Cycle helmet wearing in 2008

Transcription

1 PPR420 Cycle helmet wearing in 2008 C Sharratt, L Walter and O Anjum

2

3 Transport Research Laboratory PUBLISHED PROJECT REPORT PPR 420 Cycle helmet wearing in 2008 by C Sharratt, LK Walter and O Anjum (TRL) Prepared for: Project Record: PPRO 4/45/4 (Lot 3) Client: Cycle helmet wearing in 2008 Department for Transport, Road Safety Research and Statistics Division (Dr Sofia Marcal-Whittles) Copyright Transport Research Laboratory May 2009 This Report has been prepared for Department for Transport. The views expressed are those of the authors and not necessarily those of Department for Transport. Name Date Approved Project Manager Catherine Sharratt 21/05/2009 Technical Referee Jeremy Broughton 21/05/2009

4 When purchased in hard copy, this publication is printed on paper that is FSC (Forest Stewardship Council) registered and TCF (Totally Chlorine Free) registered.

5 Contents Executive summary 1 Abstract 3 1 Introduction 5 2 Methodology Major built-up roads survey Minor built-up roads survey Statistical significance of results 8 3 Results of the major built-up roads survey Wearing s across time Cyclist characteristics Age and gender Ethnicity Type of bike and cycling conditions Type of bike Time Weather Location Central London Recreational routes Interaction effects Adult cyclists in London during the week Adult cyclists in other cities during the week Child cyclists in other cities during the week Adult cyclists in non-city locations during the week Child cyclists in non-city locations during the week Cyclists on recreational routes at weekends 17 4 Results of the minor built-up road survey Wearing s across time Cyclist characteristics (age, gender, ethnicity) Type of bike, cycling conditions and location Observations of child cyclists Differences between child and adult cyclists Interaction effects 27 5 Conclusions Major built-up roads Minor built-up roads 30 TRL PPR 420

6 Acknowledgements 31 References 31 Appendix A Site locations 32 Appendix B Data collection forms for major built-up roads survey 34 Appendix C Letter of Authority 37 Appendix D Data collection forms for Minor built-up roads survey 39 Appendix E Wearing s by district for Minor built-up roads survey 42 Appendix F Interaction calculations, major built-up roads 43 Appendix G Interaction calculations, minor built-up roads 47 TRL PPR 420

7 Executive summary One of the Government s key objectives is to reduce dependency on the car and achieve a more sustainable and healthy transport system. The Department for Transport encourages safer cycling and under the road safety stgy it recommends that cyclists wear cycle helmets. In 2007, 136 fatal pedal cyclist casualties were recorded by the Police in Great Britain, 2,428 serious casualties and 13,631 slight casualties (DfT Scottish Executive and National Assembly for Wales, 2007). Previous research studies suggest that these statistics are an under-estimation of the actual incidence of cycling casualties and that about 50% of pedal cycle casualties have a head/face injury. Previous surveys carried out by TRL for the DfT have shown that cycling helmet wearing s on major built-up roads have risen significantly from 16% in 1994 to 30.7% in 2006 (DfT, 2008). Wearing s on minor built-up roads have increased from about 8% in 1999 to 13.8% in The observational surveys carried out at sites on major builtup roads were originally selected for their high cycle flows. The survey was first carried out in 1994 and surveys on minor built-up roads were added in 1999 to increase the sample of child cyclists, which were over-represented in this survey, and so successfully balanced the under-representation at major built-up roads. This report presents the results of further nationwide observation surveys of cyclists conducted in the Autumn of 2008 by Accent and analysed by TRL on behalf of the DfT. The aim was to monitor cycle helmet wearing s in order to assess changes over time and to inform policy on the use of cycle helmets. The survey design involved repeat observational surveys of cycle helmet wearing s at 79 fixed locations on built-up major roads around Great Britain and repeat observational surveys at 20 built-up minor roads. As this is a longitudinal survey, the methodology used in previous year s surveys was replicated and the survey documents were adopted from those used in the previous surveys. The maps provided by DfT ensured the same major route locations and minor route circuits were surveyed. It was proposed that six-hour interviewer shifts are conducted over the same sites/geographical locations as used in the previous surveys. The 2008 survey on major built-up roads showed that cycle helmet wearing was 34.3%, an increase from 30.7% in The wearing has increased each year the survey has been carried out since 1994 when it was 16.0%. The increase since 2006 is due to the increase in adult cyclists wearing helmets from 31.5% to 35.3% as the cycle helmet wearing for children has remained constant at 17.6%. Female cyclists overall continue to be observed wearing helmets more than male cyclists, as in all previous surveys. The for both male and female cyclists has increased since 2006 and over time since the first survey in The for male cyclists has increased from 15.5% to 34.1% and for female cyclists from 17.2% to 34.9%. Whilst the cycle helmet wearing for female child cyclists has increased between 1994 and 2008 from 21.9% to 32.9%, the for male child cyclists has declined over time from 16.0% to 13.1%. The for male adult cyclists was similar to female adult cyclists at 35.4% compared to 35.0%, increasing from 15.5% and 17.0% in 1994 respectively. As in previous surveys, the cycle helmet wearing s at the London sites were significantly higher at 69.5% than sites outside of London at 29.9%. On recreational routes there continued to be more cyclists wearing helmets than at other sites as well as there being more children observed on these routes. An overall model of all the variables together showed that for adult cyclists in London the strongest effect on wearing was time of day, with them more likely to wear a helmet TRL 1 PPR420

8 at peak than at off-peak times. Adult cyclists in other cities on weekdays were more likely to wear a helmet at peak times in wet weather. White adult cyclists in London, other cities, and non-cities riding tradition town bikes were more likely to wear a helmet. Child cyclists in cities excluding London were more likely to wear a helmet if female and riding a racing bike during wet weather. In non-city locations girls were also more likely to wear a helmet than boys. The overall cycle helmet wearing on minor built-up roads increased to 16.7% in 2008 from 13.8% in The has increased each year the survey has been carried out since When analysed by age and gender, the for boys aged has decreased since 1999 from 8.6% to 7.0%. The wearing on minor built-up roads was significantly higher for children than adults in 1999 but this was reversed in all subsequent surveys. In 2008 the wearing for adult cyclists was 18.2% compared to 12.0% for child cyclists. Whereas adults were more likely to cycle on the pavement during peak hours, children were more likely to cycle on the pavement at all times of the day. Adults and children were more likely to wear a helmet if cycling on the road (19.4% and 13.9% respectively) rather than the pavement (13.7% and 10.5% respectively) and both most likely when on a cycle path (24.8% and 14.5% respectively) though the difference was only significant for adults. Female children were most likely to wear a cycle helmet, followed by female adults, male adults, and male children. The helmet wearing was also higher among cyclists using additional safety aids, riding a racing bike, of white ethnic origin, or if riding on the road compared to the pavement. In summary, once again the overall cycle helmet wearing has increased. However, on major built-up roads the increase was accounted for by the increase in adult (both female and male) wearing helmets as there was no change to the child helmet wearing overall and the for male child cyclists declined. On minor built-up roads the wearing for boys aged decreased to 7% since the first survey in 1999 and the for boys aged 7-10 is still low at 11.6%. As in previously surveys, the wearing s at London sites were significantly greater than sites outside of London and were particularly higher during peak times in London and in other cities. Both adult and child cyclists were more likely to wear a helmet when cycling on the road than the pavement. TRL 2 PPR420

9 Abstract An observational survey of pedal cycle helmet wearing s was undertaken in 2008 on behalf of the Department for Transport. TRL was commissioned to conduct the seventh survey with Accent collecting the survey data. The surveys enable DfT to assess changes over time in wearing patterns and to inform policy on the use of cycle helmets. The survey involved observations at 79 fixed locations on major built-up roads and at 20 locations on minor built-up roads around Great Britain. The 2008 survey on major built-up roads showed that cycle helmet wearing was 34.3%, an increase from 30.7% in The wearing has increased each year the survey has been carried out since 1994 when it was 16.0%. The increase since 2006 is due to the increase in adult cyclists wearing helmets from 31.5% to 35.3% as the cycle helmet wearing for children remained constant at 17.6% (the same as in 1994 and 2006, having dropped in between). Other findings are reported by age and gender, ethnic origin, time of day, weather, type of bike, riding position, location, and lighting conditions. The overall cycle helmet wearing on minor built-up roads increased to 16.7% in 2008 from 13.8% in The has increased each year the survey has been carried out since When analysed by age and gender, the for boys aged has decreased since 1999 from 8.6% to 7.0%. Other findings are reported by ethnic origin, time of day, weather, type of bike, riding position, location, lighting conditions, whether riding with a passenger, whether in a school uniform, on a paper round, and the presence of additional safety aids. TRL 3 PPR420

10 TRL 4 PPR420

11 1 Introduction The Department for Transport commissioned Accent and TRL to undertake a survey of pedal cycle helmet wearing s in 2008 that built on the six previous surveys undertaken in 1994, 1996, 1999, 2002, 2004 and Accent carried out the data collection whilst TRL analysed the survey data. The survey was first undertaken by TRL in the autumn of 1994 (Taylor and Halliday, 1996) to observe the cycle helmet wearing in Great Britain on major built-up roads. This was repeated in 1996 (Bryan-Brown and Taylor, 1997), 1999 (Bryan-Brown, 2000), 2002 (Gregory, Inwood and Sexton, 2003), 2004 (Inwood, Whitley and Sexton, 2005), and 2006 (McGarry and Sheldon, 2007). In 1994, twenty-seven thousand cyclists were observed at 79 sites throughout Great Britain. 16% of cyclists were wearing cycle helmets. The overall wearing has increased each time that the survey has been repeated as shown in Table 1-1 Table 1-1 Overall wearing across time on major-built up roads % % % % % % % Overall wearing 16.0% 17.6% 21.8% 25.1% 28.2% 30.7% 34.3% When TRL was asked to repeat the survey for a third time in 1999, it was decided to make the sample more representative of the national cycling population by including a survey of minor built-up roads. The proportion of adult and child cyclists wearing a cycle helmet was found to be less on minor than on major built-up roads, thus supporting the inclusion of surveys on minor built-up roads in future surveys. Surveys on major built-up roads also tend to under-represent child cyclists in the sample. Child cyclists tend to use minor rather than major built-up roads, so coverage of children is improved by conducting a parallel survey on minor roads. The two surveys were not intended to produce a nationally representative wearing, but rather to measure trends over time. The objectives of the 2008 surveys were: To ascertain cycle helmet wearing s on major and minor built-up roads in 2008, building on previous surveys; To identify and deliver any improvements to previous surveys, in terms of methodology and/or implementation; To report the findings of the major and minor built-up roads in a format that is consistent with previous reports, including analysis of differences by age, gender and other factors; and To report changes in wearing s over time by comparing results with previous surveys. TRL 5 PPR420

12 2 Methodology 2.1 Major built-up roads survey The survey design was kept as similar to the design of the previous surveys as possible, using the sites in England, Scotland and Wales as shown in Figure 2.1 and Appendix A. It was a fixed site survey where all sites were surveyed on weekdays, unless identified as being on a recreational route, which were then observed on Sundays. Figure 2.1. Major Site Locations TRL 6 PPR420

13 Accent conducted the surveys at the same times and on the same day of the same week of the year as closely as possible to previous surveys. The survey teams were required to work at a specific site for a six-hour shift, which was either 07:000 to 13:00 or 13:00 to 19:00, with observations being made continuously. Interviewers were provided with a map showing where they needed to work. They used the observation data collection forms as used in previous surveys (see Appendix B). Interviewers collected the following details for each passing cyclist: Helmet (on head, carrying but not wearing, no) Gender of cyclist Age (child, [under 16] or adult, [16 or over]) Ethnic origin (black, white, Asian - Indian subcontinent [e.g. Pakistan, Bangladesh], Asian - South East Asian [e.g. China, Thailand, Malaysia], or other 1 ) Type of bike ridden (racing, mountain/bmx, traditional town, other 2 ) Passenger (yes- helmet worn, yes- no helmet worn, no) If more than one cyclist riding together, number in the group All interviewers were provided with a Letter of Authority from the DfT verifying that they were undertaking a legitimate survey (see Appendix C). 2.2 Minor built-up roads survey The survey on minor built-up roads was first carried out in 1999 and repeated in all subsequent surveys. The main aim was to boost the sample of child cyclists and to be more representative of the types of cycling that adults and children do on minor built-up roads. The main considerations were to observe a large sample of children in an economical way, ensuring a geographical spread, across all socio-economic groups. Whilst the surveys on major built-up roads were undertaken at fixed sites, a moving observation technique on specified routes was employed for the minor built-up roads survey. For the 2004 survey, the same compact areas (1km squares) were used as in the previous surveys within the following twenty locations: 1. Coventry 2. Crewe & Nantwich 3. Doncaster 4. Eastleigh 5. Fareham 6. Gloucester 7. Ipswich 8. Kingston 9. Kingston-upon-Hull 10. Lincoln 11. Liverpool 12. Milton Keynes 13. Peterborough 14. Portsmouth 15. Slough 16. Taunton 17. Wigan 1 Other includes any ethnic group which is not black, white, Indian subcontinent, or south east Asian. 2 Other includes any type of bike which is not a racer, mountain/bmx or traditional town bike. TRL 7 PPR420

14 18. Wirral 19. Wolverhampton 20. Worthing Previous studies had shown that the socio-economic status of an area affects helmet wearing s (Christie, 2000). The DfT s Index of Local Deprivation (Government Statistical Service, 1999) was used as a measure of the socio-economic status of an area. Child casualty s derived from Stats 19 accident data were used as a proxy for cycling levels. Districts were ranked by child casualty levels and by Deprivation Index and sampled from those with high cycling activity (i.e. many casualties), ensuring a good mix of socio-economic status and geographical areas. Within each area a route was mapped out, covering as many roads within the square as possible. Each of the selected routes was surveyed on one weekday and one weekend day by driving around the prescribed route for a six hour period (either from 07:00 to 13:00 or 13:00 to 19:00), recording the following details for each cyclist observed: Helmet (on head, carrying but not wearing, no) Gender of cyclist Age group (infant 0-6, junior 7-10, secondary 11-16, adult) Ethnic origin (black, white, Indian subcontinent [e.g. Pakistan, Bangladesh], South East Asian [e.g. China, Thailand, Malaysia]) Riding position (road, pavement, cycle path, other) Type of bike (racing, mountain/bmx, traditional town, other) School uniform (yes or no) Passenger (yes - helmet worn, yes - no helmet worn, no) Paper round (yes, no) Additional safety aids e.g. front/rear lights; fluorescent clothing; stabilisers; reflectors (yes, no) The data collection forms used by Accent can be seen in Appendix D. 2.3 Statistical significance of results Throughout the following sections, there are tables showing differences between groups of cyclists e.g. children versus adults. Significant differences between groups are indicated as shown in Table 2-1: Table 2-1 Levels of significance NS Not significant at the 5% level p<0.05 Significance of at least 5% p<0.01 Significance of at least 1% p<0.001 Significance of at least 0.1% In the main body of the text, significance levels are quoted as a p level, e.g. a p value of 0.05 or less means that the result is significant to at least the 5% level. Data are missing in some cases, so sample s are given where appropriate. TRL 8 PPR420

15 3 Results of the major built-up roads survey 3.1 Wearing s across time In the 2008 survey 20,034 cyclists were observed at the sites previously surveyed 3. Of these cyclists 34.3% were wearing helmets. The wearing has increased from the wearing observed in the 2006 survey when 30.7% of observed cyclists wore a helmet (statistically significant 4 ). As shown in Table 3.1, however, the rise is due to an increased proportion of adult cyclists wearing helmets. The wearing for child cyclists in this sample has remained the same from 2006 to Figure 3.1 shows the changes in cycle helmet wearing s between 1994 and 2008 by age and gender. Analysis of the survey data for adult males and adult females sepaly shows that the helmet-wearing of adult males increased steadily between 1994 and 2008 from 15.6% to 35.4%. Similarly, the wearing of adult females has increased steadily over the same period from 17.1% to 35.0%. Although the wearing for female children cyclists fell between 1994 and 1996, it has increased with each subsequent survey. The wearing for female children cyclists in 2008 is 32.9% compared to 21.9% in The wearing for boy cyclists had decreased with every survey between 1994 and 2004 but increased from 10.8% in 2004 to 13.6% in The wearing for boy cyclists in 2008 is 13.1%, slightly lower than in 2006 and 2.9% lower than the first survey in Figure 3.1. Changes in cycle helmet wearing s between 1994 and 2008 by age and gender 3 A total of 20,063 cyclists were observed, but 29 were excluded from all analyses as it was not recorded whether or not they were wearing a helmet. 4 Chi-square (df = 1) = , p<0.001 TRL 9 PPR420

16 Table 3.1. Cycle helmet wearing s All cyclists 1994 % % % % % % % 2008 Significance of 2006/2008 difference (χ2) 16.0% 27, % 27, % 26, % 26, % 22, % 20, % 20,034 p<0.001 Males Females 15.5% 19, % 19, % 18, % 18, % 16, % 13, % 14,413 p< % 7, % 7, % 7, % 5, % 6, % 6, % 5,621 p<0.001 Children Adults 17.6% 1, % 1, % 1, % 1, % 1, % 1, % 1,075 NS 15.9% 25, % 24, % 24, % 24, % 21, % 19, % 18,917 p<0.001 TRL 10 PPR420

17 3.2 Cyclist characteristics Age and gender The 2008 adult cycle helmet wearing of 35.3% was significantly greater than the 17.6% observed amongst children 5. The proportion of adult cyclists wearing helmets has significantly increased from 31.5% in 2006 to 35.3% 6. The proportion of children wearing a helmet whilst cycling has remained the same at 17.6% from 2006 to There was no difference between the wearing s of male and female cyclists overall or between adult cyclists but there was a significant difference between male (13.1%) and female (32.9%) child cyclists (see Table 3.2) 7. As in 2006, the wearing for male adults (35.4%) was higher than for female adults (35.0%), but the difference was not significant. The wearing of males increased from 31.8% in 2006 to 35.4% in 2008, and females increased from 30.9% to 35.0%; both increases were statistically significant 8. Table 3.2. Wearing s by age group and gender Children Wearing Adult Wearing 2008 Survey Males 13.1% % 13,545 Females 32.9% % 5, Survey Males 13.6% % 12,913 Females 29.2% % 6, Survey Males 10.8% 1, % 14,929 Females 26.3% % 6, Survey Males 12.3% 1, % 17,738 Females 24.4% % 6, Survey Males 12.7% 1, % 17,794 Females 20.9% % 6, Survey Males 13.3% 1, % 17,545 Females 17.6% % 7, Survey Males 16.0% 1, % 18,624 Females 21.9% % 7,368 5 Chi-square (df=1) = , p< Chi-square (df = 1) = , p< Chi-square (df = 1) = , p< For male data: Chi-square (df=1) = , p< For female data: Chi-square (df=1) = , p< TRL 11 PPR420

18 3.2.2 Ethnicity Since 2002, the ethnicity of the cyclist was recorded to test whether ethnicity might influence helmet wearing and significant differences were found overall 9. White cyclists were more likely to be wearing a cycle helmet (35.4%) and cyclists of other origin (16.9%) and Indian subcontinent (20.2%) were least likely to be wearing a helmet. Table 3.3 compares the wearing of cyclists by ethnic group. Table 3.3. Wearing by ethnicity Adults Children Total Wearing Wearing Wearing Black 28.4% % % 455 White 36.4% 17, % % 18,209 Indian 21.0% % % 588 subcontinent South East Asian 22.6% % % 640 Other 17.6% % % 89 Total 18,863 1,316 19,981 When child and adults were analysed sepaly, wearing s varied with ethnicity but this was statistically significant for adult cyclists only 10. White adult cyclists were more likely to be wearing helmets (36.4%). The sample s for child cyclists are low so it is difficult to draw conclusions. It is worth noting that with the exception of other and South East Asian, the wearing for all ethnic groups has increased since 2006 (see table 3.4). Table 3.4. Changes in wearing by ethnicity between 2006 and Change Wearing Wearing in wearing Black 27.6% % % White 31.5% 18, % 18, % Indian Subcontinent 19.5% % % South East Asian 25.2% % % Other 37.1% % % Total 20,658 19, Type of bike and cycling conditions Type of bike The type of bike being ridden by the cyclist was classified as racing bike, mountain bike/bmx, traditional town bike, or other. As in previous surveys, in 2008 the most common types of bikes were the mountain bike/bmx (48.5% of the bikes observed) and then traditional town bike (31.4%). Table 3.5 shows that helmet wearing varied significantly according to the type of bike ridden Chi-square (df = 4) = , p< Chi-square (df = 4) = , p< Chi-square (df = 3) = , p<0.001 TRL 12 PPR420

19 Of those cyclists on racing bikes, 56.0% wore a helmet compared with 41.2% on other bikes, 32.6% on mountain bikes/bmxs and 24.3% on traditional town bikes. Table 3.5. Wearing for child and adult cyclists by bike type Adults (N= 18,857) Children (N= 1,074) Total (N= 19,975) % Riding bike Wearing % Riding bike Wearing % Riding Bike Wearing Mountain bike 47.1% 34.2% 71.3% 13.8% 48.5% 32.6% Traditional Town 32.2% 24.5% 17.4% 19.8% 31.4% 24.3% Racing 17.7% 56.1% 7.1% 47.4% 17.1% 56.0% Other 2.9% 42.9% 4.2% 22.2% 3.0% 41.2% Total 100% 35.2% 100% 17.6% 100% 34.3% When adult and child cyclists were analysed sepaly, helmet wearing varied significantly according to the type of bike ridden 12. Over half (56.1%) of adult cyclists riding racing bikes were wearing cycle helmets, whilst under half (42.9%) of those riding on other bikes wore a helmet, just over a third (34.2%) of those riding a mountain bike/bmx and under a quarter (24.5%) on traditional town bikes. For child cyclists, the numbers observed were small but wearing also varied according to the type of bike ridden. Just under a half (47.4%) of children riding a racing bike wore a helmet, with under a quarter (22.2%) of those riding an other type of bike wearing one, just under a fifth (19.8%) of those riding a traditional town bike did, but only 13.8% of those on mountain bike/bmx wore a helmet Time As in previous years, the survey teams monitored cyclists continuously, grouping observations into quarter hour periods. Morning surveys began at 0700hrs and continued until 1259hrs. Afternoon surveys commenced at 1300hrs and finished at 1859hrs. Figure 3.2 shows how observations were distributed throughout the day during weekdays only. Over half (59.4%) of the 19,811 cyclists observed on weekdays were observed during peak time and 40.6% during off-peak times For adult cyclists, chi-square (df=3) = , p<0.001 and for child cyclists, chisquare (df=3) = , p< Peak hours were defined as hours and hours. TRL 13 PPR420

20 1,200 Number of cyclists observed 1, :30 18:00 17:30 17:00 16:30 16:00 15:30 15:00 14:30 14:00 13:30 13:00 12:30 12:00 11:30 11:00 10:30 10:00 9:30 9:00 8:30 8:00 7:30 7:00 Figure 3.2. Number of cyclists by time of day during the week in 2008 Helmets were worn by 24.6% of those observed during weekday off-peak times and 40.7% of those observed during peak times, and this difference is statistically significant 14. This difference was also significant for children and adults 15. Where the wearing for children was 22.3% (n = 641) in peak times compared to 9.4% (n = 427) in off peak times, for adults the wearing in peak times was 41.7% (n = 11,137) compared to 25.5% (n = 7,602) in off-peak times. Table 3.6 shows how the peak and off-peak wearing s have varied over the years. The wearing has increased both in peak and off-peak periods since The wearing in peak periods remains higher than in off-peak periods. Cyclists may believe they are more at risk in peak periods due to the greater volume of traffic and therefore be more likely to wear a helmet or it could be related to the type of cyclist who is cycling at those times of day. Only two sites were observed at the weekend and these sites were on recreational routes. A total of 188 cyclists were observed on these sites. The wearing was 52.1% for cyclists on these routes which were excluded from the main peak/off-peak analysis during the weekday survey. Table 3.6. Wearing by year (all cyclists in weekday survey) Year Peak Off Peak % 11.3% % 13.2% % 15.0% % 16.8% % 18.0% % 24.0% % 24.6% 14 Chi-square (df = 1) = , p< For children: Chi-square (df = 1) = , p<0.01. For adults: Chi-square (df = 1) = , p<0.01. TRL 14 PPR420

21 3.3.3 Weather The weather (wet, dry, and mixed) was recorded for 100% of the cyclists observed in A wearing of 35.3% was found in rain, 33.0% in dry weather and 51.8% when the weather was mixed. These differences were statistically significant 16. In previous surveys, more cyclists were observed wearing helmets in wet weather than in dry weather, significantly in 1996 and Location Central London The wearing at the three central London sites was 69.5% in the 2008 survey, significantly higher than at sites outside central London (29.9%) 17. The cycle helmet wearing in central London was therefore more than double that for site outside of London, replicating the findings from previous surveys. The 2008 wearing in London increased from 60.6% in 2006 to 69.5% in 2008 and outside of central London from27.2% in 2006 to 29.9% in In 2008, the cycle helmet wearing for adults in Central London was 69.5%, and 30.7% outside Central London. This represents an increase from 2006 from the corresponding wearing s of 60.9% and 28.0%. For child cyclists, the sample in Central London is small (1,075). However, there was a slight decrease in the wearing s for children in sites outside of London to 17.2% from 17.4% in 2006, though an increase from the 2004 from 13.7%. As in previous surveys, the vast majority of cyclists observed in Central London in 2008 were adults (99.4% of cyclists observed in Central London), with only a small minority of cyclists observed in Central London being children (0.3%). This compares to 93.8% adults and 6% children at sites outside of Central London. In addition, 74.8% of the sample in Central London was male compared with 71.6% elsewhere Recreational routes A total of cyclists were observed on a Sunday on the Ridgeway cycle track in Oxfordshire and on the Bath to Bristol cycle route. Significantly more cyclists wore helmets on these recreational routes (52.1%) than at other sites (34.2%) 19. Table 3.7 shows a comparison of the wearing by site in 2006 and It was found that there was a significant difference between the wearing for children observed on recreational routes and at non recreational routes in Great Britain and also for adults 20. There was a significant difference between the wearing for males observed on recreational routes and at non recreational routes in Great Britain and also for females 21. No weekday recreational route data are available, so this effect cannot be distinguished from any weekday effect. On recreational routes, 28.7% were riding in a 16 Chi-square (df = 2) = , p< Chi-square (df = 1) = , p< Although the sample is relatively small, significance testing takes account of some of the variability due to sample. A conclusion of significant difference would therefore be the same if observations were repeated even if the numbers were not. 19 Chi-square (df=1) = , p< For children Chi-square (df=1) = , p< For adults Chi-square (df = 1) = , p< For males Chi-square (df = 1) = , p< For females Chi-square (df = 1) = 8.693, p< TRL 15 PPR420

22 group compared to 4.4% of those not on recreational routes. This difference was significant 22. Table 3.7. Wearing by site in 2006 and 2008 Recreational routes Other sites Recreational routes Other sites % n % n % n % n Overall 39.4% % 20, % % 19,849 Children 54.2% % 1, % % 1,068 Adults 36.0% % 19, % % 18,739 Females 41.8% % 6, % % 5,574 Males 38.1% % 13, % % 14, Interaction effects Many of the variables discussed in Sections 3.2 to 3.4 are interrelated. For example, the two variables type of route (recreational or city) and time of week (peak/off-peak) are likely to be related: cyclists in cities may be more likely to be commuting and therefore to be travelling in peak periods and conversely, cyclists on recreational routes could be more likely to be cycling in off-peak periods. In this section we consider all the variables in one model in order to take these interrelations into account. In Sections 3.2 to 3.4 differences were observed in location, age and time of week. The data were divided into subgroups to represent these differences and each subgroup was analysed sepaly. 1 London, adults (there were only 7 child observations) and weekdays (n=2,251). 2 Other city sites, adults and weekdays (n=5,689). 3 Other city sites, children and weekdays (n=175). 4 Non-city sites, adults and weekdays (n=10,684). 5 Non-city sites, children and weekdays (n=884). 6 Weekend recreational use (there were only 7 child observations and 178 adult observations). A logistic regression model (Pallant, 2005) was used to identify which variables and interactions affect cycle helmet wearing. The factors that were fitted to the model with interaction terms included ethnicity, type of bike, weather, sex, and time of day. The detailed results are shown in Appendix F Adult cyclists in London during the week For the subgroup of adult cyclists in London during the week, the model found that time of day, ethnicity, and type of bike had effects that were statistically significant, and that the type of bike and peak interaction was significant. That is, differences were observed in cycle helmet wearing between these subgroups. Taking into account all variables, the following conclusions can be drawn from the model: Cyclists riding Mountain/BMX or Traditional Town bikes are significantly less likely to be wearing a helmet than cyclists on Racer bikes; 22 Chi-square (df = 2) = , p<0.001 TRL 16 PPR420

23 Cyclists cycling in off-peak hours are significantly less likely to be wearing a helmet than cyclists travelling in peak hours ; White cyclists in this subgroup are significantly more likely to be wearing a helmet than Black cyclists ; Cyclists of Asian Indian subcontinent ethnic origin are significantly less likely to be wearing a cycle helmet than Black cyclists; Cyclists on Mountain/BMX bikes cycling in peak hours are significantly more likely to be wearing a helmet than other subgroups Adult cyclists in other cities during the week The next model suggests that the variables time of day (peak/off-peak), weather, and type of bike account for differences in helmet wearing s in the subgroup of adult cyclists in cities other than London during the week. Helmet wearing for cyclists cycling during off-peak hours is significantly higher than those cycling in peak hours; Cyclists cycling in dry and wet conditions are significantly less likely to wear a helmet than those cycling in mixed weather conditions; Mountain/BMX, Traditional Town and Other bike users are significantly less likely to wear a helmet than those on Racer bikes; Cyclists cycling in dry or wet weather during peak hours are less likely to be wearing helmet than those cycling in mixed weather (during peak hours) Child cyclists in other cities during the week The third subgroup has few cyclists and therefore the model cannot be as complex. Three variables (and no interactions) were found to be important when accounting for differences in helmet wearing habits: sex, time of day and type of bike. Girls are significantly more likely than boys to wear cycle helmets; Child cyclists cycling in off-peak periods are less likely to wear a helmet than those in peak hours; Most children ride mountain bikes; however helmet wearing of children on all other types of bike is higher Adult cyclists in non-city locations during the week For adult cyclists in non-city locations, the model found that sex, time of day, ethnicity and type of bike were all important in distinguishing different groups of people who have different helmet wearing s. There was also an interaction of sex with bike type. Cyclists on Mountain/BMX bikes, Traditional Town bikes and others are less likely to wear a helmet than those on Racer bikes; Off-peak cyclists were less likely to be wearing a helmet than those cycling in peak hours; White and Asian-Indian subcontinent cyclists were significantly more likely than black cyclists to be wearing a helmet Child cyclists in non-city locations during the week For children cycling in non-city locations, the important discriminating factors were sex and the combined factor sex and time of day. Girls were three times more likely than boys to wear a helmet in non-city locations Cyclists on recreational routes at weekends There were insufficient data to determine a satisfactory model for recreational routes. TRL 17 PPR420

24 4 Results of the minor built-up road survey 4.1 Wearing s across time Table 4.1 shows that 4,699 cyclists were observed in the 2008 survey 23, 16.7% of whom were wearing a cycle helmet. The increase from the wearing (13.8%) in 2006 was statistically significant 24. Figure 4.1 shows the changes in cycle helmet wearing s over time by age and gender. Figure 4.1. Changes in cycle helmet wearing between 1999 and 2008 by age and gender In addition to the 134 children who were wearing a helmet, eight were observed carrying a helmet on their bike but not on their head. Seven of the eight were riding on the pavement or grass verge rather than the road and three were wearing school uniform. 23 A total of 4,741 cyclists were observed, but 42 were excluded from all analyses as it was not recorded whether or not they were wearing a helmet. 24 Chi-square (df = 1) = , p<0.001 TRL 18 PPR420

25 Table 4.1. Cycle helmet wearing s 1999 % % % % % 2008 Significance of 2006/2008 difference (χ 2 ) All cyclists 8.2% 3, % 4, % 4, % 4, % 4,699 p<0.001 Males Females 8.2% 2, % 3, % 3, % 3, % 3,612 P< % % 1, % % 1, % 1,057 NS Children Adults 9.7% % 1, % 1, % 1, % 1,116 P< % 2, % 3, % 2, % 3, % 3,540 P<0.01 TRL 19 PPR420

26 4.2 Cyclist characteristics (age, gender, ethnicity) The wearing for male cyclists in 2008 of 15.3% was greater than the wearing in 1999 of 8.2%, and the for female cyclists has increased from 8.1% in 1999 to 20.9% in Table 4.2 shows that the cycle helmet wearing for child males was 9.0% and for child females was 22.0%, increases from 6.4% and 18.4% respectively from 2006, however this was only significant for male children 25. Whilst the adult male wearing has significantly increased from 13.4% in 2006 to 17.4% in , the wearing for adult females has decreased slightly from 21.8% to 20.7% (not significant). Table 4.2. Wearing by age and gender in 2006 and Adult wearing Child wearing Adult wearing Child wearing % n % n % n % n Male 13.4% 2, % % 2, % 847 Female 21.8% % % % 259 Either 15.6% 3, % 1, % 3, % 1,106 Table 4.3 shows that, once again cyclists of white ethnic origin continue to be significantly more likely to wear a helmet than other ethnic groups 27. That said, the cycle helmet wearing has continued to increase from 2006 for all ethnic groups. The difference was statistically significant only for adult cyclists 28 and not for child cyclists. The lack of statistical significance for child cyclists may be due to a lack of statistical power because of the smaller sample. Table 4.3. Wearing by ethnic origin Category Wearing Rate (n= 4,452) Wearing (n= 4,681) Black 8.8% % 333 White 14.7% 3, % 3,994 Indian-Asian 3.3% % 298 South East Asian 2.3% % Type of bike, cycling conditions and location Table 4.4 shows that those riding on racing bikes were significantly more likely to wear a helmet than those riding on other bikes 29, reflecting the results from previously minor built-up roads surveys and the major built-up roads survey. The wearing for cyclists on each type of bike has increased since Chi-square (df = 1) = 3.965, p< Chi-square (df = 1) = , p< Chi-square (df = 4) = , p< Chi-square (df = 4) = , p< Chi-square (df = 3) = , p<0.001 TRL 20 PPR420

27 Table 4.4. Wearing by type of bike Category Wearing Rate (n=4,457 ) Wearing (n= 4,696) Racing bike 27.2% % 339 Mountain bike/ 12.9% 2, % 3,106 BMX Traditional 11.8% 1, % 1,136 town bike Other 14.9% % 115 The wearing was found to be greatest during the weekday peak hours 30, as in previous years. Table 4.5 shows the wearing at different times of the week. Table 4.5. Wearing by time and day Category Wearing Rate (n=4,473) Wearing Rate (n= 4,699) Weekday peak 17.2% 1, % 1,710 Weekday off-peak 13.9% 1, % 1,039 Weekend 10.7% 1, % 1,950 As in the all previous surveys, cyclists were found to be more likely to wear a helmet if they were riding on the road (18.5%) rather than the pavement (12.4%). However, the wearing was highest amongst those cyclists riding on a cycle path of a recreational route (22.4%) 31. In 2008 a significantly higher proportion of cyclists wore a helmet in dry weather (17.3%) than in the wet (11.5%) 32. Although the wearing observed in 2008 varied with the level of light, the difference was not significant. Table 4.6 shows the wearing by light condition. Table 4.6. Wearing by light condition Category Wearing Wearing Rate Getting light 15.4% % 280 Light 9.8% 2, % 3,885 Getting dark 9.3% % 458 Dark 7.6% % 76 Overall 10.0% 3, % 4,699 The presence of additional safety aids such as fluorescent jackets, front/rear lights and reflectors was significantly correlated with the wearing in 2008, the being 14.3% on cycles without additional safety aids and 29.2% with safety aids 33. This 30 Weekday peak hours are defined as 0700 to 0959 and 1600 to 1859 hours Monday to Friday. 31 Chi-square (df=3) = , p< Chi-square (df = 2) = 9.672, p< Chi-square (df=1) = , p< of those with additional safety aids = 758, sample of those without additional safety aids = 3,941. TRL 21 PPR420

28 replicates the findings from the previous surveys and also demonsts an increase over time in the wearing s for both cyclists using additional safety aids and those without them. Fifty-five cyclists with a passenger were observed; in a fifth of cases (21.8%) both cyclist and passenger wore a helmet, in 12.7% only the passenger did, in 3.6% only the cyclist did. In the remaining 61.8% neither cyclist nor passenger wore a helmet. Thirtythree of the 55 cyclists with a passenger were adults, 15 were aged years, three were between seven and ten, and the remaining two were less than 7 years old. Two observations were made of a baby in a cycle buggy or carrier on the bike. The table in Appendix E shows how wearing s varied by district for adults and children. There was no clear relationship between wearing and the level of deprivation. It may be due to cyclists passing through an area and not necessarily living in that area. 4.4 Observations of child cyclists The purpose of conducting the survey on minor built-up roads, as in previous surveys, was to increase the sample of child cyclists and to be more representative of the type of cycling that children do. It was found in the 2008 survey, as in previous surveys, that the number of child cyclists peaked in the morning and afternoon, associated with travel to and from school (see Figure 4.1). This year, for the first time, the wearing for children was greater during the weekday off-peak (13.5%) rather than the weekday peak (11.9%) or the weekend (11.6%), but the difference was not significant Number of cyclists :30 18:00 17:30 17:00 16:30 16:00 15:30 15:00 14:30 14:00 13:30 13:00 12:30 12:00 11:30 11:00 10:30 10:00 9:30 9:00 8:30 8:00 7:30 7:00 Figure 4.1: Child cyclists observed during 2008 survey by time of day TRL 22 PPR420

29 The cycle helmet wearing was slightly (though not significantly) higher amongst children in identifiable school uniform than those not in uniform in 2008 (though this difference was significant in all of the previous surveys). Table 4.7 summarises the sample s and wearing s split by school uniform. Although the wearing for the 31 children on a paper round was higher at 16.1% than for those not on a paper round (11.9%), the difference was not statistically significant. This was also the case in previous surveys. There was a significant difference between the wearing s of children riding different types of bike 34. Of those riding racing bike, 30.4% were wearing a helmet. For mountain bikes/bmxs, the figure was 11.8%, with more than three in four children riding this type of bike. Of those riding traditional town bikes, 9.7% wore a helmet, and 15.3% of those riding other types of bikes. The wearing for children on each type of bike has increased over time. The wearing amongst children in wet weather was 8.4% compared to 12.7% when dry and 4.9% when mixed, the opposite of the trend for cyclists on major built-up roads. The difference was not statistically significant. 34 Chi-square (df = 3) = 8.902, p<0.05 TRL 23 PPR420

30 Table 4.7: Wearing s amongst children by school uniform Children wearing uniform Wearing % Wearing % Wearing % Wearing % Wearing % Yes 18.2% % % % % 199 No 8.9% % 1, % 1, % 1, % 917 Either 9.8% % 1, % 1, % 1, % 1,116 Table 4.8: Wearing s amongst children by age group Wearing % Wearing % Wearing % Wearing % Wearing % 0-6 Infants 7-10 Junior Secondary , Adult 7.4 2, , , , ,540 Base 3,306 4,804 4,003 4,408 4,656 TRL 24 PPR420

31 4.5 Differences between child and adult cyclists The combined adult and child cycle helmet wearing observed in the current survey was significantly higher than that observed in The for children aged 7-10 and that for adults were also significantly higher than in The cycle helmet wearing on minor built-up roads was significantly 36 higher for children than for adults in 1999, but this was reversed in all subsequent surveys 37 including the current survey. The wearing has increased from 1999 to 2008 amongst children adults (table 4.9). Table 4.8 gives wearing s by age group in all surveys. Table 4.9 shows that male children aged 7-10 and continue to be less likely to wear a cycle helmet than all other groups. Although the wearing for boys aged 7-16 decreased in each survey between 1999 and 2004, since 2004 the wearing increased in 2006 and again in the 2008 survey. Table 4.9: Comparison of wearing s by age group and gender Male Female Year Age Wearing Wearing group % 29 25% % % % % 126 Adult 7.7% 1, % % 76 6% % % % % 180 Adult 11.1% 2, % % % % % % % 187 Adult 10.9% 2, % % % % % % % 193 Adult 13.4% 2, % % % % % % % 156 Adult 17.4% 2, % 793 Child and adult cyclists were more likely to wear a helmet when riding on the road than on the pavement (the differences in wearing with cycling position were significant for adults 38 but not for children). The wearing of children cycling on the road was 13.9% compared with 10.5% of those riding on the pavement. For adults the wearing 35 For children aged 7-10: Chi-square (df = 1) = , p< For adults: Chisquare (df = 1) = 9.235, p< Chi-square (df=1) = 4.87, p< : Chi-square (df=1) = 23.31, p< : Chi-square (df=1) = , p< : Chi-square (df = 1) = , p< : Chi-square (df = 1) = , p< Chi-square (df = 3) = , p<0.001 TRL 25 PPR420

32 was 19.4% on the road and 13.7% on the pavement. Both adults and children were more likely to wear a helmet when cycling on a cycle path than on the road or pavement. Table 4.10 shows that (as in previous surveys) adults were less likely than children to cycle on the pavement than the road and that for adults a greater proportion cycle on the road during peak periods than off-peak periods. This is compared to children who were more likely to cycle on the pavement than on the road during peak and off-peak times. Table Cycling in road or on pavement, by time of day Children Adults Road Pavement Other Road Pavement Other Peak 35.3% 53.7% 10.9% 64.4% 26.2% 9.4% Off-peak 36.2% 58.3% 5.5% 59.0% 31.1% 9.9% Only 20 cyclists were observed with a passenger 39, 16 of whom were adult cyclists and four were children (see Table 4.10). Both the cyclist and passenger were wearing a helmet in three of the four children cyclists who had a passenger, compared to nine of the 16 adult cyclists with a passenger (see Table 4.11). The passengers were likely to be children. There were observations of the passengers sitting on the handlebars, whilst others were on a child seat on the back of the bike or in a trailer or cycle buggy. Table Child and adult cyclists with and without a passenger Children Adults Wearing Number Wearing Numbe r Cyclist (without passenger) 12.0% % 635 Cyclist (with passenger) % 2 Passenger only 25.0% % 5 Cyclist and passenger 75.0% % 9 Other differences between the distributions and wearing s of child and adult cyclists are summarised in Table Analyses of cycle helmet wearing s throughout the report have not included passengers in the calculations. TRL 26 PPR420

33 Table Wearing and distribution of child and adult cyclists, 2008 Variable/Category Children Adults Wearing Distribution Wearing Distribution Overall wearing 12.0% 1, % 3,540 Gender Male Female Additional safety aids Present Not present Day/time Weekday peak Weekday off-peak Weekend 9.0% 22.0% 25.4% 10.1% 11.9% 13.5% 11.6% % 20.7% 30.4% 15.7% 20.7% 14.2% 18.5% 2, ,931 1, ,376 Light/dark Light Getting light Dark Getting dark Riding position Road Pavement Cycle path Other Type of bike Racing bike Mountain bike/bmx Traditional town Other Weather Dry Raining Mixed 13.2% 2.9% 20.0% 7.3% 13.9% 10.5% 14.5% 27.3% 30.4% 11.8% 9.7% 15.3% 12.7% 8.4% 4.9% % 16.8% 24.2% 19.8% 19.4% 13.7% 24.8% 22.4% 48.7% 15.2% 15.2% 21.8% 18.9% 12.3% 17.9% 2, ,199 1, , , Interaction effects As with the major roads data, there are relationships between some of the factors analysed in Sections 4.2 to 4.5 above. These factors have been combined together in a model to assess the overall effect of the variables when other information has been taken into account. Initially a model of helmet wearing s for different age groups and sex was investigated. This showed that cycle helmet s are different for males and females and differ across age groups. The wearing s for these groups are shown in Table G.1 in Appendix G. Adding additional factors to this base model showed that bike type and the absence or presence of safety aids were the factors with the greatest effect on cycle helmet wearing (explaining 8.1% and 5.1% of the variation in the model, as shown in Table G.3). The variables that were found to be important in the model when describing differences in helmet wearing s were sex, age, area, riding position, bike type, additional safety aids, ethnicity and the interactions age and sex, area and bike and sex and ethnicity. TRL 27 PPR420

34 The model shows that: Females are more likely than males to be wearing helmets; Adults are more likely than children to be wearing helmets. The wearing is highest for female children, followed by female adults, male adults, and lowest for male children; The wearing is lower for riders of Mountain/BMX bikes, Traditional Town cycles and other cycles, compared with riders of Racer bikes; The wearing is significantly lower for cyclists on the pavement / verge than on the road or cycle-path; Cyclists with additional safety kit are more likely to wear a helmet; The wearing is significantly higher for white cyclists than for cyclists with other ethnic origins Those cycling in residential and shopping areas are less likely to be wearing a helmet than elsewhere (other than round school). TRL 28 PPR420

35 5 Conclusions The analyses of the two surveys have been carried out independently, and the results cannot be combined to give overall s. However, they show that wearing s on major built up roads are much higher for both adults (35.3%) and children (17.6%) than those on the minor built up roads (18.2% and 12.0% for adults and children respectively). 5.1 Major built-up roads The 2008 survey on major built-up roads showed that cycle helmet wearing was 34.3%, an increase from 30.7% in The wearing has increased in each survey; it was 16.0% in the first survey in The increase since 2006 is due to the increase in adult cyclists wearing helmets from 31.5% to 35.3% as the cycle helmet wearing for children has remained constant at 17.6% (the same as in 1994 and 2006, although lower s were found between 1996 and 2004). Female cyclists overall continue to be observed wearing helmets more than male cyclists, as in all previous surveys. The for male and for female cyclists has increased since 2006 and over time since the first survey in The for male cyclists has increased from 15.5% to 34.1% and for female cyclists from 17.2% to 34.9%. Whilst the cycle helmet wearing for female child cyclists has increased between 1994 and 2008 from 21.9% to 32.9%, the for male child cyclists has declined over time from 16.0% to 13.1%. As in 2006, the for male adult cyclists is higher than for female adult cyclists at 35.4% compared to 35.0%, increasing from 15.5% and 17.0% respectively. White cyclists continue to be more likely to wear a helmet than cyclists of other ethnic origin groups. The wearing s for all ethnic groups have increased since 2006, except for South East Asian and other. Both adult and child cyclists were significantly more likely to wear a helmet if riding a racing bike than an other bike, though mountain bike/bmx remained the most commonly observed type of bike. For the seventh survey (in 2008), helmets were significantly more likely to be worn during peak times than off peak times of the day for both child and adult cyclists, at 40.7% and 24.6% overall. As in previous surveys, cycle helmet wearing s at the London sites were significantly higher (69.5%) than at sites outside London (29.9%). On recreational routes there continued to be more cyclists wearing helmets than at other sites as well as there being more children observed on these routes. To examine cyclists' helmet wearing patterns more fully, the interaction between helmet wearing and age and sex of cyclist, type of bike, time of use, weather and ethnicity was analysed. Taking all these key variables into account, the overall conclusions are: In most areas cyclists on Mountain bikes are significantly less likely to wear a helmet than those on Racer bikes; similarly, cyclists cycling during off-peak hours are less likely to be wearing a helmet. Female children are significantly more likely to wear a helmet than male child cyclists, and the wearing for white cyclists is significantly higher than that for black cyclists. For adult cyclists in other cities on weekdays, there was an interaction between ethnicity and type of bike (cyclists more likely to wear a helmet if white and on a Traditional Town bike). There was also an interaction between sex and bike type for adults in non-city locations. Sufficient data for recreational routes was not available. TRL 29 PPR420

36 5.2 Minor built-up roads The purpose of the survey on minor built-up roads is to increase the sample of child cyclists and to be more representative of cycling patterns. The overall cycle helmet wearing increased to 16.7% in 2008 from 13.8% in The has increased each year the survey has been carried out since When analysed by age and gender, boys aged 7-10 years and years continue to be least likely to wear a helmet and the for boys aged has significantly decreased since 1999 from 8.6% to 7.0%. As in previous surveys and the major built-up roads survey in 2008, white cyclists were significantly more likely to wear a helmet than other ethnic groups. The wearing has increased for all ethnic groups since 2006, though significant only for adult cyclists. Cyclists were significantly more likely to wear a helmet if riding a racing bike, as in previous surveys on minor and major built-up roads, with the wearing increasing since 2006 for cyclists riding each type of bike. Once again, the wearing was greatest during weekday peak hours. As in all previous surveys cyclists were significantly more likely to wear a helmet if riding on a cycle path of a recreational route, then on the road rather than the pavement. Cyclists on minor built-up roads were more likely to be wearing a helmet in dry weather than when wet.. Once again, a cyclist was significantly more likely to wear a helmet if they had additional safety aids than if they were riding a bike without any additional safety aids (29.2% compared to 14.3%). In over half of the 55 cases where a cyclist was riding with a passenger in 2008, neither cyclist nor passenger were wearing a helmet, with both cyclist and passenger wearing one in a fifth of cases. There was no significant difference in wearing amongst children on minor build-up roads with time of day, when wearing a school uniform, when on a paper round, or in different weather conditions. But those riding a racing bike were significantly more likely to wear a helmet than when riding other types of bike. The wearing on minor built-up roads was significantly higher for children than adults in 1999 but this was reversed in all subsequent surveys. In 2008 the wearing for adult cyclists was 18.2% compared to 12.0% for child cyclists. Whereas adults were more likely to cycle on the pavement during peak hours, children were more likely to cycle on the pavement during peak and off-peak times of the day. Adults and children were more likely to wear a helmet if cycling on the road (19.4% and 13.9% respectively) rather than the pavement (13.7% and 10.5% respectively) and both most likely when on a cycle path ( 24.8% and 14.5% respectively) though the difference was only significant for adults. Analyses were carried out to identify the overall effects of a number of factors (age, sex, weather, area, ethnicity, peak/off-peak, light condition, riding position and safety aids) on the propensity of cycle helmet wearing. To compensate for lack of data available, age groups were compacted to adult/child (below 16 or above). The age/sex interaction showed that female children were more likely to wear helmets; wearing s indicated that female adults, male adults and male children followed. Notable factors increasing wearing s were having additional safety aids, riding a racing bike, being of white ethnic origin and riding on a road or cycle path compared to riding on the pavement. TRL 30 PPR420

37 Acknowledgements The work described in this report was carried out by Accent and the Safety Group of the Transport Research Laboratory. The authors are gful to Jeremy Broughton who carried out the technical review and auditing of this report. References Bryan-Brown, K (2000). Cycle helmet wearing in TRL Report 487. TRL Limited, Crowthorne. Bryan-Brown, K and Taylor S (1997). Cycle helmet wearing in TRL Report 286. TRL Limited, Crowthorne. Christie, N (2000). Cycle helmet promotion worldwide. TRL Unpublished report PR/SE/094/00. Department for Transport (2000). Tomorrow s roads- safer for everyone. The Government s road safety stgy and casualty reduction targets for Department for Transport (2004). Tomorrow s roads- safer for everyone. The first three year review. The Government s road safety stgy and casualty reduction targets for Government Statistical Service (1999). Regional Trends 34. London: The Stationary Office. Gregory K, Inwood C and Sexton B (2003). Cycle helmet wearing in TRL Report 578. TRL Limited, Crowthorne. Inwood C, Whitley C and Sexton B (2005). Cycle helmet wearing in TRL Report 644. TRL Limited, Crowthorne. McGarry T and Sheldon R (2007). Cycle helmet wearing in Road Safety Research Report No.84. Department for Transport. Nagelkerke, N. J. D. (1991). A note on a general definition of the coefficient of determination. Biometrika, Vol. 78, No. 3: Office for National Statistics (2002). Journey to work table 2001 Census. Pallant, J (2005). SPSS Survival Manual. Open University Press, Berkshire. 2 nd edition. Taylor, S B and M E Halliday (1996). Cycle helmet wearing in Great Britain. TRL Report 156. TRL Limited, Crowthorne. Towner E, Dowswell T, Burkes M, Dickinson H, Towner J and Hayes M (2002). Bicycle Helmets A review of their differences: A critical review of the Literature. Safety Research Report No. 30. Department for Transport. London. TRL 31 PPR420

38 Appendix A Site locations 1 London Road roundabout, Derby 2 Lexden Road, Colchester 3a Cycle Route, Central Park, Chelmsford (westbound flow) 3b Cycle Route, Central Park, Chelmsford (eastbound flow) 4 A6 High Street, St Pauls Sq, Bedford (7am-1pm), southbound flow) 5 A6 High Street, St Pauls Sq, Bedford (1pm-7pm), southbound flow) 6 A6 High Street, St Pauls Sq, Bedford (7am-1pm), northbound flow) 7 A6 High Street, St Pauls Sq, Bedford (1pm-7pm), northbound flow) 8 London Rd., Newark, Nottingham 9 Entrance to Uni' west / Beeston Lane, Nottingham 10 Castle Boulavard, Nottingham 11 A560 Stockport Rd., Cheadle, Stockport (westbound flow) 12 A560 Stockport Rd., Cheadle, Stockport (Eastbound flow) 13 A56 Chester Rd., Stretford, Trafford, Manchester (NE bound flow) 14 A56 Chester Rd., Stretford, Trafford, Manchester (SW bound flow) 15 A347, Talbot Avenue, Dorset 16 Robin Hood Walk, Boston, Lincoln (Southbound) 17 Robin Hood Walk, Boston, Lincoln (Northbound) 18 North Bar, Beverley, Humberside 19 Holderness Rd. and Craven St, Hull 20 Ashby Road, Scunthorpe 21 Bargate, Grimsby 22 Victoria Park Rd/Queens Rd/Victoria Pk, Leicester 23 A512 Ashby Road, Loughborough, Leicestershire 24 Great Western Road / Byres Road, Glasgow (7am 1pm) 25 Great Western Road / Byres Road, Glasgow (1pm 7pm) 26 Great Western Road / Byres Road, Glasgow (7am 1pm) 27 Great Western Road / Byres Road, Glasgow (1pm 7pm) 28 Pollokshaws Rd / Langside Ave, Glasgow (7am 1pm) 29 Pollokshaws Rd / Langside Ave, Glasgow (1pm 7pm) 30 Pollokshaws Rd / Langside Ave, Glasgow (7am 1pm) 31 Pollokshaws Rd / Langside Ave, Glasgow (1pm 7pm) 32 North Road, Cardiff 33 Binley Road, Coventry 34 Waterloo Bridge, London 35 Hyde Park, London 36 Westminster Bridge, London 37 Blue Bell Road, Norwich 38a A532 Hungerford Road, Crewe 38b Padgate Lane, Bruche Ave, Warrington, Cheshire (out of Warrington) 38c Padgate Lane, Bruche Ave, Warrington, Cheshire (into Warrington) 39 Marlowe Ave, Swindon, Wiltshire 40 A590, Jubilee Bridge, Barrow-in-Furness 41 Bristol - Bath Cycle Track, Bristol (towards Bath) 42 Bristol - Bath Cycle Track, Bristol (towards Bristol) 43 Bristol A38 North of Haymarket 44 Tenterbanks, Stafford 45 A454 Willenhall Road, Wolverhampton 46 Great North Road, Newcastle 47 A638 Bawtry Road, Doncaster (southbound) 48 A638 Bawtry Road, Doncaster (northbound) 49 Union StreeVHolburn Street, Aberdeen TRL 32 PPR420

39 50 Rushmere Road, Northampton 51 Middle Meadow WaWNorth Meadow Walk, Edinburgh 52 Myton RoadlBanbury Road, Warwick 53 Ashlawn Road, Rugby, Warwickshire 54 A1057 St. Cuthberts Way, Darlington 55 Oxford Street, Liverpool 56 Ullet Road, Liverpool 57 Gasworks Pipe Bridge - Oxford Cycle Route, Oxford 58 B4O00, Ridgeway, Oxford 59 Parks Road, Oxford 60 Botley Rd and Railway Station, Oxford 61 Normanston Drive, Lowestoft, Suffolk 62 Grimwade Street, Ipswich, Suffolk 63 Inbound crossing of Gosprt to Portsmouth Ferry, Portsmouth 64 A329 Wokingham Road, Berkshire 65 Cinder Lane (Jubilee Terrace), York 66 Dean Gate, York Minster, York 67a Bridge Street, Cambridge (into city) 67b Bridge Street, Cambridge (from city) 68 Garret Hostel Lane, Cambridge 69 Silver Street, Cambridge 70 Westfield Road, Peterborough 71 Town Bridge, Peterborough 72 A430 Northgate Street/Worcester Street, Gloucester 73 Prestbury Road, Cheltenham 74 Castle Eden Walkway I Cycleway, Stockton (7am 1pm) 75 Castle Eden Walkway I Cycleway, Stockton (1pm 7pm) TRL 33 PPR420

40 Appendix B Data collection forms for major built-up roads survey TRL 34 PPR420

41 TRL 35 PPR420

42 TRL 36 PPR420

43 Appendix C Letter of Authority TRL 37 PPR420

44 Simon Maxwell Road User Safety Division Department for Transport Zone 2/13 Great Minster House 76 Marsham Street London SW1P 4DR Direct Line: Web Site: To whom it may concern, 28 July 2008 Cycle helmet wearing survey 2008 The Department for Transport (DfT) has commissioned TRL and Accent to undertake its cycle helmet wearing survey This year s survey builds on previous surveys undertaken in 1994, 1996, 1999, 2002, 2004 and These surveys of actual pedal cycle helmet wearing s enable DfT to assess changes in wearing s over time and inform policy on the use of cycle helmets. The road safety stgy Tomorrow s Roads - Safer for Everyone, published in 2000, commits DfT to monitor cycle helmet wearing s. Accent will be undertaking the data collection. Accent is a full service MRQSA accredited research agency, and I would be gful for any assistance you are able to provide to enable them to undertake the survey on our behalf. If you wish to check Accent's credentials you can call the Market Research Society free on If you would like any further information about the survey please contact Catherine Sharratt at TRL (Tel: ; csharratt@trl.co.uk), Teresa McGarry at Accent (Tel: ; teresa.mcgarry@accent-mr.com), or myself at the above address. Yours sincerely, Simon Maxwell TRL 38 PPR420