Annex A - Research Methods
What do we mean by risk?
Looking at risk
In designing the questionnaire, a number of aspects of risk needed to be taken into account. First, the level of risk (probability of a defined type of accident) may be expressed in absolute terms - e.g. the number of riders out of 100 who would have an accident in a year - or relative to the risk of other activities. Secondly, it is necessary to distinguish between estimates of objective risk derived from accident statistics, and respondents' own estimates and perceptions of risk. This survey was able to make use of results from previous research (Sexton et al, 2004b) to estimate the objective risk of riders of a given age, experience and mileage. Thirdly, risk estimates may be needed for different populations - e.g. motorcyclists in general, or riders who are like the respondent in some way ( e.g. with respect to age, experience and annual mileage). Fourthly, risk estimates for different types or severities of accident may be needed.
All these approaches were used in the questionnaires. For example, riders were asked to imagine 100 typical riders, and say how many they thought would have accidents of various severities in the next 12 months. They were then asked the same question but for riders of the same age, sex and experience as themselves, and doing a similar amount and type of riding. A further question sought riders' estimate of their own risk by asking them to imagine 100 riders exactly like themselves in every way. Other questions presented the respondent with information on risk ( e.g. the statistical risk of riders of their own age, sex and experience, or the published estimate that, per mile travelled, motorcycle riders are 25 times more likely to be killed than car users. Comparisons of motorcycling risk with the risks of other transport modes, and with the risks of sports such as hang-gliding and rock climbing were also sought.
Relationship between risk and sensation seeking
Studies by Zuckerman and others showed that risk and sensation seeking are related. Zuckerman (2000) says that "risk-taking is not the main point of sensation-seeking behaviour; it is merely the price such people pay for certain kinds of activities that satisfy their need for novelty, change and excitement.
Sensation-seeking can extend to the physical, involving unusual or extreme sports such as skydiving, hang gliding, scuba diving, auto racing, rock climbing and white-water kayaking. Motorcycling could be included as such an activity. An interest in participating in such sports or activities describes one subcategory of sensation-seeking: thrill- and adventure-seeking. Arnett (1994) developed scales measuring sensation-seeking, they include two sub-scales which measure the novelty and the intensity of the sensation.
Risk and the relationship with perceived control
Starr (2004) introduced a concept of voluntary risk. Research in this area has found that humans tolerate substantially more risk when they engage in voluntary behaviour (Sjoberg 2004). This is related to a sense of controllability where less risk is perceived in situations that are under personal control, 'Risk is sometimes defined as insufficient controllability' (Brun 1994). People are found to believe that they are more in control than they actually are. This helps to explain why riders (and drivers) think they are generally better and less at risk than the average rider (or driver), because they are in control.
Self-completion questionnaire development
The questionnaire for the survey was based on one used in an earlier study (Sexton et al, 2004). This included items on rider style, rider behaviours and rider attitudes. It was refined in order to better reflect the requirements of this project. This required some additional items on ways of measuring risk and some further items on the attitudes to perceived risks. The additional questions were determined from previous risk research as considered within the initial review process. The Arnett SSS (Sensation Seeking Scale) was used in order to provide an indication of the sensation seeking tendencies of riders (Arnett, 1994). The final form of the self-completion questionnaire is shown in Annex C.
Interview structure development
The interview structure was a combination of structured format and open-ended. The structured element was required in order to allow the use of materials and pre-considered questions. These have been used to determine the interviewees' perceptions of risk, attitudes to risk as well as basic demographic data plus information on motorbike use. The form of the structured elements relating to risk and attitudes towards risk involved the following:
- Pictures showing different risk scenarios and ask the interviewee to rate them.
- Use of a statistical model (from Sexton et al, 2004b) to predict the accident liability of the interviewee and record their response and reasons for acceptance or otherwise of that risk.
The final form of the interview schedule was determined from the review of risk measurement research as well as drawing on the experience of the project team. A small pilot of the materials was conducted with motorcyclists at the Scottish Motorcycle Show in March 2005. As a result of the pilot, the interview structure was modified slightly. The open-ended part of the interview was led by trained interviewers and explored all aspects of the risks, attitudes and experiences of the interviewee - it was not fully structured but there was a framework to guide the discussion. The interview took around 30 minutes to administer. The final form of the interview questionnaire is given as Annex D.
Sample population
In theory the sample population for this survey were all motorcyclists who ride in Scotland. However, it was not practical within the limitations of the project to select a random sample of Scottish motorcyclists.
It is well established that there are far more killed and seriously injured motorcyclists on non-built up roads and that risky rider behaviour is most likely to be observed on these roads. It is appreciated that there are more motorcycle accidents on built-up roads, but many of these are slight accidents where the fault is often the 'other' vehicle. Hence, the samples targeted were recreational riders on non-built up roads during weekends, or attending a motorcycle event ( e.g. British Superbike weekend at Knockhill). Some of these riders will also use their bikes for commuting during the week.
Samples for interviews
The interviews were conducted by teams of at least two interviewers at places where bikers gather, either before a ride or after a ride. This included cafes, meeting spots (with a catering van), motorcycle shows and at Knockhill Circuit during the British Super Bikes racing weekend.
A cross-section of different rider types was approached. Interviewers were able to conduct one or two interviews per hour and distributed questionnaires to other riders between interviews. Given that many riders only ride during the warmer months, i.e. from April to October, then the interviews were conducted during May, June and July 2005. In this period, a total of 124 interviews were conducted.
Samples for questionnaire survey
The samples for the questionnaire survey came from three main sources. Those riders who take part in the interview sample were given a questionnaire to complete and post back to the research team. This was seen as potentially useful because the interviewer cannot ask all of the questions of interest within an interview situation. Also the questionnaire contains questions relating to risk and attitudes to risk which will be different from those used in the structured part of the interview. Unfortunately, in practice there were only 10 of the interviewees who returned the self-completion questionnaire, which was too few to do any comparative analysis.
The next sample for the questionnaires also came via the interviewers. They approached riders but did not interview them, had a very brief chat about this survey and asked them to complete the questionnaire and post it back to the research team.
Questionnaires were also distributed via motorbike dealers, clubs and organisations. The purpose of the survey was explained, and it was emphasised that the research is aimed at obtaining a better understanding of motorcyclist risk acceptance and attitudes towards risk.
Finally, around 500 questionnaires were distributed to attendees at the BMF Kelso Bikefest in July 2005.
In total, over 2,000 questionnaires were distributed throughout Scotland during the summer of 2005. The riders who received questionnaires covered the spectrum of riders and their selection was opportunistic. Hence there was no way in which those riders who took or were given a questionnaire could be reminded to complete it and return it. This meant that there was no way to influence the return rate. The survey was closed in August 2005 by which time 364 completed questionnaires had been returned representing a response rate of around 17%.
Focus groups
Riders who were interviewed were asked if they would be prepared to attend a focus group and were offered a small incentive to do so. It was anticipated that 3 focus group sessions would be run with a mix of 8-10 riders in each. However, in practice so few riders agreed to attend a focus group that this stage of the research was not tenable and so was abandoned.