Companionship and Time Investment in Social Fields at ...

Companionship and Time Investment in Social Fields at Different Life Cycle Stages:

Implications for Activity and Travel Modeling and Simulation

Jae Hyun Lee Konstadinos G. Goulias

Geotrans Laboratory

Department of Geography University of California, Santa Barbara

1

Outline

• Introduction

• Data & Descriptive Analysis

• Model Specification

• Results & Findings

• Limitations

2

Introduction

• In time allocation and travel one of the most important factors is human interaction (Kitamura, 1988; Jones, 1990; Bhat & Koppelman, 1999)

3

Introduction

• Intra-household interactions are most often included in activity-travel analysis, and many researchers have shown their significant roles in explaining daily-life (Townsend, 1987; Van Wissen, 1997; Golob and McNally, 1997; Golob 1998)

• In travel behavior research we also see increasing attention on

incorporating the influence on behavior of interactions with external to the household persons (Carrasco and Miller, 2006, 2009; Larsen et al., 2006; Axhausen, 2002; Goulias and Kim, 2005; Arentze and Timmermans, 2008)

• Recent research in travel behavior emphasizes the role played by human interactions on time allocation to activities and travel (Goulias, 2007; Dubernet & Axhausen, 2014)

4

• Daily interpersonal contacts is a determinant and motivation of behavior and can be classified according to a field of action of an individual that changes gradually in one's life course (Goulias, 2009) and as an outcome of roles in social networks (Carrasco and Miller, 2006, 2009)

• E.g. Changes in my daily schedule over time

Introduction

6th Grade

8 School 15 Play soccer with schoolmates 17 Snack with schoolmates 18 English Academy

College

6 Swimming 10 Econ Geo Class 13 Group meeting for Assignment 15 Climatology Class 17 Dinner with College friends and Karaoke

Full-time Job at KOTI

9 Work 13 Meeting with Freight model team 17 Meeting with Policy Support System team 19 Dinner with KTDB colleagues

Grad student with Family

8 Swimming 11 TA for Transport modeling class 13 Office hour as TA 14 Geo-Stats Class 18 Dinner with family

5

Introduction

• Our interest includes an explicit accounting of daily contacts (with whom and for whom) in a model system of equations that are able to show the role life cycle stages play in differentiating among persons' propensity to interact with other persons in a day

• We also want to do this in a time allocation model system that allows modeling trade-offs among activities and study the day-of-week differences

6

Data & Descriptive Analysis

• CentreSim database 2002 – Sample size: 1471 persons from 718 households

– Location: Centre County, Pennsylvania

– Each person provides their two-day activity information along with answers to the question “with whom the activity was completed” and “for whom the activity was completed” for each activity

Source : Google map 7


8

• Social Network – 1) Family 2) Relatives 3) Friends 4) Businessmates 5) Schoolmates 6)

Clubmates 7) others

• Contacts Variables – Total number of people contacted in two days in a certain category of

relationship

e.g. 1) # of Family contacts in two days 2) # of Relatives Contacts in

two days, and so on.

– Total number of persons for whom each activity was conducted in two days


9


0 5 10 15 20 25 30

Life-Cycle Stages with the Number of Contacts in Two days

10


0 200 400 600 800 1000 1200 1400 1600 1800 2000

Life-Cycle stages with Total Activity Duration in Two Days

11

Data & Descriptive Analysis Life-Cycle Stages with the Number of activities for Self and Others in Two days

12

Data & Descriptive Analysis Life-Cycle stages with Total Activity Duration for Self and Others in Two Days

13

Path Model (SEM) Specification

Schematic of the Full Model with Broad Categories

Children < 17 years old College Students

Home duties Part-time with Kids

Part-time no Kid Fulltime with Kids

Fulltime no Kid Retired and Disabled

Friday Saturday Sunday

Family Relatives Friends

Club-mates

Work School Sleep

Home-leisure Home-other

Shopping Social Service

Total Travel time

Accessibility

Spatial Category

Life Cycle Category

Day of the Week Category

Number of people with whom interacted in a day

Amount of daily activity time


Travel time


Club-mates

For how many people they do something in a day

14

Path Model (SEM) Specification

Schematic of the No-Contacts Model with Broad Categories






Work School Sleep



Total Travel time

Accessibility

Spatial Category

Life Cycle Category




Travel time

15

Exogenous Variables - Life Cycle Stage

Life Cycle Category Detailed Life Cycle Stage

Children 1) Child at home, 2) Child at preschool, 3) Child at kindergarten

through 6th grade, 4) Child from 7th to 12th grade

College 5) College/University students

Homeduties 7) Homeduties with no children, 8) Homeduties with children

Parttime without kids 9) Part time worker (<40 hrs per week) with no children

Parttime with kids 10) Part time worker (<40 hrs per week) with children

Fulltime without kids 11) Full time worker (>=40 hrs per week) with no children

Fulltime with kids 12) Full time worker (>=40 hrs per week) with children

Retired & Disabled 13) Retired, 14) Disabled

All others 6) Looking for a job, 15) All other

16

Exogenous Variables - Others

• Day of the week - Friday, Saturday, Sunday; A dummy variable is defined based on the first day of interview. Ex) Friday dummy variable indicates the average duration of activities for Friday and Saturday

• Opportunity based Accessibility measure - The number of employees in all types of business establishments within three miles (median of entire travel distance in the CentreSim dataset)

Home Location

Business Establishments

Area within 3 miles from home(in network distance) 17

Results – Significant Variables







Club-mates

Work School Sleep



Total Travel time

Accessibility

Spatial Category

Life Cycle Category


Number of people with whom interacted in a day



Travel time


Club-mates

For how many people they do something in a day

18

Results – Full model

Children

College

Homeduties

Fulltime with kids

Fulltime no kids

# of family with

# of Relatives with

Work duration

School duration

Sleep duration

Home-Leisure duration

Service duration

Shop duration

Social duration Total Travel

time

Accessibility

Sunday

Friday

Saturday

Chi-square-261.110, df-63 CFI 0.973, TLI-0.924, SRMR 0.017 RMSEA 0.033 Positive significant path

Negative significant path

Parttime with kids

Parttime no kids

Retired & Disabled

# of Friends with

# of Clubmates with

# of family for

# of Relatives for

# of Friends for

# of Clubmates for

19


Children

College

Homeduties

Fulltime with kids

Fulltime no kids

# of family with

# of Relatives with

Accessibility

Sunday

Friday

Saturday



Parttime with kids

Parttime no kids

Retired & Disabled # of Friends with

# of Clubmates with

# of family for

# of Relatives for

# of Friends for

# of Clubmates for

20

Findings

• The model identifies a variety of life-cycle dependent roles of interpersonal interactions and “altruistic” activities in daily behavior

– Persons in homeduties and workers with kids are associated with altruistic activities the most

– Persons in many life-cycle groups do not participate in activities for their relatives

– Children are actively engaged with their family

– Workers without kids and college students interact with fewer family members in a day

– College students and Children are more likely to interact with their friends

– Weekend days are important for interacting with friends and clubmates

– Intra-household contacts and friends play important roles in explaining Maintenance (Service, Shopping) and Discretionary (Home-leisure, Social) activity durations

21


# of family with # of Relatives with



# of Friends with # of Clubmates with

# of family for # of Relatives for # of Friends for # of Clubmates for

22

Findings

• When persons need to participate in activities for others, they also interact with them at the same time (e.g., persons in homeduties and their children)

• Sometimes persons participate in activities for others, then interact with them later (e.g. buying a toy for their children)

• Activity participation for others that are not family members inhibits the number of interactions with family members

23


# of family with

# of Relatives with

Work duration

School duration

Sleep duration


Service duration

Shop duration

Social duration

Total Travel time

Positive significant path


# of Friends with

# of Clubmates with

# of family for

# of Relatives for

# of Friends for

# of Clubmates for

Accessibility

24

Findings

• Home based accessibility played a significant role in explaining Maintenance-Discretionary activity and total travel duration

• The with whom variables are stronger direct predictors of activity durations than the for whom variables

• The for whom variables influence activity duration through indirect effect

• Social activity durations and home-leisure activity durations were most sensitive to the inclusion of human interactions variables in the model

• Trade-off relationships among different types of activity durations were found except from Service to Travel time and Sleep to Leisure at home

25

Results – No-Contacts Model

Children

College

Homeduties

Fulltime with kids

Fulltime no kids

Work duration

School duration

Sleep duration


Service duration

Shop duration

Social duration Total Travel

time

Accessibility

Sunday

Friday

Saturday



Parttime with kids

Parttime no kids

Retired & Disabled

26

Results – Differences between two models






Work School Sleep



Total Travel time

Accessibility

Spatial Category

Life Cycle Category




Travel time

27

Results – Differences between two models






Work School Sleep



Total Travel time

Accessibility

Spatial Category

Life Cycle Category




Travel time

28

• Many significant indirect paths were found with the full model, and this confirms the need to incorporate human interactions and the paths of influence among daily activity durations

• A comparison between a model with no contacts and the full model with the daily contacts shows Fulltime workers, Parttime workers with kids, and college students are the most sensitive to the interpersonal contacts and the number of altruistic activities in time allocation

• Allocation of time to discretionary activities (particularly social activity and home leisure activity) is the most sensitive to specifications that control for number of daily contacts

Findings

29

• Aggregation of activity types confounds the influence of exogenous variables on duration (e.g., shopping), aggregation of children in the same group does not allow us to distinguish among different time allocations

• Although we find a clear influence of the day of the week on human interaction and time allocation, analyzing averages of two days masks the influence of each day of the week

Limitations

30

Acknowledgement

Funding for this research was provided by • The University of California Multicampus

Research Program Initiative on Sustainable Transportation

• The University of California Transportation Center (UCTC)

• The University of California Center on Economic Competitiveness in Transportation (UCCONNECT)

• Grants of the University of California Santa Barbara Office of Research and the College of Letters and Science.

31

Thanks very much

Questions?

32

Companionship and Time Investment in Social Fields at ...

Documents