 119878 119863 119872 119863 119871 119863 119867 119863 119878 119863 119872 119863 119871

119863 119867

Total 0 0 0 0 0 0 0 0Direct 0 0 0 0 0 0 0 0Indirect 0 0 0 0 0 0 0 0

119863 119878

Total minus0847 0 0 0 minus0995 0 0 0Direct minus0847 0 0 0 minus0995 0 0 0Indirect 0 0 0 0 0 0 0 0

119863 119872

Total 0056 minus0090 0 0 0104 minus0141 0 0Direct minus0021 minus0090 0 0 minus0036 minus0141 0 0Indirect 0076 0 0 0 0140 0 0 0

119863 119871

Total minus0004 minus0349 minus0765 0 0066 minus0427 minus0400 0Direct minus0315 minus0418 minus0765 0 minus0374 minus0484 minus0400 0Indirect 0311 minus0069 0 0 0440 0056 0 0

63 Effects of Trip Chaining on Trip Chaining Accordingto Table 9 the effects of trip-chaining characteristics oneach other also follow similar frameworks in both modelsSpecifically N W has positive effects on T W and negativeeffects on both N O and T O T W has negative effects onN O and T O and N O has negative effects on T O whichindicates that there are strong relationships and trade-offsbetween work chains and non-work chains

Note that the absolute value of effects of work chainson non-work chains of the poor is larger than that of the

nonpoor It can be explained that low-income residents haveless freedom to participate in different types of activities otherthan work due to their economic status

64 Effects of Activity Duration on Trip Chaining The esti-mation results in Table 10 show that both for the poorand nonpoor travel is derived from activity participationActivity duration also affects trip chaining behavior besidessociodemographics For example we find that number ofwork chains (N W) and the travel time of work chains (T W)


Table 9 Total direct and indirect effects of trip chaining on trip chaining


119873 119882


119879 119882


119873 119874

Total minus0318 minus0113 0 0 minus0167 minus0056 0 0Direct minus0281 minus0113 0 0 minus0150 minus0056 0 0Indirect minus0037 0 0 0 minus0017 0 0 0

119879 119874

Total minus0349 minus0302 0433 0 minus0237 minus0274 0600 0Direct minus0127 minus0253 0433 0 minus0063 minus0241 0600 0Indirect minus0222 minus0049 0 0 minus0174 minus0033 0 0

Table 10 Total direct and indirect effects of activity duration on trip chaining


119873 119882

Total minus0063 0067 minus0151 minus0226 minus0072 0076 minus0197 minus0201Direct minus0081 minus0041 minus0324 minus0226 minus0079 minus0049 minus0278 minus0201Indirect 0018 0108 0173 0 0007 0125 0081 0

119879 119882

Total minus0091 0028 minus0148 minus0356 minus0154 0014 minus0256 minus0376Direct minus0157 minus0121 minus0312 minus0281 minus0266 minus0189 minus0320 minus0314Indirect 0065 0149 0164 minus0076 0111 0203 0064 minus0062

119873 119874

Total 0071 minus0120 0237 0300 0105 minus0172 0311 0303Direct 0025 0 0329 0197 0030 0 0368 0252Indirect 0046 minus0120 minus0091 0103 0075 minus0172 minus0057 0051

119879 119874

Total 0043 minus0124 0159 0207 0086 minus0191 0259 0258Direct minus0080 minus0074 minus0031 minus0041 minus0108 minus0093 minus0013 minus0027Indirect 0123 minus0050 minus0191 0249 0194 minus0098 0272 0285

are directly affected by the 4 categories of activities We alsofind that the N W increases as the D M or D L decreaseswhile N O increases as D M or D L increases

From Table 10 we can also find that in different modelsthe effects between the same variables are different Forexample the absolute value of total effect of D M on N W issmaller in the low-income group than in the non-low-incomegroup however the corresponding absolute values of directeffect and indirect effect are larger in Model A than that inModel B

In general the absolute values of effects in Model Bare greater than their counterparts in Model A whichindicates that activity participation has larger effects on thetrip chaining characteristics in the group of non-low-incomepopulation than that of the poor This scenario is possiblybecause the sample size of the low-income group (846) ismuch less than that of the non-low-income group (7534)thus on the whole the causal relationship of activity durationand trip chaining behavior revealed in Model B is strongerthan that in Model A

7 Conclusions

This paper focuses on the activity-trip chaining behaviorof urban low-income populations in developing countriesUsing the data of residents travel survey of Nanjing City(2009) and a specific travel survey of low-income residentsof Nanjing City (2010) we proposes two structural equationmodels to investigate the relationships among sociodemo-graphics activity participation and travel behavior of bothlow-income populations and non-low-income populations ofNanjing City Based on the model outputs we analyzed fourcategories of effects of the two groups The general findingscan be summarized as follows

First on average the duration of out-of-home activitiestaken by the low-income populations is less than that of thenon-low-income populations and the less trip chains andless total travel time indicate that low-income populationsgenerally do less out-of-home activities

Second the relationships among sociodemographicsactivity duration and trip chaining of both groups can becaptured by the proposed SEM models and most of the


Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8

Figure 3 SEM path diagram for non-low-income group

estimated effects are quite similar to those reported in theliterature

Third both the structural equation models follow thesamemodeling frameworkTherefore the activity-trip chain-ing behavior of both the low-income populations and non-low-income populations shares some similarities For exam-ple sociodemographics especially household income res-idential locations age and gender significantly affects theactivity-trip chaining behavior of both the poor and thenonpoor

Finally low-income populations have some unique char-acteristics on the activity-travel behavior which are differentfrom those of the non-low-income populations For instancehousehold characteristics have more influence on the activityparticipation of low-income population the trade-off amongthe four type activities is differently in two groups the effectsof work chains on non-work chains of the poor are largerthan those of the nonpoor in general activity participationhas greater effects on trip chaining in the group of non-low-income residents than that of low-income residents

Based on these findings of travel behavior characteristicsof urban low-income populations in developing countriesthe following policies are suggested for the government andtransportation agencies

(1) Adopt transit-oriented transportation planning strat-egy such as adding new shuttle buses from low-income population concentrated areas to metro sta-tions and opening new bus lines across low-incomeneighborhoods step by step

(2) In order to reduce the monetary cost of low-incomeresidents the government can either subsidize them

directly to improve social equity or introduce twoor more bus operating companies to break themonopoly so as to improve the level of bus serviceand reduce the bus fares

(3) In the long-term planning the city should transformfrom single center pattern to polycentric developmentpattern Meanwhile the government should considerhybrid land use and create more job opportunitiesnear the residential area of low-income populationssuch that low-income residents in the urban fringewill not waste two much time on their trip chains

(4) Provide more vocational training for low-incomeadults and improve their ability of earning moneyIn addition guarantee the next generation of low-income residents can receive high quality educationand help them climb higher along the social ladderThese policies can change their inferior position oftravel fundamentally

This research offers promising insights into the activity-travel behavior of the poor and extends the need to craftingeffective transportation policies specifically for the urbanlow-income populations in developing countries Howeverthis research can be extended in terms of the followingaspects (a) conduct specific studies on the trading-off rela-tionships between in-home and out-of-home activities (b)study the interactions between activity participation andtravel chaining behavior on two or more successive days (c)consider the household level activity-travel behavior charac-teristics instead of individual level (d) adopt the proposedSEMmodel to other cities in developing countries It is hopedthat these issues and others can be addressed in the future

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

This research is supported by the National Natural ScienceFoundation of China (NSFC no 51078085 51178109 5117811051378119) Graduate Innovation Project of Jiangsu Province(No CXZZ12 0113) and the Fundamental Research Funds forthe Central Universities China The authors would like toexpress their appreciation towards Nanjing Institute of Cityamp Transport Planning Co ltd in particular for the valuableassistance in obtaining and interpreting the data used forthese models

References

[1] G Giuliano H Hu and K Lee ldquoThe role of public transit inthe mobility of low income householdsrdquo Final Report MetransTransportation Center Los Angeles Calif USA 2001 httpwwwamericandreamcoalitionorgautomobilitytransitfor-poorpdf

[2] G Giuliano ldquoLow income public transit and mobilityrdquo Trans-portation Research Record no 1927 pp 63ndash70 2005


[3] E Blumenberg and P Haas ldquoThe travel behavior and needsof the poor a study of welfare recipients in Fresno CountyrdquoPublication FHWA-CA-OR-2001-23 FHWA US Departmentof Transportation 2001

[4] K Clifton ldquoExamining travel choices of low-income popula-tionsmdashissues methods and new approachesrdquo in Proceedings ofthe 10th International Conference on Travel Behavior ResearchLucerne Switzerland August 2003

[5] N McDonald S Librera and E Deakin ldquoFree transit forlow-income youth experience in San Francisco Bay areaCaliforniardquo Transportation Research Record no 1887 pp 153ndash160 2004

[6] R Behrens ldquoUnderstanding travel needs of the poor Towardsimproved travel analysis practices in South Africardquo TransportReviews vol 24 no 3 pp 317ndash336 2004

[7] S Srinivasan and P Rogers ldquoTravel behavior of low-incomeresidents studying two contrasting locations in the city ofChennai Indiardquo Journal of Transport Geography vol 13 no 3pp 265ndash274 2005

[8] P Thakuriah P S Sriraj S Soot and Y Liao ldquoDeterminantsof perceived importance of targeted transportation services forlow-income ridersrdquo Transportation Research Record no 1986pp 145ndash153 2006

[9] J Taylor M Barnard H Neil and C Creegan The TravelChoices and Needs of Low Income Households The Role of theCar The National Centre for Social Research London UK2009 httptridtrborgviewaspxid=886473

[10] S Gao and R A Johnston ldquoPublic versus private mobility forlow-income households transit improvements versus increasedcar ownership in the sacramento California regionrdquo Trans-portation Research Record no 2125 pp 9ndash15 2009

[11] T F Golob ldquoStructural equation modeling for travel behaviorresearchrdquoTransportation Research BMethodological vol 37 no1 pp 1ndash25 2003

[12] R Kitamura J P Robinson T F Golob M A Bradley JLeonard and T van der Hoorn ldquoA comparative analysis of timeuse data in theNetherlands andCaliforniardquo in Proceedings of the20th PTRC Summer Annual Meeting Transportation PlanningMethods pp 127ndash138 1992

[13] X Lu and E I Pas ldquoSocio-demographics activity participationand travel behaviorrdquo Transportation Research A Policy andPractice vol 33 no 1 pp 1ndash18 1999

[14] T F Golob ldquoA simultaneous model of household activity par-ticipation and trip chain generationrdquo Transportation ResearchB Methodological vol 34 no 5 pp 355ndash376 2000

[15] A R Kuppam andRM Pendyala ldquoA structural equations anal-ysis of commutersrsquo activity and travel patternsrdquo Transportationvol 28 no 1 pp 33ndash54 2001

[16] J-H Chung and Y Ahn ldquoStructural equation models of day-to-day activity participation and travel behavior in a developingcountryrdquo Transportation Research Record no 1807 pp 109ndash1182002

[17] M Yang W Wang X Chen T Wan and R Xu ldquoEmpiricalanalysis of commute trip chaining case study of ShangyuChinardquo Transportation Research Record no 2038 pp 139ndash1472007

[18] S S V Subbarao andKV Krishna Rao ldquoTrip chaining behaviorin developing countries a study of Mumbai MetropolitanRegion Indiardquo European Transport paper 3 no 53 pp 1ndash302013

[19] M Yang W Wang G Ren R Fan B Qi and X ChenldquoStructural equation model to analyze sociodemographicsactivity participation and trip chaining between householdheads survey of Shangyu Chinardquo Transportation ResearchRecord no 2157 pp 38ndash45 2010

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

MathematicsJournal of


Mathematical Problems in Engineering

Hindawi Publishing Corporationhttpwwwhindawicom

Differential EquationsInternational Journal of

Volume 2014

Applied MathematicsJournal of


Probability and StatisticsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of


Mathematical PhysicsAdvances in

Complex AnalysisJournal of


OptimizationJournal of


CombinatoricsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of


Operations ResearchAdvances in

Journal of


Function Spaces

Abstract and Applied AnalysisHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of Mathematics and Mathematical Sciences


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Algebra

Discrete Dynamics in Nature and Society



Decision SciencesAdvances in

Discrete MathematicsJournal of


Volume 2014 Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Stochastic AnalysisInternational Journal of


The organization of the paper is as follows In thefollowing section we briefly review the relevant literature onthe topic of this study In Section 3 we introduce the principleof structural equation modeling Section 4 describes the dataused in this paper Section 5 presents the development andcalibration of themodels Section 6 presents selected findingsof model estimation results Finally the paper ends withconclusions and future research directions

2 Literature Review

In the literature travel behavior of urban low-income res-idents has not been studied much due to the limited dataSeveral representative studies are listed as follows

Giuliano et al examined the use of public transit by low-incomehouseholds and they declared that public transit is nota reasonable substitute for the private vehicle formost peoplepoor or not poor [1 2] Blumenberg and Haas claimed thatwelfare recipients with unlimited access to automobiles havehigher employment rates and report fewer transportationproblems [3] Clifton presented a few of the challenges facingthose interested in the intersection between poverty andtravel behavior and introduced opportunities to explore low-income travel using some new approaches [4] McDonaldet al found that the free-bus pass program increased low-income studentsrsquo bus ridership and after-school participationthey also found that the increases in bus use were greateramong free-bus pass holders in areas with high levels of busservice and among high school students [5]

The above studies are all based on the data of developedcountries Recently some researches started to focus on thelow-income populations in developing countries Using anactivity diary survey administered in Cape Town a city ofSouth Africa Behrens found that travel occurring by nonmo-torized modes for non-work purposes and during off-peakperiods is considerable They also argued that restricting thefocus of analysis to motorized work and peak period trip-making can create serious misconceptions of the true natureof travel behavior particularly of low-income households[6] Srinivasan and Rogers surveyed 70 households whichwere located in two different parts of Chennai (a city ofIndia) The results indicated that residents in the centrallylocated settlement were more likely to use nonmotorizedtravel modes than the peripherally located residents [7]

More recently researchers paid more attention to themobility of low-income populations Thakuriah et al pro-posed an index of perceived service importance (PSI) toevaluate low-income transit services [8] Taylor et al focusedon the role of the car of the travel choices and needs of low-income households and they concluded that the car clearlyplays an important role in the lives of low-income households[9] Gao and Johnston examined possible impacts of carownership promotion versus transit improvements on jobaccessibility work trips and traveler benefits for low-incomehouseholds [10]

However most of the travel behavior studies pertainingto low-income populations are trip-based and ignored theeffects of activity participation on travel behavior None of

these studies have examined trip chaining behavior of thepoor

After 30 years of development activity-based traveldemand modeling has been widely used in travel behavioranalysis but there are few activity-based travel demandmod-els specifically for low-income populations Meanwhile dueto the availability of improved software structural equationmodeling (SEM) has become an effective tool to modeltravel behavior especially in the field of activity-based traveldemand modeling [11]

Kitamura et al was the first to apply SEM in modelingjoint demand for activity duration and travel Based on theCalifornia time use survey data they confirmed a negativefeedback of commute time to non-work activities [12] Luand Pas described the development estimation and inter-pretation of a model relating sociodemographics activityparticipation and travel behavior Using the structural equa-tion modeling methodology with the endogenous variablesof travel behavior indices a complex set of interrelationshipsamong the variables of interest is estimated simultaneouslyThey found that travel behavior can be better explained byincluding activity participation endogenously in the modelthan through sociodemographics alone [13] Golob estimateda joint model of work and non-work activity duration fourtypes of trip chains and three measures of travel time expen-diture In this model maximum likelihood (ML) estimationwas applied to Portland data and the effects of in-home workand residential accessibility were also explored [14]

More recently using activity-based travel survey datacollected in the Washington DC metropolitan area Kup-pam and Pendyala carried out an exploratory analysis ofcommutersrsquo activity and travel patterns to investigate andestimate relationships among sociodemographics activityparticipation and travel behavior The model estimationresults show that significant trade-offs exist between in-homeand out-of-home activity participation [15] Chung and Ahnused structural equation models to analyze the day-to-dayactivity participation and travel behavior in a developingcountry They confirmed that activity patterns are signifi-cantly different on weekdays and weekends Furthermorethey found that during weekdays there are some day-to-dayvariations in the patterns of activity participation and travelbehavior [16]

Due to the modern and hectic life style travel behaviorof people is becoming complex day by day especially in fastdeveloping countries Therefore the better understandingof trip chain decision making is necessary to transporta-tion researchers and policy makers In the last few yearsresearchers began to study the trip-chaining behavior indeveloping countries [17 18] However to the best of ourknowledge there is still no research specifically for the low-income populations in developing countries who are themain focuses of this study

To sum up previous studies have confirmed that activityparticipation had a significant relationship with travel behav-ior particularly in timeuseHowever lesswork has beendoneto explore the impact of activity participation on trip chain-ing specifically for low-income populations Since previousstudies have primarily focused on households in developed



3 Methodology



y = By + Γx + 120577 (1)




Σ (120579) = [Σyy (120579) Σyx (120579)Σxy (120579) Σxx (120579)

]

= [

(Ι minus Β)minus1

(ΓΦΓ1015840


1015840

(Ι minus Β)minus1

ΓΦ

ΦΓ1015840

(Ι minus Β)minus1

1015840

Φ

]

(2)




4 Data Description










Exogenous variables






Activity duration


Trip chaining





119873 people

1 15 18 18 39 05 052 104 123 141 1489 198 2033 444 525 665 5076 674 877ge4 283 334 1000 921 123 1000

119873 Kid0 632 747 747 6613 878 8781 200 236 983 892 118 996ge2 14 17 1000 29 04 1000

Income


119873 Car0 692 818 818 5790 769 7691 150 177 995 1640 218 986ge2 4 05 1000 104 14 1000

119873 bike

0 225 266 266 1654 220 221 507 599 865 3315 440 6602 104 123 988 1988 264 923ge3 10 12 1000 577 77 1000

119873 ebike

0 343 405 405 3330 442 4421 347 41 816 3314 44 8822 140 165 981 833 111 992ge3 16 19 1000 57 08 1000





Job




Age



Educational level








D S

D H

D M D L

NONW

T W T O

Trip chaining

Socio-demographics

Activity duration









Models 120594

2 DF 119875 120594

2DF RMSEA1 GFI2 CN3





0

Σ =


0




Income

Sex

Job

Edu

Lic

Ic

1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

00

0

0

0

0

0

00

00

00

00

0

00

Big_zone

0

0

N_kid

00

0

0

N_people

00

1

D_H

N_W

N_O

D_S

D_M

D_L

T_W

T_O

e1

e2

e3

e4

e5

e6

e7

e8






119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874













119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874




119863 119867


119863 119878


119863 119872


119863 119871









119873 119882


119879 119882


119873 119874


119879 119874




119873 119882


119879 119882


119873 119874


119879 119874





7 Conclusions





Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8














Acknowledgments


References




























Volume 2014




Journal of











Journal of


Function Spaces






Algebra











3 Methodology



y = By + Γx + 120577 (1)




Σ (120579) = [Σyy (120579) Σyx (120579)Σxy (120579) Σxx (120579)

]

= [

(Ι minus Β)minus1

(ΓΦΓ1015840


1015840

(Ι minus Β)minus1

ΓΦ

ΦΓ1015840

(Ι minus Β)minus1

1015840

Φ

]

(2)




4 Data Description










Exogenous variables






Activity duration


Trip chaining





119873 people

1 15 18 18 39 05 052 104 123 141 1489 198 2033 444 525 665 5076 674 877ge4 283 334 1000 921 123 1000

119873 Kid0 632 747 747 6613 878 8781 200 236 983 892 118 996ge2 14 17 1000 29 04 1000

Income


119873 Car0 692 818 818 5790 769 7691 150 177 995 1640 218 986ge2 4 05 1000 104 14 1000

119873 bike

0 225 266 266 1654 220 221 507 599 865 3315 440 6602 104 123 988 1988 264 923ge3 10 12 1000 577 77 1000

119873 ebike

0 343 405 405 3330 442 4421 347 41 816 3314 44 8822 140 165 981 833 111 992ge3 16 19 1000 57 08 1000





Job




Age



Educational level








D S

D H

D M D L

NONW

T W T O

Trip chaining

Socio-demographics

Activity duration









Models 120594

2 DF 119875 120594

2DF RMSEA1 GFI2 CN3





0

Σ =


0




Income

Sex

Job

Edu

Lic

Ic

1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

00

0

0

0

0

0

00

00

00

00

0

00

Big_zone

0

0

N_kid

00

0

0

N_people

00

1

D_H

N_W

N_O

D_S

D_M

D_L

T_W

T_O

e1

e2

e3

e4

e5

e6

e7

e8






119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874













119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874




119863 119867


119863 119878


119863 119872


119863 119871









119873 119882


119879 119882


119873 119874


119879 119874




119873 119882


119879 119882


119873 119874


119879 119874





7 Conclusions





Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8














Acknowledgments


References




























Volume 2014




Journal of











Journal of


Function Spaces






Algebra












Exogenous variables






Activity duration


Trip chaining





119873 people

1 15 18 18 39 05 052 104 123 141 1489 198 2033 444 525 665 5076 674 877ge4 283 334 1000 921 123 1000

119873 Kid0 632 747 747 6613 878 8781 200 236 983 892 118 996ge2 14 17 1000 29 04 1000

Income


119873 Car0 692 818 818 5790 769 7691 150 177 995 1640 218 986ge2 4 05 1000 104 14 1000

119873 bike

0 225 266 266 1654 220 221 507 599 865 3315 440 6602 104 123 988 1988 264 923ge3 10 12 1000 577 77 1000

119873 ebike

0 343 405 405 3330 442 4421 347 41 816 3314 44 8822 140 165 981 833 111 992ge3 16 19 1000 57 08 1000





Job




Age



Educational level








D S

D H

D M D L

NONW

T W T O

Trip chaining

Socio-demographics

Activity duration









Models 120594

2 DF 119875 120594

2DF RMSEA1 GFI2 CN3





0

Σ =


0




Income

Sex

Job

Edu

Lic

Ic

1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

00

0

0

0

0

0

00

00

00

00

0

00

Big_zone

0

0

N_kid

00

0

0

N_people

00

1

D_H

N_W

N_O

D_S

D_M

D_L

T_W

T_O

e1

e2

e3

e4

e5

e6

e7

e8






119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874













119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874




119863 119867


119863 119878


119863 119872


119863 119871









119873 119882


119879 119882


119873 119874


119879 119874




119873 119882


119879 119882


119873 119874


119879 119874





7 Conclusions





Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8














Acknowledgments


References




























Volume 2014




Journal of











Journal of


Function Spaces






Algebra













Job




Age



Educational level








D S

D H

D M D L

NONW

T W T O

Trip chaining

Socio-demographics

Activity duration









Models 120594

2 DF 119875 120594

2DF RMSEA1 GFI2 CN3





0

Σ =


0




Income

Sex

Job

Edu

Lic

Ic

1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

00

0

0

0

0

0

00

00

00

00

0

00

Big_zone

0

0

N_kid

00

0

0

N_people

00

1

D_H

N_W

N_O

D_S

D_M

D_L

T_W

T_O

e1

e2

e3

e4

e5

e6

e7

e8






119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874













119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874




119863 119867


119863 119878


119863 119872


119863 119871









119873 119882


119879 119882


119873 119874


119879 119874




119873 119882


119879 119882


119873 119874


119879 119874





7 Conclusions





Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8














Acknowledgments


References




























Volume 2014




Journal of











Journal of


Function Spaces






Algebra















Models 120594

2 DF 119875 120594

2DF RMSEA1 GFI2 CN3





0

Σ =


0




Income

Sex

Job

Edu

Lic

Ic

1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

00

0

0

0

0

0

00

00

00

00

0

00

Big_zone

0

0

N_kid

00

0

0

N_people

00

1

D_H

N_W

N_O

D_S

D_M

D_L

T_W

T_O

e1

e2

e3

e4

e5

e6

e7

e8






119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874













119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874




119863 119867


119863 119878


119863 119872


119863 119871









119873 119882


119879 119882


119873 119874


119879 119874




119873 119882


119879 119882


119873 119874


119879 119874





7 Conclusions





Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8














Acknowledgments


References




























Volume 2014




Journal of











Journal of


Function Spaces






Algebra












119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874













119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874




119863 119867


119863 119878


119863 119872


119863 119871









119873 119882


119879 119882


119873 119874


119879 119874




119873 119882


119879 119882


119873 119874


119879 119874





7 Conclusions





Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8














Acknowledgments


References




























Volume 2014




Journal of











Journal of


Function Spaces






Algebra












119863 119867


119863 119878


119863 119872


119863 119871


119873 119882


119879 119882


119873 119874


119879 119874




119863 119867


119863 119878


119863 119872


119863 119871









119873 119882


119879 119882


119873 119874


119879 119874




119873 119882


119879 119882


119873 119874


119879 119874





7 Conclusions





Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8














Acknowledgments


References




























Volume 2014




Journal of











Journal of


Function Spaces






Algebra












119873 119882


119879 119882


119873 119874


119879 119874




119873 119882


119879 119882


119873 119874


119879 119874





7 Conclusions





Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8














Acknowledgments


References




























Volume 2014




Journal of











Journal of


Function Spaces






Algebra










Income

Sex

Job

Edu

D_H

N_W

N_O

D_S

Lic

Ic

D_M

D_L

1

T_W

T_O1

1

1

1

1

1

1

N_car

N_bike

N_ebike

Age

0

0

0

0

0

0 0

0

00

00

Big_zone0

0

0

00

00

N_kid

0

00

00

N_people

0

0

0

e1

e2

e3

e4

e5

e6

e7

e8














Acknowledgments


References




























Volume 2014




Journal of











Journal of


Function Spaces






Algebra


































Volume 2014




Journal of











Journal of


Function Spaces






Algebra









research article activity-trip chaining behavior of urban low...

Documents