Response to the IAM: Analysis of traffic signal compliance, by mode

I started this little project back in the summer of 2011 and intended to publish it when I got around to studying( a greater number of locations to improve the reliability of the results. In the mean time however, The Institute of ‘Advanced’ Motorists has published a ‘study‘ along similar lines. The IAM’s study makes this incredibly small scale study look positively academically rigorous in comparison, so I have decided to share my findings early. The counting involved was incredibly tedious anyway…

1. Introduction

Traffic signal non-compliance by cyclists, commonly referred to as ‘red light jumping’ (RLJ) is widely perceived as a problem by non-cyclists in the UK. Despite the popularity of this belief there is little in the way of empirical evidence to confirm or refute this claim, with the exception of a study in London by Transport for London in 2007 [1], which did not look at non-compliant behaviour by operators of other classes of vehicle, showed cyclists’ traffic signal non-compliance rate at 16 per cent.

At the time of publication, there does not appear to have been a study which looks at traffic signal non-compliance rates in all vehicle types on UK roads which takes the different characteristics of vehicle types into account. There are notable differences in size, speed and mass between different classes of vehicles, the most pronounced of which is the difference between the characteristics of cycles and motor vehicles. In order to provide meaningful, comparable results it is important that the experiment be designed to take these differences into account. Presented here is a method for comparing traffic signal non-compliance by vehicle type which takes into account the differences in vehicle characteristics to allow useful comparisons between modes to be made for the first time.

2. Materials and Methods:

2.1 Choice of suitable locations:

Locations for the study of traffic signal compliance were chosen to meet the following criteria; Locations had to include a four-way signalised junction and provide a location where video footage could be recorded without this being obvious to the subjects studied. The locations chosen were within a 5 km radius of Sandbar on Grafton Street in Manchester, UK, for ease of transportation. Video footage was captured at peak times (between 7am-9am & 5pm-7pm) on weekdays. This was due to the levels of cycling in Manchester being too low to be detectable at other times of the day and days of the week. Transport between locations where video footage was recorded was performed via a 2009 Raleigh Tourist De Luxe.

2.2 Video footage capture:

Video footage was captured using a NV-GS27 Mini-DV video camera (Panasonic) mounted on a tripod in a position obscured from the view of road users to prevent the subjects becoming aware of the study and subsequently altering their behaviour. Video footage was captured in 576i format and transferred to PC for playback using Nero Vision software.

2.3 Analysis:

The total vehicle throughput of the southbound lane(s) of two were counted for the following modes; bicycle, car, taxi, bus, van, HGV and motorcycle, where, “car” includes private hire vehicles and vans whose form factor is the same as that of a car and ‘taxi’ includes Hackney Carriages only. Traffic signal compliance was monitored as the number of vehicles whose operators were, “Able to ignore the red ‘stop’ signal if they so wished”; for motor vehicles (except motorcycles) this was recorded as the number of vehicles whose ability to pass through a red ‘stop’ signal was not impaired by the presence of another motor vehicle positioned in front of their vehicle. For example, a single carriageway where motor traffic had formed two queues at a red ‘stop’ signal would have two motor vehicles whose operators are able to proceed through the red ‘stop’ signal; the two vehicles at the front of the queue. In the event of a motor vehicle operator proceeding far enough beyond the first ‘stop’ line at the junction during a red ‘stop’ phase to permit the subsequent motor vehicle operator to also pass through the first ‘stop’ line, the number of vehicles whose operators were, “Able to ignore the red ‘stop’ signal if they so wished” was subsequently increased. Due to the much reduced width of their vehicles, cycle and motorcycle operators are by and large able to position their vehicles at the front of the queue, regardless of the presence of other vehicles waiting at the red ‘stop’ signal. In order to make reasonable inter-modal comparisons, all cyclists who arrived at the junction during the red ‘stop’ phase were recorded as being, “Able to ignore the red ‘stop’ signal if they so wished.”

Figures were recorded in a Microsoft ‘Excel 2010’ spreadsheet, which was used for all subsequent analyses.

3. Results

Figure 1a: Part one of an approximately 20 minute film observing the behaviour of southbound traffic participants at the signalised junction between Wilmslow Road and Wilbraham Road/Moseley Road in Manchester, UK

Figure 1b: Part one of an approximately 20 minute film observing the behaviour of southbound traffic participants at the signalised junction between Wilmslow Road and Wilbraham Road/Moseley Road in Manchester, UK

Figure 2a. Part one of an approximately 20 minute film observing the behaviour of southbound traffic participants at the signalised junction between Upper Brook Street and Hathersage Road in Manchester, UK

Figure 2b. Part two of an approximately 20 minute film observing the behaviour of southbound traffic participants at the signalised junction between Upper Brook Street and Hathersage Road in Manchester, UK.

Two junctions in Manchester, UK, were filmed each for a period of approximately twenty minutes. Locations were chosen based on criteria described in section 2.1. The first of these was at the intersection of Wilmslow Road and Wilbraham Road/Moseley Road, the video collected at this location is shown in figures 1a and 1b. The second of these locations was the intersection of Upper Brook Street and Hathersage Road, the video collected at this loaction is shown in figures 2a and 2b. All video footage was collected between 12/07/2011 and 16/07/2011.

The traffic signal compliance rates observed in the footage were noted for each mode of transport for the southbound lanes as described in section 2.3. The non-compliance figures for the intersection between Wilmslow Road and Wilbraham Road/Moseley Road is shown in Table  1. The non-compliance figures for the intersection between Upper Brook Street and Hathersage Road is shown in Table  2. The combined non-compliance figures for both intersections is shown in Table  3.

Total number Able to proceed through red Proceeded past first stop line Proceeded through junction Percentage proceeded past first stop Percentage proceeded through junction
Bicycle 41 14 3 1 21.4 7.1
Car 246 21 6 0 28.6 0.0
Taxi 6 1 0 0 0.0 0.0
Bus 16 0 0 0 0.0 0.0
Van 10 1 0 0 0.0 0.0
HGV 0 0 0 0 0.0 0.0
Motorcycle 2 0 0 0 0.0 0.0

Table 1. Compliance with traffic signals at the intersection between Wilmslow Road and Wilbraham Road/Moseley Road, by vehicle type.

Total number Able to proceed through red Proceeded past first stop line Proceeded through junction Percentage proceeded past first stop Percentage proceeded through junction
Bicycle 18 9 1 0 11.1 0
Car 239 12 4 0 33.3 0
Taxi 7 0 0 0 0.0 0
Bus 4 1 1 0 100.0 0
Van 9 0 0 0 0.0 0
HGV 0 0 0 0 0.0 0
Motorcycle 1 0 0 0 0.0 0

Table 2. Compliance with traffic signals at the intersection between Upper Brooks Street and Hathersage Road, by vehicle type.

Total number Able to proceed through red Proceeded past first stop line Proceeded through junction Percentage proceeded past first stop Percentage proceeded through junction
Bicycle 59 23 4 1 17.4 4.3
Car 485 33 10 0 30.3 0.0
Taxi 13 1 0 0 0.0 0.0
Bus 20 1 1 0 100.0 0.0
Van 19 1 0 0 0.0 0.0
HGV 0 0 0 0 0.0 0.0
Motorcycle 3 0 0 0 0.0 0.0

Table 3. Compliance with traffic signals for all intersections, by vehicle type.

4. Discussion

The data presented here present a radically different picture of traffic signal compliance by vehicle type than that which is perceived by the general population of the United Kingdom. Whilst there is a common belief, particularly amongst tabloid-reading members of the British public that cyclists as a group are particularly unlikely to comply with traffic signals, this perception does not correlate with the results of this empirical study.

Only one vehicle during the two periods of observation passed completely through a junction whilst a traffic signal was in a red ‘stop’ phase, and this vehicle was a bicycle. Fifteen vehicles proceeded to pass beyond the ‘stop’ line whilst a traffic signal was in a red ‘stop’ phase, of which four were bicycles, ten were cars and one was a bus. Whilst it may appear that cycles are over-represented in this group due to their relatively low number (10% of traffic during the observation periods) their physical characteristics allow a disproportionately high number of cycles to reach the front of the queue, giving their operators the choice of whether or not to comply with the traffic signal. This is a choice which the operator of a motor vehicle in a queue of other motor vehicles does not have, with only motor vehicle operators at the front of queues able to choose not to comply with a traffic signal if they so wish. The design of the analysis used in this study, described in section 2.3 takes this important consideration into account.

The results presented here show that during the observation period, 17.4 per cent of cyclists who were able to choose whether or not to comply with traffic signals chose not to do so, compared with 30.3 per cent of motorists and 100 per cent of bus drivers.

5. Conclusions

This study represents the first meaningful empirical evidence to challenge the popular perception that cyclists have uniquely strong disregard for compliance with traffic signals. The results of this study are broadly in line with previous observations of traffic signal compliance by cyclists [1]. The method presented here for the analysis of traffic signal compliance by a variety of different vehicle classes allows, uniquely, for the different physical characteristics of differing vehicle types to be accounted for in the study, allowing a fair an meaningful comparison of non-compliance rates to be made for different modes. It is hoped that this method can be used in future studies, to avoid the risk of an ‘apples vs. oranges’ comparison being made which produces results skewed in favour of one transport mode or another due to a failure to account for the different characteristics of different vehicle types.

It is hoped that further, larger-scale studies using this method would build upon these results whilst accounting for a wider range of locations, junction designs and road types. It is hoped that a larger sample size will provide meaningful data on traffic signal compliance by modes which were not observed in sizeable quantities during this study, such as buses, motorcycles, vans and HGVs.

6. References

[1] TfL: RNPR Traffic Note 8, Proportion of Cyclists Who Violate Red Lights in London (June 2007)

Advertisements

25 thoughts on “Response to the IAM: Analysis of traffic signal compliance, by mode

  1. Nice work. I still think the statistics are a bit low to make meaningful conclusions but certainly there is something there.

    *puts academic reviewer hat on* Is it nessessary to quote the absolute numbers to a decimal place. I’m not sure it’s possible to have a fraction of any of the vehicles involved. Unless you have an argumentative couple on a tandem.

  2. Interesting, and I can see a good, simple point- if cyclists jump lights it’s ‘cos they’re in a position to do so. But do you think think your evidence will have the slightest impression? Sorry to be negative, and I don’t really know what to say. I’ve just spent so long pointing out that however much cyclists break rules (if they do) they don’t kill people. Not much reception so far. Sorry.

    • Of course not, it merely serves to highlight what we all already suspected, the accusation that cyclists are particularly unlikely to comply with traffic signals is based on existing anti-cyclist prejudice. But it is still nice to have something like this to point people towards when they do start spouting such bile about cyclists.

  3. Nice work. Though I agree that the numbers are a bit low, it’s great to see such work going on at grass roots level. Just by taking up the challenge, you’ve raised some interesting issues, such as the ability of cyclists to potentially jump lights, unlike motorists. There’s more to do in this direction – traffic lights are designed for motorised traffic, not cyclists – so it can be argued that this alone makes cyclists less likely comply with the idea of “playing” the role of vehicular cyclists. The legal alternative at a red light, after all, is to push your bike across the road a la pedestrian mode. It’s a small step from this to “scootering” across, and another small step to arguing “may as well cycle” if there is no traffic. This is the logic, I would suggest, behind at least some RLJing.

    Similarly, pavement cycling, again demonised, has a clear logic – lack of cycling infrastructure where it is most needed. Most pavement cycling that I witness is alongside busy roads where any risk assessment worth its salt would conclude “better safe than legal”. Behind all the statistics is a crucial issue – we are not comparing like with like when comparing cyclists with motorists using what is motorist-centric infrastructure.

    PS It might be of interest for you to have a look at our study of cyclist-pedestrian interactions in a pedestrian zone in Darlington from a few years back. We took a similar approach, using video evidence.

    http://www.grassick.net/cycling_downloads/CCTV%20STUDY.pdf

    A gathering of the grass roots studies somehwere on the net?

  4. I note as a cycleist that I am far more likely to (be tempted to) ‘RLJ’, even straight across the junction, *outside* rush-hour… at night or on a Sunday morning when there is no or little other detectable traffic.

  5. The IAM’s study isn’t worthy of a response.

    Advanced motorists should concern themselves with improving the standard of motoring of non-advanced motorists, many of whom routinely break the speed limit, use phones, park on pavements and generally act in a loutish, careless and often criminal manner resulting in thosands of deaths and serious injuries every year.

    That the IAM thinks it important to devote even one minute to the alleged ‘problem’ of red light jumping cyclists demonstrates they are advanced in one thing only – the sense of persecution many motorists feel because of the expensive and miserable time they have behind the wheel and because they aren’t allowed to do as they like all the time.

  6. I am not sure which question or problem your study is addressing. You start with “Traffic signal non-compliance by cyclists, commonly referred to as ‘red light jumping’ (RLJ) is widely perceived as a problem by non-cyclists in the UK.” and later you refer to “the popular perception that cyclists have uniquely strong disregard for compliance with traffic signals.” There are two separate ideas here, One is that non-cyclists tend to see non-compliance by cyclists as a problem. The other is that many people believe that cyclists are more likely to ignore traffic lights than other types of road user.

    You don’t provide any evidence that either of those descriptions is true, but your study does try to estimate the frequency of cyclist non-compliance relative to other vehicles’ non-compliance and I think you imply that if people only knew the actual frequencies their perceptions and beliefs would change. Have I understood this accurately?

    There is at least one observational study that has looked at compliance by cyclists. Published in 2007 by Transport for London just here: http://www.tfl.gov.uk/assets/downloads/businessandpartners/traffic-note-8-cycling-red-lights.pdf With more resources they have observed more junctions and for longer. They provide figures that support the suggestion that quite a lot of cyclists do ignore red lights at junctions in parts of south London. Compliance in their study varied by gender and by time of day. But figures of somewhere between one in eight to one in three cyclists were observed choosing to ignore red lights across the types observed.

    That study says nothing about other vehicles but I wonder why you introduced that refinement to your study? My reading of the incoherent rants about red-light jumping is that they are genuine expressions of some kind of anxiety about real experience and real emotions. They are not logical arguments that could be settled by arithmetic. The people (like me) who find rule breaking by cyclist disturbing have real anxiety that is frequently triggered by cyclists who don’t follow the rules. It is a phenomenological reality.

    The real research problem, to my mind, is our failure to understand or reduce anxiety levels on our roads. Questions of law enforcement, road design, access to training and patience with others’ concerns (misguided or otherwise) seem to me to be far more important that arithmetic and counting. Numbers have their humble place, but in social and psychological research they are most effective when tempered with careful qualitative study. In political matters the simple facts have great potential for causing arguments and it’s only rarely that they help to resolve them.

    • I really don’t understand your point, samsaundersbristol. You are saying that facts don’t matter, only impressions or emotions? That’s a bit of a strange idea. Surely both matter, and it’s important to have true facts measured with proper methodologies to counter the emotional arguments. It seems like you haven’t fully read the Original Post, or else why do you re-quote same TfL study as was already referenced?

      • Thanks for the questions Dave. I am saying a number of things, Firstly facts are frustratingly elusive. Secondly, facts presented as part of a campaign are rarely decisive: feelings and values are generally stronger than facts. Thirdly, I am saying that the anxiety expressed by people about cyclists flouting traffic regulations is a real anxiety. The fear or irritation they feel when a cyclist startles them on a pavement or passes them at red light is probably not contrived. We cannot remove those feelings by telling them they need not worry or by pointing out that cars do worse things. It’s a pragmatic thing.

        I’m afraid I missed the reference to the TfL study in the blog. My mistake.

        My main concern was to point out that a better case for better roads could be made if we stopped trying to justify the kinds of rule breaking that upset people so much. It would probably do cycling a lot of good if we focussed on positive demands that helped everybody.

  7. Red Alert: Motorists Drive Through 278 Red Traffic Lights a Minute – 22nd August 2011
    5.2 million drivers have passed through a red traffic signal in the last month
    Amber gamblers: Almost one in ten drivers doesn’t slow down for amber lights
    New research from Direct Line car insurance reveals that motorists are driving through 12 million red lights each month on British roads – the equivalent of running 278 red lights every minute *.
    Over 5.2 million (14 per cent) motorists admit that they drive through an average of two red traffic lights each month. Over 760,000 (two per cent) motorists habitually drive through red lights if they feel the road is clear and there is no traffic.
    As a result of their reckless behaviour at red lights, drivers risk clocking up over 36 million penalty points (three for each offence) and fines from fixed penalty notices totalling £721** million every month.
    A worryingly high number of motorists, dubbed ‘amber gamblers’, are putting their lives and other road-users at risk by failing to slow down for traffic lights about to turn red. Almost one in ten (nine per cent) motorists don’t reduce their speed when approaching amber lights, with four per cent admitting to putting the accelerator to the floor to race through the lights.
    Over 1.5 million (four per cent) motorists admit to getting ‘a thrill’ when driving through amber traffic lights that they know will have turned red before they have passed through the crossing or junction. Over 500,000 drivers risk being ‘rear ended’ as they stamp on their brakes as an automatic response whenever they see amber traffic lights…..
    http://www.directline.com/about_us/news_22082011.htm
    http://www.webcitation.org/65guz89OR

  8. It would definitely be interesting to see this approach applied somewhere with denser traffic. I see all kinds of traffic jumping red lights but I wonder if that’s because I am “programmed” to look out for it with my cyclist hat on…

  9. Nice work.

    Perhaps another way to analyse the data would be vehicles (including bicycles) passing the signals at “stop”, which strictly speaking includes the amber phase. You could then collect variables data in the form of time after the amber signal starts that a vehicle crosses the line, and could then perform a significance test for smaller sample sizes. This can be used to compare junctions as a whole, as well as the individual sub-populations.

  10. Pingback: Auto Express inadvertently proves that drivers of cars are more likely to jump red lights than cyclists | As Easy As Riding A Bike

  11. Pingback: Driving and (smart)phone use | Chester Cycling

  12. Pingback: In appreciation of Andrew Mitchell | Chester Cycling

  13. Pingback: Response to LTDA: Analysis of signal compliance, by mode | As Easy As Riding A Bike

Leave a comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s