The 'perception of risk' vs 'real risk' of riding a bike in Melbourne

With increasing urban population and transport congestion it is important for a city to have bike riding as an attractive travel choice. The biggest barrier for people to choose riding a bicycle is the perceived lack of safety. How safe, or unsafe, people feel about riding a bike has a direct relationship with their willingness to ride.

With Melbourne still striving to become a ‘world-class cycling city’, we wanted to explore how our perceptions of riding risk compared with the real risk of riding at locations across Melbourne. For example, were there locations with a high perception of risk and low real risk?

With CrowdSpot collaborators The Squeaky Wheel, we came up with BikeSpot and applied for a TAC Road Safety Community Grant. We also had the support of 13 Government Partners and 16 Community Groups.

The BikeSpot Project launched on 14 Feb 2016 and gave people had the option to add either ‘Safe Spots’ or ‘Unsafe Spots’ and also the ability to ‘Support’ (vote) and add comments on Spots added by other users.

The map was open for 2.5 months until 30 April 2016. We received over 8,000 submissions from over 2,000 people.

The Top 'Unsafe' and 'Safe' Spots

By adding the number of comments and ‘supports’ a spot received we can determine the top 10 perceived ‘Unsafe’ and ‘Safe’ Spots across the entire city.

Perception of risk Vs real risk

To further understand differences in the 'perception of risk' and 'real risk' of riding a bicycle in Melbourne we can compare the top 10 unsafe spots with historical crash statistics (VicRoads CrashStats).

It is important to note that VicRoads CrashStats data is generated through police crash reporting. As a result, these crashes tend to involve more serious injuries and therefore minor crashes often go unreported. CrashStats is still thought to be an accurate representation of bicycle crash locations, as was confirmed in the previous Cycle Stories project from 2013, where crowdsourced crashes were compared to CrashStats.

To define whether locations are deemed low, medium or high real risk we have applied the following categories based on the number of crashes that took place at that location under the same circumstances:

Low real risk: < 5 crashes

Medium real risk: 5–15 crashes

High real risk: > 15 crashes

For a more detailed analysis of each location view the original post on Medium

Seven out of the top 10 Unsafe Spots have ‘low real risk’ when compared to CrashStats data. Five of the low real risk locations are associated with mixing different transport modes. These locations present a high perception of risk but low real risk, possibly due to increased caution when riding and their relative predictability (you know which direction a car is travelling in a lane). This is in contrast to high real risk locations such as Chapel St and Sydney Rd, where dooring crashes are sudden and very unexpected.

Not surprisingly, the majority of Top 10 Safe Spots are all located along routes with strong mode separation.

Factors increasing peoples perception of riding risk

Mixing with different transport modes

Changing traffic conditions

Fear of unpredictable circumstances

Do people actually ride at these locations?

There is a common thought that crashes might be lower at locations where there is a high-perceived risk due to people choosing not to ride at those locations. The map below includes Strava volume data from a sample week in late 2015. The blue lines indicate low rider volume and orange/red lines indicate high rider volume. The majority of the top BikeSpot locations correspond with high rider volume numbers.

Thus, while it appears that people continue to ride along bicycle corridors with high-perceived risk (likely due to convenience), a relatively low crash history may also be a result of extra caution being taken at these locations.

Top Project Findings

• There is a great variation between the ‘perceived risk’ and ‘real risk’ amongst the top 10 'Unsafe' Spots. This is likely due to greater caution being taken when riding at these locations. 
• Bike riders are willing to accept certain levels of risk to achieve the major benefit of convenience.
• The results support mode separation as the majority of top 10 Safe Spots are located along routes with strong mode separation.
• Amongst existing confident riders, there are environments in Melbourne that are either uncomfortable or very stressful to ride in, despite not having a reported high crash history.
• Locations with a high-perceived risk can prevent existing riders from choosing to ride more often and deter others interested in riding from riding altogether.
• Crash statistics only tell part of the story. To become a world-class cycling city we need to also improve locations where people don’t feel safe.

Links to further analysis

The full article was first published on Medium

The full project report can be downloaded

You can explore the data via the project spatial visualisation

BikeSpot in the media

The Age (11 July)

Channel 7 News (11 July)

ABC 774 Radio Interview (12 July)

Crikey (13 July)

The Herald Sun (21 July)