際際滷

際際滷Share a Scribd company logo

Dynamics of Social Interaction Frequency: Role of Geography and Accessibility

Robin Lovelace
Robin Lovelace

This presentation won the postgraduate paper prize for a TGRG-sponsored submission at the RGS-IBG conference 2013.

1 of 19
Download to read offline
Dynamics of Social Interaction Frequency: Role of Geography and Accessibility Fariya Sharmeen
Background People are space bound. Social networks also are embedded in a spatial setting. Studies on social networks recognize distance in geographical space as a determinant of distance in network links in terms of tie strength between an ego and alters. Thus geography matters for social interactions as much as anything else / department of the Built Environment 27-8-2013 2
Background Which geographical features? Distance to alters Accessibility to Transport infrastructures Accessibility to Facilities Degree of Urbanization History of interaction Dynamics of distance / department of the Built Environment 27-8-2013 3
Background Why this is important? 1. Acknowledge the role of geographical features in explaining social interaction 2. Comprehensive modelling approach 3. can provide useful direction to sustainable community design and can strengthen social cohesion policies. / department of the Built Environment 27-8-2013 4
Theory and Concepts Accessibility to facilities and transport infrastructures Egos socio demographics h o m Ego-Alter tie o characteristics p h il y Geographical distance with alter Social Interaction History of social interaction path dependency / department of the Built Environment 27-8-2013 5 Degree of urbanization Changes in geographical distance with alter life course approach a c c e s s i b il it y
Research Questions To what extent are social interaction/activity frequencies path dependent? Does a change in geographical distance with alters influence social interaction/activity frequency? What are the effects of urbanization and accessibility to social/recreational facilities on social interaction/activity frequency? What are the effects of life cycle-events on social interaction/activity frequency? / department of the Built Environment 27-8-2013 6
Data Event based retrospective survey 1.Socio demographics 2.Present social network 3.Dynamics in social network 4.Dynamics in Activity Travel schedule Netherlands 700 respondents September 2011 8458 ties / department of the Built Environment 27-8-2013 7
Data Cetral Bureau of Statistics, Netherlands Degree of urbanization and Accessibility to facilities within neighbourhood defined by 4 digit postcode 2006 Open street map calculate distance to city/town/village center / department of the Built Environment 27-8-2013 8
Methodology Ordered Logit model n y ijt = 硫 1 itk x1 itk + 竜 ij k =1 Extension 1: Random effects model n y ijt = 硫 n 1itk n x1itk + 硫 2 ijtk x 2 ijtk + 硫 3 itk x3 itk + 留 i + 竜 ij k =1 k =1 k =1 Extension 2: Path dependence model n n n n yijt = 硫1itkx1itk +硫2ijtkx2ijtk + 硫3itkx3itk + 硫2ij(t1)k x2ij(t)k + yij(t1) +留i +竜ij k=1 / department of the Built Environment k=1 k=1 27-8-2013 9 k=1
Descriptives / department of the Built Environment 27-8-2013 10
Results Ordered logit model of face-to-face interaction frequency Dependent variable face o face interaction frequency (0=none, 1=less, 2=once a month, 3=2-3 times a month, 4=once a week, 5=2-3 times a week, 6=daily) Model 1: Basic model Model 3: Model 2: Random effects Random with land use effects and accessibility model indicators Model 4: Random effects with state dependence Model 5: Random effects with interaction variables -12276.56 -12262.65 Log likelihood function -14589.34 -14262.48 -14245.97 Restricted log -15514.46 -14589.34 -14562.30 likelihood Chi square 1850.24*** 653.71*** 632.65*** -12582.74 -12567.12 612.36*** 608.93*** # of Parameters 34 35 44 49 54 AIC/N 3.458 3.381 3.379 2.915 2.912 i=Exp(-AIC /2) 6.70 4.85 3.30 1.34 1 / department of the Built Environment 27-8-2013 11
Results 6 groups of variables I1: Egos socio demographics I2: Ego-alter tie characteristics I3: Life cycle events I4: Density and Aceessibility indicators I5: History of social interactions and distance I6: interaction effects I1: Egos socio demographics AGE21-30 Gender: male No car Driving license: yes HH with children HH without children HH with friends Work hr per week Size of Social network # club membership / department of the Built Environment 27-8-2013 12 硫 0.352 -0.007 0.319 0.144 -0.081 -0.151 -0.080 0.001 -0.103 0.082 *** *** *** ** *** *** *** t 7 -0.25 7.91 4.19 -2.17 -3.67 -1.6 1.33 -7.45 3.45
Results I2: Ego-alter tie characteristics Same age group Same education level Same gender Age difference: upto 2 category Education level difference: upto 2 category Relation: neighbour Relation: friend Relation: family Relationship strength: weak Relationship strength: medium Distance: 0-1 km Distance: 2-5 km Distance: 6-30 km / department of the Built Environment 27-8-2013 13 硫 -0.119 0.218 0.169 0.011 -0.028 0.203 0.239 0.272 -0.893 -0.315 2.739 0.284 0.257 *** *** *** *** *** *** *** *** *** *** *** t -4.57 5.07 6.24 0.46 -0.73 4.69 8.2 7.49 -19.13 -12.46 46.67 7.21 6.98
Results I3: Life cycle events Event: change in work/study Event: children start school Event: change in residence Event time: 6-12 months Event time: 0-6 months I4: Density and accessibility indicators Distance to city centre (log) Rural area Suburban area #Shopping within 1 km #Caf辿/Restaurant within 1 km #Attractions within 10 km Distance to highway (log) Distance to train station (log) Distance to public green (log) / department of the Built Environment 27-8-2013 14 硫 0.056 0.092 -0.108 -0.161 -0.202 硫 -0.011 0.064 -0.057 0.014 -0.002 -0.015 0.034 -0.031 0.001 *** *** *** * * * * *** *** * * t 0.81 1.12 -2.71 -4.6 -6.1 t 1.68 1.92 -1.86 1.67 -6.53 -2.41 1.79 -1.76 -0.62
Results I5: History of social interaction and distance F2F before: daily F2F before: weekly F2F before: monthly NOWFAR NOWCLOSE 3.342 2.200 0.812 -0.425 -0.059 I6: Interaction effects -0.347 0.412 0.104 -0.144 0.251 *** *** *** ** 硫 Neighbour*change in residence NOWFAR*change in residence NOWFAR*change in work/study NOWCLOSE*change in work/study NOWCLOSE*change in residence Threshold parameter Mu(01) Mu(02) Mu(03) Mu(04) Mu(05) t 硫 硫 1.632 2.483 3.116 3.938 5.003 t 66.96 96.24 114.0 142.4 163.2 / department of the Built Environment Std Deviation random effect Sigma 27-8-2013 15 硫 t 0.482 20.82 89.85 60.84 17.39 -2.13 -1.01 t *** *** *** *** -5.07 2.35 0.56 -3.48 5.27
Answers to Research Questions Face-to-face social interaction frequencies are strongly path dependent. Findings support that change in geographical distance affect social interaction frequency. Varied effects of degree of urbanization, accessibility to facilities and transport infrastructures were observed and reported. Social interaction/activity frequency dynamics are subject to the type of event. / department of the Built Environment 27-8-2013 16
Conclusion The study aims to model face to face social interaction frequency taking the effects of accessibility, degree of urbanization, and path dependency. It shows that inclusion of local geographical and accessibility indicators and long term dynamics improves model performance. The findings add to the long debated body of literature in sociology about distance decay in social relationships. The novel contribution is the inclusion of the effects of changes induced by life-cycle events. / department of the Built Environment 27-8-2013 17
Next? In depth analysis of purpose and location-type choice Alters home location Preferences, limitations of individuals Ethnic and migrated neighbourhoods / department of the Built Environment 27-8-2013 18
Thank you Fariya Sharmeen f.sharmeen@tue.nl

More Related Content

Dynamics of Social Interaction Frequency: Role of Geography and Accessibility

  • 1. Dynamics of Social Interaction Frequency: Role of Geography and Accessibility Fariya Sharmeen
  • 2. Background People are space bound. Social networks also are embedded in a spatial setting. Studies on social networks recognize distance in geographical space as a determinant of distance in network links in terms of tie strength between an ego and alters. Thus geography matters for social interactions as much as anything else / department of the Built Environment 27-8-2013 2
  • 3. Background Which geographical features? Distance to alters Accessibility to Transport infrastructures Accessibility to Facilities Degree of Urbanization History of interaction Dynamics of distance / department of the Built Environment 27-8-2013 3
  • 4. Background Why this is important? 1. Acknowledge the role of geographical features in explaining social interaction 2. Comprehensive modelling approach 3. can provide useful direction to sustainable community design and can strengthen social cohesion policies. / department of the Built Environment 27-8-2013 4
  • 5. Theory and Concepts Accessibility to facilities and transport infrastructures Egos socio demographics h o m Ego-Alter tie o characteristics p h il y Geographical distance with alter Social Interaction History of social interaction path dependency / department of the Built Environment 27-8-2013 5 Degree of urbanization Changes in geographical distance with alter life course approach a c c e s s i b il it y
  • 6. Research Questions To what extent are social interaction/activity frequencies path dependent? Does a change in geographical distance with alters influence social interaction/activity frequency? What are the effects of urbanization and accessibility to social/recreational facilities on social interaction/activity frequency? What are the effects of life cycle-events on social interaction/activity frequency? / department of the Built Environment 27-8-2013 6
  • 7. Data Event based retrospective survey 1.Socio demographics 2.Present social network 3.Dynamics in social network 4.Dynamics in Activity Travel schedule Netherlands 700 respondents September 2011 8458 ties / department of the Built Environment 27-8-2013 7
  • 8. Data Cetral Bureau of Statistics, Netherlands Degree of urbanization and Accessibility to facilities within neighbourhood defined by 4 digit postcode 2006 Open street map calculate distance to city/town/village center / department of the Built Environment 27-8-2013 8
  • 9. Methodology Ordered Logit model n y ijt = 硫 1 itk x1 itk + 竜 ij k =1 Extension 1: Random effects model n y ijt = 硫 n 1itk n x1itk + 硫 2 ijtk x 2 ijtk + 硫 3 itk x3 itk + 留 i + 竜 ij k =1 k =1 k =1 Extension 2: Path dependence model n n n n yijt = 硫1itkx1itk +硫2ijtkx2ijtk + 硫3itkx3itk + 硫2ij(t1)k x2ij(t)k + yij(t1) +留i +竜ij k=1 / department of the Built Environment k=1 k=1 27-8-2013 9 k=1
  • 10. Descriptives / department of the Built Environment 27-8-2013 10
  • 11. Results Ordered logit model of face-to-face interaction frequency Dependent variable face o face interaction frequency (0=none, 1=less, 2=once a month, 3=2-3 times a month, 4=once a week, 5=2-3 times a week, 6=daily) Model 1: Basic model Model 3: Model 2: Random effects Random with land use effects and accessibility model indicators Model 4: Random effects with state dependence Model 5: Random effects with interaction variables -12276.56 -12262.65 Log likelihood function -14589.34 -14262.48 -14245.97 Restricted log -15514.46 -14589.34 -14562.30 likelihood Chi square 1850.24*** 653.71*** 632.65*** -12582.74 -12567.12 612.36*** 608.93*** # of Parameters 34 35 44 49 54 AIC/N 3.458 3.381 3.379 2.915 2.912 i=Exp(-AIC /2) 6.70 4.85 3.30 1.34 1 / department of the Built Environment 27-8-2013 11
  • 12. Results 6 groups of variables I1: Egos socio demographics I2: Ego-alter tie characteristics I3: Life cycle events I4: Density and Aceessibility indicators I5: History of social interactions and distance I6: interaction effects I1: Egos socio demographics AGE21-30 Gender: male No car Driving license: yes HH with children HH without children HH with friends Work hr per week Size of Social network # club membership / department of the Built Environment 27-8-2013 12 硫 0.352 -0.007 0.319 0.144 -0.081 -0.151 -0.080 0.001 -0.103 0.082 *** *** *** ** *** *** *** t 7 -0.25 7.91 4.19 -2.17 -3.67 -1.6 1.33 -7.45 3.45
  • 13. Results I2: Ego-alter tie characteristics Same age group Same education level Same gender Age difference: upto 2 category Education level difference: upto 2 category Relation: neighbour Relation: friend Relation: family Relationship strength: weak Relationship strength: medium Distance: 0-1 km Distance: 2-5 km Distance: 6-30 km / department of the Built Environment 27-8-2013 13 硫 -0.119 0.218 0.169 0.011 -0.028 0.203 0.239 0.272 -0.893 -0.315 2.739 0.284 0.257 *** *** *** *** *** *** *** *** *** *** *** t -4.57 5.07 6.24 0.46 -0.73 4.69 8.2 7.49 -19.13 -12.46 46.67 7.21 6.98
  • 14. Results I3: Life cycle events Event: change in work/study Event: children start school Event: change in residence Event time: 6-12 months Event time: 0-6 months I4: Density and accessibility indicators Distance to city centre (log) Rural area Suburban area #Shopping within 1 km #Caf辿/Restaurant within 1 km #Attractions within 10 km Distance to highway (log) Distance to train station (log) Distance to public green (log) / department of the Built Environment 27-8-2013 14 硫 0.056 0.092 -0.108 -0.161 -0.202 硫 -0.011 0.064 -0.057 0.014 -0.002 -0.015 0.034 -0.031 0.001 *** *** *** * * * * *** *** * * t 0.81 1.12 -2.71 -4.6 -6.1 t 1.68 1.92 -1.86 1.67 -6.53 -2.41 1.79 -1.76 -0.62
  • 15. Results I5: History of social interaction and distance F2F before: daily F2F before: weekly F2F before: monthly NOWFAR NOWCLOSE 3.342 2.200 0.812 -0.425 -0.059 I6: Interaction effects -0.347 0.412 0.104 -0.144 0.251 *** *** *** ** 硫 Neighbour*change in residence NOWFAR*change in residence NOWFAR*change in work/study NOWCLOSE*change in work/study NOWCLOSE*change in residence Threshold parameter Mu(01) Mu(02) Mu(03) Mu(04) Mu(05) t 硫 硫 1.632 2.483 3.116 3.938 5.003 t 66.96 96.24 114.0 142.4 163.2 / department of the Built Environment Std Deviation random effect Sigma 27-8-2013 15 硫 t 0.482 20.82 89.85 60.84 17.39 -2.13 -1.01 t *** *** *** *** -5.07 2.35 0.56 -3.48 5.27
  • 16. Answers to Research Questions Face-to-face social interaction frequencies are strongly path dependent. Findings support that change in geographical distance affect social interaction frequency. Varied effects of degree of urbanization, accessibility to facilities and transport infrastructures were observed and reported. Social interaction/activity frequency dynamics are subject to the type of event. / department of the Built Environment 27-8-2013 16
  • 17. Conclusion The study aims to model face to face social interaction frequency taking the effects of accessibility, degree of urbanization, and path dependency. It shows that inclusion of local geographical and accessibility indicators and long term dynamics improves model performance. The findings add to the long debated body of literature in sociology about distance decay in social relationships. The novel contribution is the inclusion of the effects of changes induced by life-cycle events. / department of the Built Environment 27-8-2013 17
  • 18. Next? In depth analysis of purpose and location-type choice Alters home location Preferences, limitations of individuals Ethnic and migrated neighbourhoods / department of the Built Environment 27-8-2013 18