Commuting mode choice in transit oriented development: Disentangling the effects of competitive neighbourhoods, travel attitudes, and self-selection
Kamruzzaman, M., Shatu, F.M., Hine, J. Turrell, G., 2015, in Transport Policy
doi:10.1016/j.tranpol.2015.06.003
Location |
Brisbane, Queensland, Australia |
Population |
Other (specify) |
Sample size |
3537 |
Factor analysis type |
exploratory factor analysis, oblimin rotation |
Stepwise regression |
no |
Removal of insignificant variables |
yes |
Reviewed by |
AR |
Abstract
This research identifies the commuting mode choice behaviour of 3537 adults living in different types of transit oriented development (TOD) in Brisbane by disentangling the effects of their "evil twin" transit adjacent developments (TADs), and by also controlling for residential self-selection, travel attitudes and preferences, and socio-demographic effects. A TwoStep cluster analysis was conducted to identify the natural groupings of respondents' living environment based on six built environment indicators. The analysis resulted in five types of neighbourhoods: urban TODs, activity centre TODs, potential TODs, TADs, and traditional suburbs. HABITAT survey data were used to derive the commute mode choice behaviour of people living in these neighbourhoods. In addition, statements reflecting both respondents' travel attitudes and living preferences were also collected as part of the survey. Factor analyses were conducted based on these statements and these derived factors were then used to control for residential self-selection. Four binary logistic regression models were estimated, one for each of the travel modes used (e.g. public transport, active transport, less sustainable transport such as the car/taxi, and other), to differentiate between the commuting behaviour of people living in the five types of neighbourhoods. The findings verify that urban TODs enhance the use of public transport and reduce car usage. No significant difference was found in the commuting behaviour between respondents living in traditional suburbs and TADs. The results confirm the hypothesis that TADs are the "evil twin" of TODs. The data indicates that TADs and the mode choices of residents in these neighbourhoods is a missed transport policy opportunity. Further policy efforts are required for a successive transition of TADs into TODs in order to realise the full benefits of these. TOD policy should also be integrated with context specific TOD design principles. © 2015 Elsevier Ltd.
Factors
Models
Dependent variable |
Mode of transport to work: car/taxi/motorcycle |
Model type |
binary logit |
Sample size |
3537.0 |
R2 |
nan |
Adjusted R2 |
|
Pseudo R2
(Nagelkerke)
|
0.27 |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
nan |
Log-likelihood at constants |
nan |
Log-likelihood at convergence |
-1362.535 |
Variable |
Odds_ratio |
p-value |
Constant |
9.61
|
<0.05 |
Neighborhood tpye(ref: traditional suburbs) |
nan
|
nan |
Urban TOD |
0.69
|
<0.05 |
Travel attitudes and preferences |
nan
|
nan |
Negative perception about PT |
1.4
|
<0.05 |
Car dependent |
2.06
|
<0.05 |
Travel Time( ref: less than 15 minutes) |
nan
|
nan |
15-30 min |
1.59
|
<0.05 |
30-60 min |
0.63
|
<0.05 |
More than 60 min |
0.46
|
<0.05 |
Socio-demographics |
nan
|
nan |
Female (ref: male) |
1.32
|
<0.05 |
Car availability (ref: yes, always) |
nan
|
nan |
Yes, sometimes |
0.25
|
<0.05 |
No (ref: yes, always) |
0.33
|
<0.05 |
Do not drive (ref: yes, always) |
0.13
|
<0.05 |
Current living arrangment(ref: living alone) |
nan
|
nan |
Couple living with > = 1 children |
1.3
|
<0.05 |
Health Status |
0.84
|
<0.05 |
Dependent variable |
Mode of transport to work: public transport |
Model type |
binary logit |
Sample size |
3537.0 |
R2 |
nan |
Adjusted R2 |
|
Pseudo R2
(Nagelkerke)
|
0.44 |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
nan |
Log-likelihood at constants |
nan |
Log-likelihood at convergence |
-1080.13 |
Variable |
Odds_ratio |
p-value |
Constant |
0.08
|
<0.05 |
Neighbourhood type(ref: traditional suburbs) |
nan
|
nan |
Urban TOD |
1.46
|
<0.05 |
Built environment indicators |
nan
|
nan |
Net employment density |
0.99
|
<0.05 |
Travel attitudes and preferences |
nan
|
nan |
Negative perception about PT |
0.48
|
<0.05 |
Sensitive to environmental externalities |
0.86
|
<0.05 |
Car dependent |
0.44
|
<0.05 |
Reasons to choose current address |
nan
|
nan |
Accessibility and mobility of places |
1.29
|
<0.05 |
Natural environment |
0.87
|
<0.05 |
Travel time(ref: less than 15 minutes) |
nan
|
nan |
15-30 min |
10.29
|
<0.05 |
30-60 min |
41.8
|
<0.05 |
More than 60 min |
59.91
|
<0.05 |
Socio-demographics |
nan
|
nan |
Female (ref: male) |
1.49
|
<0.05 |
Age |
0.97
|
<0.05 |
Car availability (ref: yes, always) |
nan
|
nan |
Yes sometimes |
5.12
|
<0.05 |
Do not drive (ref: yes, always) |
1.97
|
<0.05 |
Level of education(ref: upto year 12) |
nan
|
nan |
Diploma/certificate |
0.73
|
<0.05 |
Household size |
0.84
|
<0.05 |
Dependent variable |
Mode of transport to work: active transport |
Model type |
binary logit |
Sample size |
3537.0 |
R2 |
nan |
Adjusted R2 |
|
Pseudo R2
(Nagelkerke)
|
0.18 |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
nan |
Log-likelihood at constants |
nan |
Log-likelihood at convergence |
-1139.875 |
Variable |
Odds_ratio |
p-value |
Constant |
0.11
|
<0.05 |
Neighbourhood type(Ref: traditional suburbs) |
nan
|
nan |
Activity centre TOD |
0.71
|
<0.05 |
TAD |
0.55
|
<0.05 |
Built environmental indicators |
nan
|
nan |
Net employment density |
1.01
|
<0.05 |
Intersection density |
0.53
|
<0.05 |
Travel attitudes and preferences |
nan
|
nan |
Sensitive to environmental externalities |
1.15
|
<0.05 |
Car dependent |
0.54
|
<0.05 |
Travel time (ref: less than 15 minutes) |
nan
|
nan |
15-30 min |
0.41
|
<0.05 |
30-60 min |
0.79
|
<0.05 |
Socio-demographics |
nan
|
nan |
Female(ref: male) |
0.7
|
<0.05 |
Car availability (ref: yes, always) |
nan
|
nan |
Yes, sometimes |
2.93
|
<0.05 |
Do not drive(ref: yes, always) |
3.45
|
<0.05 |
Level of education (ref: upto year 12) |
nan
|
nan |
Graduate and over |
1.41
|
<0.05 |
Current living arrangement(ref: living alone) |
nan
|
nan |
Other |
2.18
|
<0.05 |
Health Status |
1.24
|
<0.05 |
Dependent variable |
Mode of transport to work: other |
Model type |
binary logit |
Sample size |
3537.0 |
R2 |
nan |
Adjusted R2 |
|
Pseudo R2
(Nagelkerke)
|
0.15 |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
nan |
Log-likelihood at constants |
nan |
Log-likelihood at convergence |
-413.785 |
Variable |
Odds_ratio |
p-value |
Constant |
0.2
|
<0.05 |
Neighbourhood type(ref: traditional suburbs) |
nan
|
nan |
Urban TOD |
0.44
|
<0.05 |
Activity Centre |
0.33
|
<0.05 |
Potential TOD |
0.38
|
<0.05 |
Built environmental indicators |
nan
|
nan |
Cul-de-sac density |
0.11
|
<0.05 |
Travel attitudes and preferences |
nan
|
nan |
Car dependent |
1.37
|
<0.05 |
Travel time (ref: less than 15 minutes) |
nan
|
nan |
15-30 min |
0.16
|
<0.05 |
30-60 min |
0.17
|
<0.05 |
Socio-demographics |
nan
|
nan |
Female (ref: male) |
0.43
|
<0.05 |
Emploment status: full time (Ref: part time) |
0.5
|
<0.05 |
Level of education(ref: upto year 12) |
nan
|
nan |
Graduate and over |
0.54
|
<0.05 |
Health status |
1.3
|
<0.05 |