Cycling participation is promoted by all three tiers of government as a space efficient transport mode and as a means to maintain physical and mental health. Governments commit varying levels of expenditure to encouraging cycling through infrastructure, education and promotion. HOwevber, determining whether these investments are achieving the desired outcomes can be tricky, particularly if we’re interested in affecting population-level changes. These wider changes will be affected by many things, including social trends, changes in local and regional traffic levels and even demographic shifts1.
Common approaches to measuring cycling activity include:
- automatic and manual counts of bicycle riders at fixed points in the network, often along shared paths or cycleways which are commonly used by riders,
- the Census, which obtains data on modes used to travel to work on a single day in August every five years (most recently in 2016), and
- the Cycling Participation Survey (CPS).
The CPS is one of the few surveys that has been conducted continuously since 2011, and so provides a method of tracking cycling participation over recent years. The survey fieldwork is conducted by Market Solutions and analysis by CDM Research.
We differentiate here between cycling travel (synonymous with activity or trips) and participation:
- travel is the amount of cycling that is occurring; that is, the number of trips made by bicycle - or even the kilometres or minutes ridden, over some period of time, and
- participation is simply a Yes/No question as to whether the individual has ridden a bicycle at all over some period (e.g. the past week, month or year).
Measuring travel isn’t trivial; there are very significant definitional and survey issues in tracking cycling activity such that probably the only practical way of doing this is via a large scale travel survey like VISTA or HTS. By contrast, participation is much easier to measure and less likely to be subject to recall issues. It’s fairly easy to ask someone when they last rode a bicycle - but much harder to ask how many bike trips they’ve made. The downside is that participation doesn’t provide a direct measure of cycling activity. And, in theory, we may measure no change in participation but an increase in activity if pre-existing bike riders were simply to ride more often.
What is the Cycling Participation Survey?
“The strategy’s aim is to double the number of people cycling in Australia by 2016.”
This target is interpreted as a participation measure - that is, the denominator is people not trips. By this definition it is possible that, for example, cycling travel may increase even if participation were not to change (i.e. as a result of those already riding doing so more often).
The survey was repeated biennially for all states and territories between 2011 and 2017. The results for the last year the survey was completed nationally are available from Austroads. The survey suggested that cycling participation has declined marginally at the national level since 2011, although some jurisdictions have bucked the trend and reported steady or increasing participation. Clearly, this finding is contrary to the objectives of the strategy.
What is the survey methodology?
The CPS is conducted as a telephone interview with randomly selected households. Both landline and mobile phone numbers are used so as to have coverage of non-landline households. While it would be significantly cheaper to conduct the survey online it would be far less representative of the population.
An important characteristic of the survey is that it seeks to obtain coverage of all age groups. To do this the main respondent (that is, an individual randomly selected aged 15+ from the household using the “next birthday” method3) is asked about cycling participation for themselves and all other household members aged over two (it is assumed those aged under two do not ride). This coverage of young children is significant; this group has cycling participation rates much higher than adults such that not covering this cohort would significantly underestimate overall cycling participation.
What is defined as cycling participation?
This isn’t as simple as it may appear. An individual is defined as having participated in “cycling” if, in the past week, month or year they have:
- ridden a bicycle in a public place (e.g. road, footpath or in a park) and/or in a private place (e.g. their backyard, on a farm),
- the bicycle may have more than two wheels (e.g. tricycle or 3+ wheeled cargo bicycles), and may have a motor (e.g. e-bikes) or even an engine - but any vehicle that requires registration is excluded (e.g. mopeds and motorcycles),
- the bicycle need not have pedals - for example, balance bikes used by small children are included but scooters are excluded, and
- for the person to have “participated” in cycling they must have contributed to propelling the bicycle - so, for example, a young child in a trailer or seat on an adult bicycle ridden by a parent would be excluded but the parent would be included.
This fairly broad, and indeed arbitrary, definition of cycling covers much more than cycling as a form of transport, and indeed even cycling in public places. From a health perspective whether the cycling occurs in a backyard or on the streets is irrelevant, but from a transport perspective it is likely that only riding in public - and then only on roads and paths that is likely to be of interest.
What does the participation survey tell us?
The picture at a national level is one of steady declines in participation between 2011 and 2017 (Fig. 1). These decreases are most clearly illustrated when measuring participation over the previous month and year; around 27.1% of Australians were estimated to ride at least once a month in 2011 compared to just under 22% in 2017, and from 40% to 34% when measured as riding at least once over the past year.
These declines are fairly consistently observed across the eight states and territories, albeit with some variation across the years (Fig. 2). That aside, the evidence is abundantly clear that the National Cycling Strategy target of doubling cycling participation was not achieved.
What has been happening at the local level?
A number of local governments have commissioned the survey in their areas over multiple years. Doing so enables analysis at the local level which is not feasible using the state or national data. The data for these councils, anonymised into general areas, is shown in Fig. 3. While trends are sometimes difficult to ascertain, and often changes across years are within sampling uncertainty, we suggest the following trends are evident:
- For the Inner City 1 location there was growth in participation between 2011 and 2013, with stable to slight declines since that time. This is fairly consistent with the investments made by that council, which involved considerable new infrastructure in the 2010-12 period but with only marginal further improvements since that time.
- Inner suburban 2 has experienced significant residential and employment growth in their urban core, such that it is possible the changing demographics of this area has contributed to the observed growth (there being little in the way of dedicated cycling infrastructure).
- The Outer Suburban 1 site has not investment in significant infrastructure over the period, such that the steady decline in participation is fairly similar to the wider picture across Australia.
- Overall, the picture is nuanced and trends fairly limited. This points to the challenge of affecting population-level change; to achieve such widespread change requires a level of investment, and indeed political committment, that is lacking. However, this is not to say limited investments at a sub-regional level cannot encourage increased cycling participation along confined corridors.
Who rides a bike?
Males are far more likely to ride than females; around 20.4% of males ride in a typical week compared to 10.7% of females (Fig. 4). Cycling participation appears to have declined among both males and females since 2011.
Age is strongly correlated with cycling participation; young children aged under 10 are most likely to have ridden recently, followed by teenagers (Fig. 5). There’s a precipitous drop in cycling participation as children transition to adulthood - and in many cases a driving licence. The key finding here is that cycling participation is greatest among a seemingly invisible group (children) and not the highly visible stereotype of the middle aged male in lycra.
Another way of looking at these trends is to examine the changes across age groups between genders, as shown in Fig. 6 for data from 2017. The precipitous drop between teenagers and young adults is evident, but so too is the decline in participation among teenage girls. At least when measured over the past week and month this is a profoundly different trend compared with boys, who maintain or even increase their participation rate as teenagers. Moreover, girls aged under 10 have essentially identical cycling participation to boys - the gender difference only appears at later ages.
The changing participation rate cannot be attributed to a reduction in household access to a working bike; as shown in Fig. 7 the proportion of households across Australia that didn’t have access to at least one working bicycle was 44.9% in 2011 and 44.2% in 2017. The measured difference is within the sampling error.
When bicycle ownership extends beyond around half of households the cycling participation rate appears to increase (Fig. 8), but clearly a large proportion of households with acccess to a bicycle use it only rarely.
Respondents to the survey who had ridden in the past month were asked whether they had ridden at least once for recreation and/or for transport. More than 80% had ridden for recreation and only around 30% had ridden for transport (e.g. commuting to work or shopping). The relatively low transport usage is typical of cycling cultures where cycling is considered a peripheral mode. Countries with high cycling rates, such as the Netherlands and Denmark, tend to have much higher levels of transport cycling. Nonetheless, if we consider that just under a quarter of the population ride in any one month, and that around 80% of these do so for recreation, this makes cycling one of the most popular forms of physical activity - and certainly more popular than many organised sports such as football or cricket.
At current cycling rates the proportion riding for transport seems to only weakly be related to the overall cycling participation rate (Fig. 10). We suspect this be related to two factors:
- in middle and outer suburban areas the cycling participation rate is driven by the presence of families, from which the children ride regularly (and, presumably, the parents are also more likely to ride than in the absence of children), and
- in the inner city areas there is a much higher propensity to ride for transport but this is somewhat rebalanced by the lower proportion of families.
What could goverments do to encourage cycling?
The extended version of the survey asks respondents who had ridden in their area at least once in the past year a number of questions about how they perceive riding in the area. Among these questions respondents were asked to rate a number of interventions councils and State governments could consider to improve conditions for riding. These surveys almost always pointed to the provision of more off-road paths and cycleways as the highest priority (Fig. 11). Lower local speed limits were almost always rated lowest, probably reflecting the limited community understanding the impact speed has on safety.
For example, the young tend to ride more than the elderly. So, in the absence of any other change, the ageing population in Australia of itself would be expected to reduce the cycling participation rate.↩
The person in the household who picks up the phone is unlikely to be random. To redress this a simple method to ask to speak to the person aged 15+ with the next birthday in the household. This is the next birthday method.↩