Kevin Manaugh presented on factors that influence walking behavior. He discussed how walkability is affected by both the built environment and personal characteristics like motivations, needs, and culture. Manaugh identified different clusters of walkers based on their motivations and found their satisfaction varied by distance, with some groups more satisfied by longer walks. He concluded policies to increase walking should consider both infrastructure and appealing to different motivations to improve satisfaction and mode share equitably.

1. Kevin Manaugh
PhD Candidate McGill School of Urban Planning
Pro Walk/Pro Bike Conference September 11, 2012
SCHOOL OF URBAN PLANNING

2. Context
 Why do people walk?
 Quality of walking environment
 Convenience
 Necessity
 Environmental Awareness
 Cultural or family values
 Financial Constraints
 Enjoyment
 Exercise
 Social connections

3. Walkability/Propensity to Walk
 A vast literature explores walkability from a built
environment standpoint
 A vast literature also explores propensity to walk
from a behavioural/psychological perspective
 However, attempts to integrate these strands are less
common
 Many research and policy contexts ignore issues of
motivation and satisfaction

4. Big Questions
 Built form versus personal, household, neighbourhood,
cultural characteristics in walking decisions
 Does satisfaction with walking trips vary by personal
motivations? In addition to distance, slope, safety etc.

5. 1) Context/Research Questions
 How well do existing walkability indices explain the
variance in the choice to walk?
 Does this vary by trip purpose and socio-
demographic factors?
 What are the social equity implications of this?

6. 100
90
80
Percentage of walking trips
70
60
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10
Quality of Walking Environment

7. 100
90
80
Percentage of walking trips
70
60
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10
Quality of Walking Environment

8. Percentage of shopping trips by walking by decile
60
50
40
Walkscore
30 WI (800 m)
Walk Oppotunities
Pedshed (800m)
20
10
0
1 2 3 4 5 6 7 8 9 10

9. Percentage of shopping trips by walking by decile
60
581 trips
50
271 by foot
40
Walkscore
30 WI (800 m)
Walk Oppotunities
Pedshed (800m)
20
468 trips
24 by foot
10
0
1 2 3 4 5 6 7 8 9 10

10. Percentage of School trips by walking by decile
40
35
30
25
Walkscore
20 WI (800 m)
Walk Oppotunities
Pedshed (800m)
15
10
5
0
1 2 3 4 5 6 7 8 9 10

11. Percentage of School trips by walking by decile
40
35
253 out of
917 trips
30
25
Walkscore
20
105 out of WI (800 m)
1063 trips Walk Oppotunities
Pedshed (800m)
15
10
5
0
1 2 3 4 5 6 7 8 9 10

12. Sensitivity Analysis
Probabilities calculated at the mean* by walkscore deciles
No car low Retired Wealthy Middle age no Middle Class Large Young Wealthy
income no kids Kids Families Families
First Decile 72.1% 36.1% 12.6% 21.4% 30.6% 29.7% 18.5% 3.3%
Fifth Decile 74.8% 65.2% 38.4% 43.6% 43.6% 49.7% 35.8% 16.2%
Tenth Decile 78.0% 89.4% 79.5% 74.1% 61.0% 74.1% 63.1% 63.2%
*36 year old female making a 734 meter (average length) shopping trip

13. Urban Form Content
Streets Destinations
Intersections Parks
Sidewalks Transit
Trails Schools
Resident
Needs
Desires
Expectation
Culture

14. Urban Form Content
Streets Destinations
Intersections Parks
Sidewalks Transit
Trails Schools
Resident
Needs
Desires
Expectation
Culture

15. Urban Form Content
Streets Destinations
Intersections Parks
Sidewalks Transit
Trails Schools
Resident
Needs
Desires Walkability
Expectation
Culture

16. Conclusions (Part 1)
 Equity Issues
• People with limited choices are walking in
neighborhoods that are not ideal for walking
• Generalized indices (and performance measures)
might miss this distinction
• People walking does not necessarily equal good
walking environment
• Greater observed response among wealthy
households should not imply directed policy
response

17. Lessons Learned
 Walkability is not “one size fits all” but depends on:
 TripPurpose
 Socio-economic factors
 Gender
 Age
 Can perhaps best be described as a “match”
between built form factors and needs, preferences,
and desires of local residents.
 What next?

18. Part 2
Does Distance Matter? Exploring the links
between motivations and satisfaction in
walking trips

19. Context
 How do values and motivations relate to satisfaction
with walking trips?
 Much of travel behaviour research focuses on built
environment and proximity/accessibility issues
 Motivations to engage in active transportation and
derived satisfaction are often ignored

20. Context
 Most utility-maximization frameworks assume that
travel time and distance are elements of a trip to
be minimized
 However, recent research has highlighted the fact
that this may not always be the case
 For example, do ‘environmentalists’ or ‘exercise
junkies’ show a different response to trip
characteristics?

21. Data
 Survey
 Description of commute
 Motivations for using chosen mode
 Residential choice factors
 Trip satisfaction
 GIS
 Slope
 Other walkability variables
 Trip Distance

22. Methodology
 Correlations among values, motivations, satisfaction
levels, and trip characteristics
 Clustering of respondents by motivations to engage
in active transportation

23. Initial Findings
 No significant relationship between satisfaction and
distance travelled or slope of path
 Clustering of respondents by (self-reported)
motivation for walking

24. Cluster membership
Elevation
change Very Min (m) Max (m)
Cluster Count (m) satisfied (%) Distance (m)
Active cost Mimimizers 134 61.0** 19.4%**** 2034.2* 335.5 6068.6
Close Cost Mimimizers 88 29.7 25.0%**** 958.3 337.0 2354.3
Active Environmentalists 53 57.0** 52.8%*** 1801.3* 327.0 4020.9
Convenience 224 24.6 35.7% 846.3 194.5 3267.4
Close and exercise 106 30.9 40.6%*** 963.7 26.5 2862.5
Convenience and exercise 66 51.7** 39.4%*** 1675.6* 431.6 3561.6
*Statistically significant (ANOVA) F(5,671) = 61.18, p < .01(in relation to non-asterisks)
** (ANOVA) F(5,671) = 37.926, p < .01(in relation to non-asterisks)
***Chi-square (5, N = 671) = 27.58, p = .0001, higher than expected value
**** Chi-square (5, N = 671) = 27.58, p = .0001, lower than expected value

25. 100
90
80
70
60
Satisfaction
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10
Distance

26. 100
90
80
70
60
Satisfaction
50
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10
Distance

27. 70%
60%
50%
Percent Very Satisfied
40%
30%
20%
10%
0%
1 2 3 4 5
Quintile of Distance

28. 70%
Expon. (Close Cost Mimizers)
Expon. (Convience)
60%
Expon. (Convinience and
exercise)
50%
Percent Very Satisfied
40%
30%
20%
10%
0%
1 2 3 4 5
Quintile of Distance

29. 70%
Expon. (Active cost Mimizers)
60%
Expon. (Close and exercise)
50%
Percent Very Satisfied
40%
30%
20%
10%
0%
1 2 3 4 5
Quintile of Distance

30. 70%
Expon. (Active Environmentalists)
60%
50%
Percent Very Satisfied
40%
30%
20%
10%
0%
1 2 3 4 5
Quintile of Distance

31. 70%
Expon. (Active cost Mimizers)
Expon. (Close Cost Mimizers)
Expon. (Active Environmentalists)
Expon. (Convience)
60%
Expon. (Close and exercise)
Expon. (Convinience and exercise)
50%
Percent Very Satisfied
40%
30%
20%
10%
0%
1 2 3 4 5
Quintile of Distance

32. 70%
Expon. (Active cost Mimizers)
Expon. (Close Cost Mimizers)
Expon. (Active Environmentalists)
Expon. (Convience)
60%
Expon. (Close and exercise)
Expon. (Convinience and exercise)
50%
507 metres 2817 metres
Percent Very Satisfied
40%
30%
20%
10%
0%
1 2 3 4 5
Quintile of Distance

33. Conclusions
 People walk for a variety of reasons and
motivations (many of which have nothing to do with
built environment factors)
 People's satisfaction with walking is correlated with
these motivations
 Satisfaction rates are generally high

34. Conclusions
 Some people, particularly those with more
environmental awareness and propensity to
exercise, are more satisfied with longer distance
and greater slope

35. Conclusions
 This might have important implications about how
walking behavior is understood, predicted, and
modeled, particularly in terms of further expanding
utility maximization models to include preferences

36. Conclusions
 Policy implications: Is the goal to:
 Increase mode share? (GHG and CC)
 Increase total walking? (Population Health)
 Improve satisfaction of those already walking? (Equity)

37. Conclusions
 What does this all mean?
 What are the most important, effective, efficient levers
to increase walk/bike mode share?
 Policies, social marketing, infrastructure?

38. Kevin Manaugh kevin.manaugh@mail.mcgill.ca
SCHOOL OF URBAN PLANNING