Promoting walking and cycling as an alternative to using cars: systematic review

Criterion

Study

Adelaide^w15-18

Boston^w38-40

California
(cashing out)^{w41 w42}

California (telecommuting)^w49

Camden-Islington^w19

Delft^w23-30

Detmold-Rosenheim^w31-33

England
(bypasses)^{w36 w37}

England
(20 mph zones)^w35

Eugene^w22

Frome^w10

Glasgow^{w1 w2}

Gloucester^w11

Maidstone^w20

Perth^w3-9

Phoenix^w21

San Francisco^w45-47

Stockton^w34

Tampere^{w50 w51}

Trondheim^{w43 w44}

Voorhout^w48

Århus^w12-14

Were "before" and "after" data obtained from the same, or comparable, groups or areas?

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

Were outcomes compared with an appropriate control group or area, and if so, were control data collected in a comparable way?

P

P

P

P

P

P

P

P

P

P

P

Were participants, groups or areas randomly allocated to receive the intervention or the control?

P

P

P

Was the study sample randomly recruited from the study population with a response rate of at least 40%, or otherwise shown to be representative of the study population?

P

P

P

P

P

P

P

P

P

P

P

P

Were the results based on a minimum sample size of at least 100 people in each group or wave?

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

Were outcomes studied in a cohort or panel of respondents with an attrition rate of less than 30%?

P

P

P

P

Were confidence intervals or the results of significance tests reported?

P

P

P

P

P

P

P

P

P

P

P

Did the assessment of transport behaviour consider more than one type of trip, or trips made over a period longer than one day or a "typical" day?

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

Were the effects of the intervention assessed at least three months after the start of the intervention?

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

Were "before" and "after" data collected at seasonally comparable times of the year, or if not, was the study a controlled study?

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

Number of criteria met

4

4

8

4

8

7

6

3

5

4

9

9

9

7

7

5

7

5

6

7

7

7

Study

Year

Intervention

Study population

Study design

Primary outcome

Adelaide^w15-18

1998

Tailored feedback on travel diaries with suggestions on changing travel patterns, supported with customised information (Travel Blending)

Households living, working or visiting in two neighbourhoods of Adelaide

Uncontrolled prospective panel study

Reported frequency of, and time spent on, all trips in a seven-day travel diary by mode

Boston^w38-40

1978

Car restriction, subsidised bus services and pedestrianisation in central business district

Employees in city centre office buildings in Boston

Uncontrolled repeated cross-sectional study

Reported mode of journey to work on day before survey

California

(cashing out)^{w41 w42}

1992

State legislation requiring employers with at least 50 staff to "cash-out" the cost of rented parking spaces

Employees at workplaces in urban South California

Controlled repeated cross-sectional study

Reported mode of all journeys to work over five consecutive days

California (telecommuting)^w49

1993

Voluntary use of neighbourhood telecommuting centres as an alternative to commuting to their usual workplace

Registered users of telecommuting centres in California

Retrospective study using participants as their own controls

Reported mode of all trips recorded over two periods of three consecutive days

Camden-Islington^w19

NR

Site-specific advice to participating schools from a school travel coordinator

Pupils in primary schools in two London boroughs

Cluster randomised controlled trial

Reported mode of journeys to school on one day

Delft^w23-30

1982

Upgrading and increased connectivity (+3.3 km) of cycle route network

Households in suburbs of Delft

Controlled repeated cross-sectional study with nested panel study of a subset of respondents to both survey waves

Reported mode of all trips of residents aged 10 and over on one of a number of specified days covering all the days of the week

Detmold-Rosenheim^w31-33

1981

Bicycle-friendly demonstration project in two towns, mainly consisting of planning and building improvements to cycle route network (+31 km, +13 km respectively)

Households in both towns

Controlled repeated cross-sectional study

Reported mode of all trips of residents aged 10 and over on one of a number of specified days covering all the days of the week

England

(bypasses)^{w36 w37}

1992

Construction of bypasses, followed by a variety of traffic calming measures and enhanced walking or cycling facilities in each town centre

Residents of six small towns in England

Uncontrolled repeated cross-sectional study

Reported main mode of residents’ journeys to town centres

England

(20 mph zones)^w35

1996

Construction of 20 mph (30 km/h) zones, enforced using a range of engineering measures

Residents of neighbourhoods in six towns in northern England

Uncontrolled repeated cross-sectional study

Stated change in travel patterns

Eugene^w22

1994

Promotion of alternative transport modes through workplace transport coordinators and transport fairs, free bus passes, and rewards for staff using alternative modes

State employees in Oregon

Uncontrolled repeated cross-sectional study

Reported most commonly used mode for commuting

Frome^w10

2001

Individualised marketing of alternative modes of transport to households showing an interest in using them (TravelSmart: see Perth)

Households in Frome

Controlled prospective panel study with common baseline dataset

Reported main mode of all household trips, expressed in terms of estimated trips per person per year by mode

Glasgow^{w1 w2}

1998

Self-help pack to promote active commuting (Walk In to Work Out) containing written interactive materials based on the transtheoretical model of behaviour change, e.g. advice on choosing routes, personal safety, safe cycle storage, activity diary, map

Employees at three public sector organisations in Glasgow

Randomised controlled trial with intention-to-treat analysis

Reported time spent walking to work in seven-day recall physical activity diary, progression to higher stage of change, and prevalence of cycling

Gloucester^w11

2001

Individualised marketing of alternative modes of transport to households showing an interest in using them (TravelSmart: see Perth)

Households in a suburb of Gloucester

Controlled prospective panel study with common baseline dataset

Reported main mode of all household trips, expressed in terms of estimated trips per person per year by mode

Maidstone^w20

1994

Campaign using mass media and community activities to raise awareness of alternative modes of transport

Households on trunk route corridors approaching Maidstone and Tunbridge Wells

Controlled repeated cross-sectional study

Reported frequency of all household trips in a typical week by mode

Perth^w3-9

2000

Individualised marketing of alternative modes of transport to households showing an interest in using them (TravelSmart), using a tailored combination of e.g. public transport information, cycle route map, walking information booklet with motivational challenge chart, sometimes followed up with home visits

Households in South Perth and Victoria Park

Controlled repeated cross-sectional study of households using a mixture of survey sources to compile the baseline dataset, a combination of random and quota sampling for follow-up, and analysing outcomes in a sample representative of all local households irrespective of their interest or participation in the intervention

Reported main mode of all household trips, expressed in terms of estimated trips per person per year by mode

Phoenix^w21

1988

Campaign using mass media and community activities to promote voluntary no-drive day

Drivers living in Maricopa County and commuting to work

Uncontrolled repeated cross-sectional study

Reported mode of all commuting journeys

San Francisco^w45-47

NR

Neighbourhood-based car-sharing cooperative

Members and aspiring members in San Francisco

Controlled repeated cross-sectional study

Reported mode of all trips on any two days selected by the respondent

Stockton^w34

1985

New shared pedestrian and cycle route (4 km)

Secondary school pupils in Stockton

Partially-controlled repeated cross-sectional study

Reported usual mode of journey to school

Tampere^{w50 w51}

NR

Participation in a trial of the effects of walking and cycling to work on physical fitness and blood lipids

Car or bus commuters in Tampere

Randomised controlled trial [no actual intervention to promote modal shift except for participation in trial]

None [included for health effects only]

Trondheim^{w43 w44}

1991

Toll ring for motor vehicles inbound towards the city centre, Monday to Friday between 0600 and 1700

Households in Trondheim

Uncontrolled prospective panel study

Reported mode of all trips made by residents aged over 13 recorded in a one-day travel diary

Voorhout^w48

1997

Opening of a railway station in a commuter town for the first time

Households in Voorhout

Uncontrolled prospective panel study

Reported mode of (a) all activities and (b) work and school activities of residents aged over 12 on two consecutive days selected by the respondent

Århus^w12-14

1995

"Inveterate motorists" invited to try to use bike and bus as much as possible in exchange for free bike, free bus pass and other accessories and information for one year (Bikebusters)

Car commuters in Århus

Uncontrolled prospective panel study

Reported mode of all weekday trips over one week

Validity score

Study

Sample size ^a

Response

rate ^b

Follow up

rate ^c

Follow up

(months) ^d

Outcome measure

Findings ^e

Targeted behaviour change programmes

9

Glasgow^{w1 w2}

102 commuters

89%

66%

6

Reported time spent walking to work in seven-day recall physical activity diary

Among those who had not walked to work at the start of the study, those in the intervention group (n=14) reported spending a significantly greater mean time per week walking to work than controls (n=12) (125 mins vs 61 mins)

There was also a significant increase in the reported mean time spent walking to work per week, in favour of the intervention group, among those who already walked to work (intervention group (n=61): 52 min to 79 min, control group (n=43) 50 min to 60 min)

Analysis of covariance (using logarithms) resulted in an estimated average relative increase in the time spent walking to work at six months, for someone given the intervention, of 1.93 (95% confidence interval 1.06 to 3.52) times any increase in walking time for a corresponding control who walked the same amount at baseline

Reported time spent cycling

There was no difference in the reported average weekly minutes spent cycling between cyclists in the intervention group (n=9) and control group (n=9)

Progression to higher stage of change

A significantly larger percentage of the intervention group (49%) had progressed to a higher stage of active commuting behaviour change compared with the control group (31%: difference 18% (95% confidence interval 5% to 32%). At 12 months, 25% of the intervention group had progressed to the "action" or "maintenance" stages (95% confidence interval 17% to 32%)

7

Perth^w3-9

706 households

75%

NA

6

Reported main mode of all household trips, expressed in terms of estimated trips per person per year by mode extrapolated from a one-day travel survey

Intervention group: estimated absolute modal shift of +5.5% from a baseline of 13.9%. Changes in all mode shares were statistically significant whether using trips or persons as the denominator (walking: P<0.01 in either case; cycling: P<0.01 on trips, P<0.10 on persons; car driver: P<0.01 in either case; car passenger: P<0.01 on trips, P<0.10 on persons). Changes sustained at 18-month follow-up (but comparable control group data not shown)

Reported time spent walking and cycling

Estimated mean time spent walking increased from 10 to 13 minutes per person per day. Estimated mean time spent cycling increased from two to three minutes per person per day

Control group: estimated absolute modal shift of –2%. Based on reported proportions, the change in car share is likely to be significant (estimated 95% confidence intervalI: 0% to +6%) and the changes in walking and cycling shares are not ^f

9

Frome^w10

749 households ^g

74%

80%

3

Reported main mode of all household trips, expressed in terms of estimated trips per person per year by mode extrapolated from a one-day travel survey

Estimated absolute modal shift of +3.6% from a baseline of 31.1% after adjustment for changes in control group. Decrease in car driver share of all trips was statistically significant (P<0.05)

9

Gloucester^w11

624 households ^g

66%

76%

3

Reported main mode of all household trips, expressed in terms of estimated trips per person per year by mode extrapolated from a one-day travel survey

Estimated absolute modal shift of +4.4% from a baseline of 28.7% after adjustment for changes in control group. Decrease in car driver share of all trips was statistically significant (P<0.05)

7

Århus^w12-14

150 commuters

NR ^h

88%

11

Reported mode of all weekday trips over one week

Estimated absolute modal shift of +25.3% from a baseline of 18.4%. Insufficient data to judge statistical precision of results

6

Tampere^{w50 w51}

35 commuters

NR

96%

2.5

None [included for health effects only]

4

Adelaide^w15-18

403 households

NR ^h

35-50%

1

Reported frequency of, and time spent on, all trips in a seven-day travel diary by mode

Among households that participated in the intervention, there was an overall reduction in all trips of 8% and 11% in Dulwich and Christies Beach respectively. Car driver trips decreased by 10.2% and 14.6% respectively, and car passenger trips decreased by 9.4% and 8.6% respectively. Walking trips increased by 1.0% in Dulwich and decreased by 2.0% in Christies Beach. Time spent walking increased by 8.2% in Dulwich. Cycling trips decreased by 11.0% in Dulwich and increased by 20.9% in Christies Beach. Insufficient data to judge statistical precision of results

Agents of change and publicity campaigns

8

Camden-Islington^w19

714 school pupils

85%

NA

14

Reported mode of journeys to school on one day

Pupils in intervention schools were more likely to travel to school by car (24.1% vs 22.5%) and less likely to walk to school (69.9% vs 71.0%) than pupils in control schools. There was no significant difference in the odds of a pupil travelling to school on foot, by bike or on public transport between intervention and control schools (odds ratios and 95% confidence intervals: unadjusted 0.98 (0.54, 1.76); adjusted for baseline characteristics 1.20 (0.81, 1.82); adjusted for baseline and other covariates 0.98 (0.61, 1.59)

7

Maidstone^w20

761 households

20%

NA

24

Reported frequency of all household trips in a typical week by mode

Intervention area: average number of weekly journeys by car increased (12.74 to 12.82, NS), on foot decreased (5.02 to 4.95, NS), and by bike decreased (0.75 to 0.45, P<0.05)

Control area: average number of weekly journeys by car decreased (12.83 to 12.10, NS), on foot increased (6.72 to 6.85, NS), and by bike decreased (1.03 to 0.56, P<0.10)

5

Phoenix^w21

701 drivers

NR

NA

7

Reported mode of all commuting journeys

Estimated absolute modal shift of +1% following the current year’s intervention from a baseline of 5% (+3.5% from a baseline of <2.5% when compared to the previous year’s baseline survey; a proportion of this shift occurred between the waves of the intervention). Based on reported proportions, the changes over the two year period for all modes are likely to be significant (estimated 95% confidence intervals: car –2% to –6%, walking +2% to +4%, cycling 0% to +2%), but for the current study year only the change in walking share is likely to be significant (estimated 95% confidence interval 0% to +2%) ^f

4

Eugene^w22

263 commuters

NR

NA

9

Reported most commonly used mode for commuting

Estimated absolute modal shift of 0 from a baseline of 2-6% depending on area of residence. Insufficient data to judge statistical precision of results

Engineering measures

7

Delft^w23-30

1937 households

Before: 68% After: NR

NA

36

Reported mode of all trips of residents aged 10 and over on one of a number of specified days covering all the days of the week

In the main intervention area, cycling share increased from 40% to 43% of all trips, the frequency of bike trips increased by 4% and the frequency of car trips did not change. A comparison of similar trips made by a sub-panel of respondents in the intervention area who participated in both survey waves (a sample described by authors as "biased", sample size not reported) found a positive modal shift of 0.6% of all trips from a baseline of 66.2%; 8.8% of cycling trips after the intervention had been shifted from other modes, of which 4.4% came from walking and 3.3% came from the car. In a secondary intervention area which received only improvements to the bike route to the city centre, cycling mode share increased from 38% to 39%. In the control area, the frequency of car trips increased by 15% and the frequency of bike trips did not change. Insufficient data to judge statistical precision of results.

6

Detmold-Rosenheim^w31-33

Detmold:

583 households

Rosenheim:

598 households

Detmold: 53%

Rosenheim: 62%

NA

60

Reported mode of all trips of residents aged 10 and over on one of a number of specified days covering all the days of the week

Detmold: Estimated absolute modal shift of -7% (unadjusted) or –5% (adjusted for changes in age structure and car ownership) from a baseline of 41%. Insufficient data to judge statistical precision of results

Rosenheim: Estimated absolute modal shift of 0 (unadjusted and adjusted) from a baseline of 49%. Insufficient data to judge statistical precision of results

5

Stockton^w34

2946 school pupils

73%

NA

17

Reported usual mode of journey to school

Estimated absolute modal shift of –2% from a baseline of 81%. Based on reported proportions, the changes in car and cycling shares are likely to be significant (car: estimated 95% confidence interval +1% to +3%; bike: estimated 95% confidence interval –1% to –3%) and the change in walk share is not ^f

"Despite the overall decline in the numbers cycling to school, the schools within the ‘catchment area’ had cyclist percentages 2-6 times higher than those for the ‘control’ schools outside the catchment area (4 to 6% compared with 1-2%". The authors did not report a before-and-after comparison of cycling mode share in the notional intervention and control areas

5

England

(20 mph zones)^w35

200-350 residents

in each of six zones

NR ^h

NR

12

Stated change in travel patterns

In three of the zones a "significant proportion" (27%, 11% and 9% respectively) of respondents in the 12-month follow-up survey said that the intervention had made them more likely to walk. However, after comparing the reported frequencies of actual journeys by purpose and mode before and after the intervention, the authors concluded that the introduction of the 20 mph zones did not appear to have influenced the frequency or purpose of walking or car trips. Respondents did not indicate any increase in cycle use following zone implementation

4

Boston^w38-40

5449 commuters

31%

NA

24

Reported mode of journey to work on day before survey

Estimated absolute modal shift of +0.3% to +0.9% from a baseline of 6.0%. ⁱ Based on reported proportions, the change in car share is likely to be significant (estimated 95% confidence interval –4% to –7%) and the change in walking share is not. ^f

3

England

(bypasses)^{w36 w37}

1446 residents

NR

NA

12-28

Reported main mode of residents’ journeys to town centres

Estimated absolute modal shift of –3% from a baseline of 55%. Based on reported proportions, the change in walking share is likely to be significant (estimated 95% confidence interval 0% to –8%) and the changes in car and cycling shares are not ^f

Financial incentives

8

California

(cashing out)^{w41 w42}

1694 commuters

>90%

NA

12-36

Reported mode of all journeys to work over five consecutive days

Intervention workplaces: estimated absolute modal shift of +1% from a baseline of 2.8%. Overall change in distribution between all modes was significant at each workplace separately (P<0.01)

Control workplace: estimated absolute modal shift of –1%. Overall change in distribution between all modes not significant (p>0.10)

7

Trondheim^{w43 w44}

1900 households

77%

62%

12

Reported mode of all trips made by residents aged over 13 recorded in a one-day travel diary

Estimated absolute modal shift of –2.6% for all trips from a baseline of 35.9%. Insufficient data to judge statistical precision of this result

The mean number of reported daily trips per traveller decreased from 4.46 to 3.92 (-12.2%; P<0.05). Within this overall decrease, the percentage decreases in walking and cycling trips were greater than the decreases in car trips as driver and as passenger (-28.1%, -14.9%, -5.6% and –14.3% respectively). All decreases were statistically significant at P<0.01 except for that for car trips as driver, which was significant at P<0.05

The proportions of respondents who reported any walking or cycling trips decreased (walking –6.6%, P<0.01; cycling –2.6%, P<0.10)

Providing alternative services

7

San Francisco^w45-47

247 car share club members

22%

NA

9

Reported mode of all trips on any two days selected by the respondent

Estimated absolute modal shift of 0 from a baseline of 48.4%. Based on reported proportions, the difference in combined walking and cycling shares between intervention and control groups is likely to be significant (estimated 95% confidence interval –4% to -13%) and the difference in car share is not ^f

Car and combined walking and cycling shares in both intervention and control groups increased between the "before" and "after" surveys: car share increased from a pooled baseline of 5.2% to 22.2% (intervention) and 22.8% (control), and the combined walking and cycling share increased from a pooled baseline of 36.1% to 39.8% (intervention) and 48.4% (control). Author’s analysis of journeys excluding journeys in car share club vehicles found that the differences in changes in private car, walking, and cycling shares between the intervention and control groups were not significant (p>0.10)

7

Voorhout^w48

197 households

56%

59%

12

Reported mode of (a) all activities and (b) work and school activities of residents aged over 12 on two consecutive days selected by the respondent

Estimated absolute modal shift of +5.0% for all activities (from a baseline of 42.1%) and +1.9% for work and school activities (from a baseline of 29.0%). Changes in overall distribution of mode choice for both activity categories were significant (P<0.001) ^j

4

California (telecommuting)^w49

72 commuters

35%

NA

NA

Reported mode and distance of all trips recorded over two periods of three consecutive days

Estimated absolute modal shift of –0.2% from a baseline of 6.1% (statistical tests not reported). Absolute distance travelled by walking and cycling on telecommuting days was 24% lower than on control days ^k

Category

Study

Effect sought

Findings

Targeted

Glasgow^{w1 w2}

SF-36

The sample mean scores on three subscales of the SF-36 increased significantly (P<0.05) in the intervention group compared with the control group: mental health (from 72 to 76), vitality (from 57 to 64) and general health (from 71 to 76). There was no significant change in the sample mean scores on the other five subscales

Accidents

"There were no adverse effects noted from this intervention such as traffic accidents"

Targeted

Tampere^{w50 w51}

Fitness

Mean walking speed for commuting journeys increased from 5.8 km/h to 6.2 km/h and the mean cycling speed increased from 17.6 km/h to 20 km/h

Net changes in intervention group compared with control group: maximal aerobic power and maximum treadmill time increased significantly (+4.5%, P=0.02 and +10.3%, P<0.001 respectively). Heart rate and blood lactate at submaximal standard work load decreased significantly (P=0.04 and P=0.002 respectively)

Blood lipids

HDL cholesterol increased (+5%, P=0.06); no significant change in total cholesterol or triglyceride concentrations

Weight

No changes in either group

Targeted

Århus^w12-14

Blood pressure

Prevalence of "normal blood pressure" increased from 87% to 90%, prevalence of "high and slightly high blood pressure" decreased from 13% to 10%

Weight

No change in average weight overall (79 kg), but men lost an average of 200 g and women gained an average of 1 kg

Cholesterol

Prevalence of "normal" cholesterol level decreased from 66% to 61%; prevalence of "grey zone" cholesterol level increased from 18% to 23%; prevalence of "high" cholesterol level increased from 15% to 16%

Fitness

Prevalence of "poor" or "low" self-rated fitness decreased from 65% to 51%; prevalence of "average" self-rated fitness increased from 19% to 32%; prevalence of "good" or "high" self-rated fitness increased from 16% to 17%

Smoking

No change in smoking prevalence ("one third")

Health measures were assessed before and after the end of the project. No details of methods, participation rates or tests of statistical significance were reported

Engineering

Stockton^w34

Accidents

The absolute number of accidents to cyclists increased in both a notional catchment area for the cycle route and in a control area elsewhere in the town (catchment area: from 23 to 26; control area: from 34 to 46). No tests of statistical significance were reported. The authors reported a "statistically significant" shift in the distribution of all accidents from the catchment area to the control area

Engineering

England

(20 mph zones)^w35

Community severance

"There is anecdotal evidence from discussions held with groups of residents in the two largest zones that it was easier to cross the main roads after the speed limits had been reduced to 20 mph… No increase in adult or child street activity has so far been apparent… the numbers [reporting that they talked to neighbours and friends in the street] did not change significantly"

Noise

"The majority of respondents felt that noise levels had stayed the same as a result of the introduction of the 20 mph zone. In the Warrington zone, where the most stringent measures have been introduced, almost three-quarters of those questioned three months after implementation felt that noise levels had been reduced"

Accidents

The mean annual number of accidents decreased in all zones by between 32% and 100%. The overall decrease was described as "significant" in the summary of results but no tests of statistical significance were reported

Engineering

England

(bypasses)^{w36 w37}

Disturbance

The proportion of respondents who rated vibration, fumes and noise as "big problems" decreased from 8%, 13% and 9% (respectively) before the intervention to 4%, 5% and 4% after the intervention. No tests of statistical significance were reported

Accidents

The mean annual number of accidents decreased in all six towns. The annual number of accidents to pedestrians and cyclists decreased in five towns and increased in one town. No tests of statistical significance were reported

Category

Study

Reported findings

Evidence suggests positive

effects greater among

Targeted

Glasgow^{w1 w2}

"There were no gender, age or distance to work effects… However, more women than men responded to the opportunity to be involved"

Females

Targeted

Perth^w3-9

"There was a minor increase in the number of cyclists" [meaning that the increase in cycling trips was mostly the result of a larger number of bicycle trips per cyclist per day]. The intervention had no effect in the "not interested" group. The car share decreased in all age/sex groups studied, but particularly among women of working age (under 20: absolute decrease of 1% from a baseline share of 5%; males 20-60: absolute decrease of 3% from a baseline share of 38%; females 20-60: absolute decrease of 8% from a baseline share of 43%; over 60: absolute decrease of 2% from a baseline share of 14%)

Existing cyclists

Females of working age

Targeted

Frome^w10

The car share decreased in all age/sex groups studied (under 20: absolute decrease of 1% from a baseline share of 11%; males 20-60: absolute decrease of 2% from a baseline share of 25%; females 20-60: absolute decrease of 2% from a baseline share of 34%; over 60: absolute decrease of 2% from a baseline share of 30%)

Targeted

Gloucester^w11

The car share decreased by in all age/sex groups studied (under 20: absolute decrease of 1% from a baseline share of 16%; males 20-60: absolute decrease of 3% from a baseline share of 39%; females 20-60: absolute decrease of 3% from a baseline share of 30%; over 60: absolute decrease of 1% from a baseline share of 15%)

Adults of working age

Targeted

Århus^w12-14

There was a higher cycling mode share (and combined walking and cycling mode share) for men than for women at follow-up (combined shares approximately 53% and 49% respectively, estimated from graphs). Participants from households with more than one car had a lower cycling mode share (and combined walking and cycling mode share) than those from households with only one car (combined shares approximately 48% and 53% respectively, estimated from graphs). The authors also reported that "use of the bicycle increases with age; use of the bicycle is independent of income; use of the bicycle drops the larger the number of adults in the household" [data not reported]. The authors also reported a small negative modal shift in trips made by participants’ spouses, who now had greater access to the family car (estimated absolute shift of –4% from a baseline of 35%)

Males

Households with only one car

Older adults

Smaller households

Targeted

Adelaide^w15-18

In stepwise logistic regression, household composition was shown to have a minor influence on the estimated reductions in vehicle emissions resulting from the intervention. The higher the proportion of retired persons in the household and the older the mean age for adults, the less the estimated reduction

Households without older adults

Engineering

Delft^w23-30

"The increase of bicycle mobility cannot be ascribed to a larger number of people using the bicycle, but is the result of a larger number of bicycle trips per cyclist per day. This finding is consistent with the findings… that many non-cyclists are captive users of their modes… and do not have freedom of choice. Bicycle use by men increased more than that by women"

Existing cyclists

Males

Engineering

Detmold-Rosenheim^w31-33

Modal shifts were stratified by age, sex, occupational group and bike and car ownership in both towns (in this analysis, the car mode share included motorised two-wheelers). In both towns, the modal shift was more positive (or less negative) among males, non-bike owners and car owners than among females, bike owners and non-car owners respectively. In Rosenheim, the modal shift was more negative among those aged over 65 compared with those in younger age groups, and among pensioners compared with those in other occupational groups.

Males

Non-bike owners

Car owners

People below retirement age

Engineering

Stockton^w34

Boys were more likely than girls to cycle to school, both before and after the intervention

Males

Engineering

England

(bypasses)^{w36 w37}

The decrease in walking mode share was seen in both males and females (absolute decreases of 2% and 5% respectively), in adults both under and over the age of 60 (4% and 6% respectively), and in people with access to no car, one car, and more than one car (2%, 1% and 6% respectively)

Financial

California

(cashing out)^{w41 w42}

The policy was said to have had a redistributive effect on income: some firms had previously offered higher parking subsidies to higher-paid employees, but after the intervention the benefits were offered at a flat rate to all employees