A CROSS-SECTIONAL STUDY OF THE PREVALENCE OF USE OF VISIBILITY AIDS BY EDINBURGH CYCLISTS AND THEIR ASSOCIATION WITH SELF-REPORTED ACCIDENTS / NEAR ACCIDENTS

STUART CAMPBELL

UNIVERSITY OF EDINBURGH COLLEGE OF HUMANITIES AND SOCIAL SCIENCE MORAY HOUSE SCHOOL OF EDUCATION BSc (HONS) APPLIED SPORT SCIENCE DISSERTATION SPRT10033 2012

Comments below from DdF, Spokes, based o a fairly quick read.


Really interesting data.

Obviously the sample is to a large extent self-selected, and also it is taken at a specific time of year - Feb/March 2012 (which could affect some results significantly).  Has a very high student element.  However, even if the results are not too generalisable, 542 cyclists mainly from the Edinburgh area are going to give a useful local picture of  a significant section of Edinburgh cyclists.

p3 - Helmets.  Noakes (1995) is quoted but is not in the Bibliography.  I found the Noakes abstract, not the full paper.   It seems highly implausible that "up to 85% of cycling fatalities caused by head injuries could be prevented by appropriate wearing of cycle helmets" unless "up to" has a special meaning - quite apart from the question of whether you are more likely to have a crash if helmeted.

p28 fig 3 - Interesting that cycle lanes come out signficantly higher than 'segregated' cycle lanes.   However you had an open-ended question, and classifying what is written and knowing what is intended is very difficult indeed.

p29 - Amazed that over 50% of accidents were with another bike!

p31 fig 7 - Princes St concentration almost certainly highlights the tramline problem.  Also, impressionistically many other roads, such as in the Southside, have much higher bike use than Princes St, which makes this all the more worrying. Furthermore, the sample is heavily student-dominated, whereas it is perhaps more likely that commuters etc are in Princes St, so the Princes St results could understate the true position.    The map ties up with the large number of reported bike/tramline related crashes we know of during the 18months or so that Princes St was open with tramlines installed.

p31 fig 8 - remarkable for ice to come out as top factor.  Survey period (last 3 years) does include 2 very bad winters, with long periods of snow/ice, but still surprising.

p32 Fig 9 - near accidents.  It's very hard to recall near-accidents, so I wouldn't place much reliance on the timeline in this Figure.  Respondents are probably remembering recent incidents rather than what happened months or years ago.  I certainly found that when completing the survey myself - had to think hard, and then only could remember one or two recent incidents.   I would imagine Fig 4 (p29) is more reliable - it certainly was in my case where I easily recalled one specific bad crash resulting in broken kneecap.

p33 Fig 10 - much more reliability can be placed on this than on Fig 9.  Interesting how different it is from Fig 8 (factors in actual crashes).

p33 Table 7 - Interesting that a significant relationship was nearly apparent between wearing a hi-viz jacket and being involved in a near accident - this is noted but not discussed.   The 'common-sense' view would be that hi-viz makes a near miss less likely, whereas this result is in the other direction.  It seems to tie in with the research suggesting that if you 'look' safe/confident then drivers may pass closer.  [the result for helmet-wearing in table 13 is also in that direction, though only marginally so, and certainly not statistically significant].

No such relationship found for reflectors or lights, but they tend to be for nighttime visibility, so are fairly different to Hi-Viz which is strongly visible in the daytime.

p37 Overall visibility score - I didn't find how this was calculated.   However, it might have been worth separating out daytime and nighttime visibility aid types, especially given my point under p33 above.

p39 Table 13 Significance of music in real crashes sounds like an important message.

p43 Hazards - There is a conclusion that helmets should be promoted for icy conditions.  Certainly a fall from the bike, rather than a crash with a car, is where a helmet is most likely to be useful.  However, there is no discussion of whether having a helmet makes one more likely to risk cycling when it is really too icy to do so.  The paper recommends that in icy conditions one should consider not cycling at all, or cycle more carefully.  It then goes on to suggest it might be ok to cycle if you do wear a helmet.   Are these somewhat contradictory recommendations likely to reduce or increase overall injuries?  - that is unknown, but the dilemma is not recognised or discussed, and it is assumed that using a helmet can only have positive consequences.

p43 Factors - male, MTB and (particularly) music issues are worth publicising.

p49 Conclusions - see p33 and p37 above.  It may be that nightime/daytime visibility aid type distinction is of sufficient relevance that it would have appeared in the conclusions had it been studied.