Can social psychological models be used to promote bicycle helmet use among teenagers? A comparison of the Health Belief Model, Theory of Planned Behavior and the Locus of Control

Abstract

Problem: The bicycle helmet use rate is still low among teenagers despite the cumulating evidence that bicycle helmets can prevent cyclists from serious injuries and death. The objective of this study was to investigate the usefulness of the Health Belief Model (HBM; Health Education Monographs, 2 (1974) (1), Theory of Planned Behavior (TPB; Ajzen, I. (1988). Attitudes, personality and behavior. Open University Press, Milton Keynes) and Locus of Control model (LC; Psychological Monographs, (1966) (80) in understanding the intention to use bicycle helmet use among bicycle helmet owners. Method: Data were collected at two schools in Helsinki, Finland. Students (N=965) completed a questionnaire including three social psychological models applied to helmet use. Models were compared by structural equation modeling techniques. Summary: Results showed that the TPB and LC model fitted the data well, whereas fit of the HBM model was lower than the fit of TPB and LC models. All components of TPB and external LC orientation were significantly related to the intention to use a helmet. TPB together with LC model provide a promising theoretical framework for helmet use promotion campaigns. Practical suggestions for future bicycle helmet campaigns were provided.

Introduction

Numerous studies have shown that a bicycle helmet is an effective way of preventing head injury. In their recent meta-analytical study, Attewell, Glase, and McFadden (2001) summarized international studies published between 1987 and 1998 and investigated the effectiveness of bicycle helmets. The summary odds ratio estimate for efficacy was 0.40 for head injury, 0.42 for brain injury, 0.53 for facial injury, and 0.27 for fatal injury. Based on these results, Attewell et al. concluded that wearing a bicycle helmet while cycling reduces risk of head and brain injuries as well as facial and fatal injuries. Similar findings have been reported also in Finnish studies. Karttunen and Karkola (1992) studied all fatal cycling crashes between 1986 and 1990 and found that almost half of the fatalities among adult bicyclists could have been prevented by using a helmet. Olkkonen (1998) reported that high helmet wearing rates were clearly related to a smaller number of head injuries.

Although wearing a bicycle helmet seems to reduce the risk of head injury considerably, the bicycle helmet wearing rates are still rather low in countries in which wearing a bicycle helmet is not mandatory. In 1999, the helmet wearing rate for the whole of Finland was only 21% (Parkkari, 2000). In Lindahl's (1998) survey conducted in urban areas of Sweden, the estimated percentage of bicycle mileage during which a helmet was worn was 17.2% in 1997 and 18.4% in 1998. The small increase in helmet use within a year was not statistically significant. In a recent telephone survey conducted among 1,000 cyclists in the United States, Rodgers (2000) reported that 42.9% of cyclists claimed that they use a bicycle helmet “always or almost always.” The corresponding wearing rate was 27.3% in 1991, indicating that the number of cyclists has somewhat increased. It should be noted, however, that almost equal (43.0%) number of cyclists used a bicycle helmet “never or almost never” (Rodgers, 2000). Hence, the number of cyclists using a bicycle helmet is still very low.

Rodger's (2000) study was based on self-reports of helmet use in telephone interviews, which might explain relatively high wearing rates. It is also possible that in the United States cycling is more a leisure-time activity than an essential transportation mode when compared to Scandinavia, which may also be reflected in helmet use. The sample included both adult and child cyclists, which may partly explain the relatively high bicycle helmet wearing rate. Previous studies show that the helmet wearing rate is highest among primary school children. In Povey, Frith, and Graham's (1999) study conducted in New Zealand, the helmet wearing rate was over 80% for children (before helmet wearing became mandatory in 1994), whereas teenagers and adults seem to use bicycle helmets less frequently Cameron et al., 1994, Ekman et al., 1997, Povey et al., 1999, Schuffham & Langley, 1997, Sissons Joshi et al., 1994, Wood & Milne, 1988. The low helmet wearing rate is especially a problem among teenagers who cycle frequently. In Scandinavian countries, for example, a bicycle is an important means of transportation among teenagers, because distances are often too long for walking but optimal for cycling (Lehmuskoski, Vähä-Rahka, & Voltti, 2001). Therefore, teenagers should be considered as a main target group for bicycle helmet wearing campaigns and studies.

Some safety experts have suggested that providing free or low-cost bicycle helmets to teenagers would increase the bicycle helmet wearing rate significantly. In Coron, McLaughlin, and Dorman's (1996) study, 32.9% of students found bicycle helmets too expensive and, therefore, did not wear one while cycling. Based on this finding, Coron et al. (1996) suggested that reduced bicycle helmet prices may be effective in increasing the bicycle helmet wearing rate among college students. Although having a helmet is a precondition for using one, several studies show that it does not automatically lead to high wearing rate. Everett Price, Bergin, and Groves (1996) reported that only 55% of bicycle helmet owners wore their helmet during most of their rides. Jacques (1994) reported a wearing rate of 44% for children who own a bicycle helmet and concluded that providing helmets would not result in expected increase in helmet use due to lack of rider motivation. Despite this mismatch between the bicycle helmet ownership and wearing rate, many studies about bicycle helmet use have been conducted among cyclists in general, without asking if they have a helmet or not. It is likely that factors leading to the decision to buy a bicycle helmet are partly different than those related to the intention to use one. Among teenagers, it can be inferred that the purchase decision is mainly up to parents, whereas the intention to wear a helmet depends mostly on a teenager him/herself. In the present study, factors influencing the intention to wear a bicycle helmet were studied in a large sample of upper secondary and high school students having a bicycle helmet.

As Jacques (1994) stated, bicycle helmet use seems to depend mostly on a cyclist's motivation to use bicycle helmets rather than environmental, exposure, or cost related factors. Social psychological theories, like the Theory of Planned Behavior (TPB; Ajzen, 1988), Health Belief Model (HBM; Rosenstock, 1966, Rosenstock, 1974), and Locus of Control (LC; Rotter, 1966, Wallston et al., 1978) provide a potentially fruitful framework to understand why cyclists are so unwilling to use bicycle helmets even if they have one. Illustrations of these models can be found in Fig. 1, Fig. 2, Fig. 3. In their prospective study among 97 schoolboy cyclists, Arnold and Quine (1994) reported that the components of HBM explained 53% of variance in bicycle helmet use. In the same study, Health Locus of Control (health behavior targeted LC) variables did not have significant effects on bicycle helmet use when added to HBM. In a more recent study, Quine, Rutter and Arnold (1998) compared the ability of TPB and HBM to predict bicycle helmet use in a sample of school children. Again, the sample consisted only of boys. Results showed that TPB (43% of variance explained) had greater predictive utility with less redundancy than HBM (18% of variance explained). Based on these encouraging results, Quine, Rutter, and Arnold (2001) conducted a follow-up study based on TPB components. Ninety-seven child cyclists were randomly assigned to intervention or control conditions. The intervention group was presented with a booklet containing a series of persuasive messages based on TPB. According to the results, at 5-month follow-up, none of the control children had taken up helmet wearing, while 25% of the intervention group had (Quine et al., 2001). Based on this finding, the authors concluded that social cognition theories such as the TPB could be valuable in the design of effective interventions to change health behaviors.

In the studies by Arnold and Quine (1994) and Quine et al. (1998), the social psychological models were tested solely in a sample of schoolboys, because they ride a bicycle more and are, therefore, more exposed to crash risk than schoolgirls. In Finland and other Scandinavian countries, however, cycling is popular among girls. According to the Finnish cycling statistics, 6–17 year old boys travel 26% of their trips by a bicycle whereas the same figure for girls is 21%. Since Finnish girls seem to ride a bicycle almost as much as boys, girls should be included in studies about bicycle helmet use.

The HBM Rosenstock, 1966, Rosenstock, 1974 focuses on two aspects of health behavior: threat perception and behavioral evaluation (see Fig. 1). Threat perception includes two components, susceptibility to a cycling crash and anticipated severity of the consequences of a crash (e.g., likelihood of a head injury). Behavioral evaluation consists of two distinct sets of beliefs, those related to barriers (e.g., inconvenience, peer pressure) to helmet use and those concerning benefits (e.g., increased safety). In addition to threat perception and behavioral evaluation, also “cues to action” and “health motivation” components were included to the HBM. “Cues to action” refers to triggers to use a bicycle helmet (e.g., helmet in a visible place) whereas health motivation refers to one's readiness to be concerned about health matters (e.g., Becker, Haefner, & Maiman, 1977). Unlike the other components in the model, health motivation is not specific to cycling but measures how much a person values good health and safety in general.

According to TPB (described in Fig. 2), the immediate cause of volitional behavior is one's intention to engage that behavior. Behavioral intention is, in turn, determined by a person's attitude toward the behavior concerned and by his or her subjective norm. Attitudes are a person's overall evaluations of the behavior (e.g., “it is good to use a bicycle helmet”) whereas subjective norm consists of a person's beliefs about whether significant others (e.g., parents, peers) think that he or she should engage that behavior (Conner & Sparks, 1996). For example, a teenager may think that his or her parents like to see him or her using a bicycle helmet whereas he or she might expect his or her peers to see bicycle helmet use as ridiculous. According to TPB, perceived behavioral control forms the third predictor of intentions. Perceived behavioral control refers to a person's perception of the extent to which performance of the behavior is easy or difficult (Conner & Sparks). As Fig. 2 shows, perceived behavioral control has both direct and mediated (by behavioral intention) effects on behavior in TPB.

Recent studies have suggested that it is important to distinguish affective from instrumental attitudes in relation to behaviors, which might reasonably be considered as having an affective component Lawton et al., 1997, Parker et al., 1998, Parker et al., 1996. Instrumental attitudes are related to benefits (e.g., “it is beneficial to use a bicycle helmet”) and disadvantages (e.g., “it is uncomfortable to use a bicycle helmet”) of using a bicycle helmet. Emotional attitudes refer to emotions related to helmet use (e.g., “using a helmet feels stupid”). In this study, instrumental and emotional attitudes were measured separately.

The third model tested in this study was LC (Rotter, 1966). In our current study, the LC model included two components, which were external and internal LC belief orientation (see Fig. 3). Internal LC refers to belief that events (e.g., crashes) are consequences of one's own actions and therefore controllable, whereas people with external LC orientation believe that events are unrelated to their behavior and beyond their personal control. Supposedly, “internals” can be expected to be more motivated to use bicycle helmets than “externals,” because they believe that a bicycle helmet would reduce the risk of a head injury in the case of a cycling crash.

LC can be measured either as a general orientation in life or limited to a specific domain. Because the amount of variance in health behavior explained by general LC has been typically low, domain-specific scales have been developed Conner & Sparks, 1996, Lefcourt, 1991. One of the most frequently used instruments for measuring health locus of control is the Multidimensional Health Locus of Control scale (MHLC) developed by Wallston et al. (1978). In the present study, however, we found MHCL still too general and decided to target the LC measure strictly to cycling situation. Since the sample consisted of teenagers, questions about beliefs about causes of illnesses might not be appropriate. For example, the risk of heart diseases and lung cancer is a very distant and vague health risk for a 15-years old smoker, but young cyclists expose themselves to the risk of a cycling crash and head injury every time they ride a bike without a helmet.

In previous studies applying HBM, TPB, or LC, these three models have not been tested simultaneously by using the same sample of bicycle helmet owners. The aim of the present study was to compare the feasibility of the HBM, TPB, and LC in explaining teenager bicycle helmet owners' intention to use their helmets by using structural equation modeling techniques.