|
City: |
Edmonton, Alberta, Canada |
||||||||||||||||||||||||||||||
|
Organization: |
-
Department of Civil and Environmental Engineering,
University of Alberta -
Department of Resource Economics
and Environmental Sociology, University of Alberta |
||||||||||||||||||||||||||||||
|
Project
Start Date: |
August 19-25, 2014 (Survey date). Data of project
initiation not specified. |
||||||||||||||||||||||||||||||
|
Project
End Date: |
2018 (Publication) |
||||||||||||||||||||||||||||||
|
Reference: |
Cabral, L., A. M., & Parkins, J.R. (2018). “Bicycle
ridership and intention in a northern, low-cycling city”, Travel Behaviour
and Society, Vol. 13, pp. 165-173. |
||||||||||||||||||||||||||||||
|
Problem: |
Edmonton is the northernmost North American city with a
metropolitan population over one million and endures long, cold, and snowy
winters. Combined with a high car-dependency, sprawling cityscape, and poor
road maintenance, this translates to a low cycling rate. This case study used
public survey data to try and quantify the effects of infrastructure density,
traffic attitude, perceived control over time and distance, and traffic
stress tolerance perception on (1) cycling for utility purposes, (2) the
intention to cycle more frequently, and (3) the use of an active mode of
transportation, specifically for a northern and low-cycling city. The purpose
was to see if the explanatory variables/determinants either differed from or
remained significant in accordance with reviewed literature that studied
cycling behaviour in warmer cities. |
||||||||||||||||||||||||||||||
|
Technical
Solution: |
Three
empirical models (one each for finding determinants of 1) bicycle use
(utility cycling), 2) cycling intent, and 3) active mode of transportation
behaviour) were developed to describe cycling behaviour using binary
logistic regression. The data analysis was performed using IBM’s SPSS
software. The table below summarizes the data preparation using responses
from the public survey dataset.
Control
variables:
gender, age, employment status, education level, income, and whether the
respondent lived with children. General
Indications from the survey: -
15% of the respondents can be considered utility cyclists whereas 44% use an
active primary or secondary mode of transportation. -
More than a third of respondents would strongly like to cycle more often. |
||||||||||||||||||||||||||||||
|
Datasets
Used: |
646 responses to a bike
ridership survey conducted in 2014 by the City of Edmonton. This is publicly
available data drawn from a regularly surveyed panel of citizens called the
Edmonton Insight Community (who have signed up voluntarily). The responses
required answers to the eight questions pertaining to the three outcome
variables as well as the five explanatory variables. |
||||||||||||||||||||||||||||||
|
Outcome: |
Most
variables were significant and in line with other study findings in the
current literature that focused on cycling in warmer climates. Results
suggested the importance of perceived safety in deciding or intending to
cycle, as well as perceived time and distance of travel. 1) Respondents with a higher traffic
stress tolerance and stronger agreement that many destinations are within
cycling distance (perceived distance) are positively associated with the
outcomes of 1) being utility cyclists, 2) having strong cycling intent, and
3) active travel use. 2) As respondents agree to the
statement of not having time to cycle to their destinations (perceived time),
the negative correlation on all three outcome variables means that these
respondents are much less likely to be utility cyclists, intend to cycle, or
use active modes in general. 3) Higher density of cycling
facilities is significantly associated only with the use of an active mode of
transportation, and not with utility cycling behavior or the intent to cycle.
This was an unexpected outcome since most transportation literature generally
agrees that cycling is supply/facility driven. 4) Traffic attitude appeared to be
significantly correlated to the intention to cycle more often (as respondents
agree more with the proposition that the streets around their house have too
much traffic to support safe cycling, they also have a stronger intention to
cycle more frequently). This outcome was also unexpected. However, the high
statistical significance of this variable was explained by including an
intermediate variable that identified if respondents who perceive there is
too much traffic generally agree that, if it was safer to cycle on the road,
they would ride more often. The high level of agreement erased the initial
correlation and suggested that this intent was conditional upon safety. Control
variables: - Generally insignificant - Being a student is positively
correlated with the intent to cycle more often - Unemployed people are less likely
to intend to cycle more often - Males are more likely to be utility
cyclists - Older people are less likely to
have cycling intention Conclusions: -
Despite lower amounts of ridership and climatic differences, most of the
variables explored have similar impacts in Edmonton on the 3 dependent
variables as have been observed in other cities studied in current literature -
Safety-conditional traffic attitude suggests that people would like to cycle
more often but feel that conditions are not safe enough (strongly implies
that better cycling infrastructure investment is needed) - Environmental
values such as poor winter maintenance, darkness, lack of driver awareness,
and weather should be considered in future research |
||||||||||||||||||||||||||||||
|
Issues
that arose: |
-
The responses were taken from a panel of citizens who had voluntarily signed
up for the Edmonton Insight Community and were not representative of the
entire municipal population. Thus, the panel was heavily skewed towards
fulltime workers (70% vs 57% of municipal population) and lacked high school
and post-secondary students (2% vs 10% of municipal population) -
Incomplete responses, generally linked to missing answers to the profiling
questions such as gender and income, were automatically removed from the
binary logistic analysis by the SPSS software, yielding a final sample of
N=550 respondents. -
The original intent of the study was to model different levels of cycling
intensity amongst utility cyclists. The low number in the sample (N = 96) of
respondents identifying as utility cyclists meant that exploratory models
were unable to accurately distinguish between different levels of cycling
intensity. The binary variable was therefore utilized as a best replacement. |
||||||||||||||||||||||||||||||
|
Status: |
Terminated |
||||||||||||||||||||||||||||||
|
Entered
by: |
October 28th, 2020. Connor Bayne,
connor.bayne@mail.utoronto.ca |
||||||||||||||||||||||||||||||
CEM1002,
Civil Engineering, University of Toronto
Contact: msf@eil.utoronto.ca