1. Background
Driving accidents are one of the main health problems threatening individuals’ health. A Global report on road safety in 2018 stated that the road traffic events annually account for 1.35 million deaths (1). In Iran, road traffic accidents have the highest fatality rates worldwide (2), and the disability-adjusted life years (DALYs) caused by road accidents are more frequent than other diseases such as cardiovascular diseases and cancers (3).
Evidence has revealed that seatbelts reduce serious crash-related injuries and deaths by about half as seatbelts saved about 15,000 lives in 2016 (4, 5). Wearing seatbelts varies considerably from one country to another, and the highest levels of seatbelt use have been reported in developed countries (6). Moreover, the rate of seatbelt use varies on rural or urban roads among drivers and occupants. In this regard, a study in Turkey reported the rates of seatbelt use on rural and urban roads to be 71% and 21%, respectively (7). In a study in Iran, the rates of seatbelt use were reported to be 81% - 91% for drivers and 34% - 44% for front-seat passengers (8). Although wearing seatbelts is mandatory, it seems that many front-seat passengers, especially the ones on urban roads, do not use seatbelts. Accordingly, it is of great importance to detect the psychological factors affecting such a behavior.
The seatbelt use is influenced by psychological factors such as attitudes, beliefs, and intentions (9-12). Social psychological theories are useful in identifying such factors. One of the most-widely adopted psychological theories in describing high-risk behaviors such as seatbelt use is the theory of planned behavior (TPB) (3, 7, 13).
According to TPB, the main determinant of behavior is intention, and the direct determinants of intention are attitudes toward performing a behavior, subjective norm associated with a behavior, and perceived control (14). In the extended TPB, attitude encompasses experiential attitude (emotional assessment of seatbelt use) and instrumental attitude, which is determined by the individual’s beliefs about the outcomes of wearing seatbelts (behavioral beliefs), weighted by evaluating those outcomes (e.g., the expected consequences of seatbelt use) (15). In this theory, perceived norms include descriptive norms (believing in whether famous individuals use seatbelts) and subjective norms determined by his or her normative beliefs (i.e., whether important referent individuals approve or disapprove seatbelt use, weighted by an individual’s motivation to comply with those referents) (14, 15). According to the extended TPB, instead of perceived behavioral control, personal agency is examined, which encompasses self-efficacy (i.e., an individual’s assessment of one’s ability to use a seatbelt by overcoming obstacles) and perceived behavioral control (PBC) determined by control beliefs concerning the presence or absence of facilitators and barriers to seatbelt use, weighted by the perceived power of each control factor to facilitate or inhibit using seatbelts (15).
Although several studies have revealed that TPB constructs are predictors of seatbelt use (3, 7, 13), there have been contradictions in the findings of previous studies. For example, in Simsekoglu’s study, perceived behavioral control was not a predictor (7), while Tavafian found that perceived behavioral control was one of the strongest predictors of using seatbelts (3). In this regard, there is little evidence available on behavioral change theories, and further research is needed to better understand factors affecting the seatbelt use (12).
2. Objectives
To the best of our knowledge, no study has considered all the constructs of the extended TPB in predicting the seatbelt use in front-seat passengers. Accordingly, the main objective of this study was to assess how much the constructs of extended TPB can predict seatbelt use in front-seat passengers on urban trips.
3. Methods
3.1. Participants
In this cross-sectional study, the participants encompassed 328 men and women residing in Bushehr, Southwest of Iran, who had traveled on urban roads during the past week as a front-seat passengers using personal cars. The participants were selected using a multistage, stratified cluster random sampling method. Bushehr is divided into 75 districts according to the municipal classification. In this study, 15 districts were randomly selected, and the sample size of each stratum was determined proportional to the number of households residing in the concerned district. The eligible participants were invited to participate in the study until the required sample size was reached for the stratum. Sampling was performed using two questioners from March to May 2018 at two shifts in the morning and afternoon by visiting the participants’ houses.
The sample size was determined by adopting the precision power method and using a minimum sample size formula. Brooks and Barcikowski developed this formula to calculate the sample size (16).
where, ε is the acceptable absolute shrinkage; ε = 0.1 R2, p is the number of explanatory variables, and R2 is a priory effect size adopted from Tavafian et al. study (3). The sample size was estimated to be 270 persons; however, the sample size was set to be 350 to well-fit the findings.
In the sampling process, 350 participants were evaluated, 22 of whom refused to participate in the study; as such, they were excluded from the study. The subjects, who were excluded from the study, completed the demographic section of the questionnaire, and the data analysis revealed no significant difference between these subjects and the participants in this regard. The final sample size in the data analysis process was 328) response rate = 93.7%).
3.2. Instruments and Measures
In this study, the questionnaire consisted of three sections: (1) demographic variables; (2) TPB constructs, including experiential, and instrumental attitude, descriptive, and subjective norms, self-efficacy, and PBC; and (3) outcomes including intention and seatbelt use.
The demographic variables included age, gender, marital status, job, level of education, wearing a seatbelt in an accident as a front-seat passenger, a history of penalties for not wearing seatbelts as a front-seat passenger, personal opinion about the impact of laws on the seatbelt use in a front-seat passenger, and the history of seatbelt use as a driver.
To identify relevant behavioral attributes or outcomes, normative referents, and facilitators and barriers to seatbelt use, structured interviews were performed (15, 17). Structured interviews were held with the participation of 32 eligible participants. This collected data were then consisted of the questionnaire content.
Ten health education experts reviewed the first draft of the questionnaire. The extent of experts’ agreement on the questions was assessed using content validity ratio (CVR). CVR ≥ 0.62 represented appropriate content validity (18). CVR for each question ranged from 0.64 to 1, and the total CVR value of 60 items was estimated to be 0.89, indicating an acceptable value.
The reliability of the questionnaires was examined using Cronbach’s alpha coefficient for a sample of 34 participants. The reliability coefficients were 0.42, 0.79, 0.86, 0.78, 0.74, and 0.93 for experiential, and instrumental attitude, descriptive, and subjective norms, PBC, and self-efficacy, respectively. Following the item analysis, one item on the experiential attitude was deleted, and Cronbach’s alpha coefficient was estimated to be 0.72.
In this study, the constructs were evaluated using indirect measures (15). The experiential attitude was assessed for eight items with a bipolar scale (scored from -2 to +2). Instrumental attitude was also assessed for seven paired items of behavioral belief / outcome evaluation (bipolar-scale score).
Descriptive norms were assessed using five items with a bipolar scale, and subjective norms included five paired items of normative belief (bipolar scale)/motivation to comply (unipolar scale). Moreover, self-efficacy was assessed by seven items with a unipolar scale, and PBC contained eight paired items of control belief (bipolar scale)/ perceived power (bipolar scale). The scales were scored based on the guideline of the TPB questionnaire (15).
TPB outcomes included the intention to use seatbelts. Such an intention was assessed by five items with a unipolar scale. The behavior of using a seatbelt was assessed with two questions on the use or non-use of the seatbelts, as well as the use of a seatbelt for enforcing law or other reasons. The questions were scored based on a five-point Likert scale (ranging from never to always).
3.3. Data Analysis
The data collected from 328 questionnaires were analyzed by the statistical package for social sciences (SPSS) version 22.0. Descriptive statistics were used to assess the demographic variables of the participants. Then the participants were divided into three groups: Group 1, the participants who self-reported using seatbelts for enforcing the law; Group 2, the participants who used seatbelts for the sake of personal safety (not enforcing law); and Group 3, the participants who did not use seatbelts. The mean differences of the scores for the TPB constructs among the three groups were determined using one-way MANOVA. Chi-squared test was also used to evaluate the relationship between using seatbelt and qualitative demographic variables. Multiple linear regression analysis was conducted to assess the predictive constructs of intention to use seatbelts. Logistic regression analysis was also run to assess the predictive constructs of seatbelt use.
Shapiro-Wilks test was also used to check the normality of the continuous data. The distribution of data was normal for most of the constructs, including instrumental attitudes (P = 0.076), subjective norm (P = 0.203), PBC (P = 0.353), and self-efficacy (P = 0.086). The cases of non-normality can be disregarded with regard to the large sample size.
4. Results
A total of 328 participants aged 19 - 65 years old with the mean age of 38.51 (SD = 11.26) years took part in this study. Two hundred and thirty-one (70.4%) participants were female, and most of these individuals were married (77.4%, n = 254). A majority of the respondents (41.2%, n = 135) had college degrees. Moreover, 145 participants (44.2%) in this study were housekeepers. Only 21% (n = 69) of the participants had a history of road accident. Forty-nine participants (69.6%) had their seatbelt worn at the time of the accident. In addition, 55 (16.8%) participants had previously been fined by police for not using seatbelts as front-seat passengers.
Of all the subjects, 28.7% (n = 94) persons reported that they did not use seatbelts as front-seat passengers, 27.7% (n = 91) used seatbelts for enforcing laws, and 43.6% (n = 143) used seatbelts for personal safety. There were no significant differences between the participants who did not use seatbelts and those who used seatbelts for enforcing law or personal safety with regard to the participants’ demographic characteristics (Table 1).
Table 1.Distribution of Demographic Variables Under Different Conditions of Using Seatbelts
| Characteristics | Using Seatbelt for Law Enforcement (N = 91), No. (%) | Using Seatbelt for Safety (N = 143), No. (%) | Not Using Seatbelt (N = 94), No. (%) | χ2 | P Value |
|---|---|---|---|---|---|
| Sex | 0.13 | 0.937 | |||
| Male | 27 (27.8) | 41 (42.3) | 29 (29.9) | ||
| Female | 64 (27.7) | 102 (44.2) | 65 (28.1) | ||
| Married status | 4.99 | 0.288 | |||
| Single/divorced | 17 (23) | 33 (44.6) | 24 (32.4) | ||
| Married | 74 (29.1) | 110 (43.3) | 70 (27.6) | ||
| Education level | 7.78 | 0.455 | |||
| Less than diploma | 24 (30.8) | 35 (44.9) | 19 (24.3) | ||
| Diploma | 32 (27.8) | 53 (46.1) | 30 (26.1) | ||
| Academic | 35 (25.9) | 55 (40.7) | 45 (33.3) | ||
| Job | 12.63 | 0.556 | |||
| Housekeeper | 43 (29.7) | 63 (43.4) | 39 (26.9) | ||
| Employee | 10 (22.7) | 20 (45.5) | 14 (31.8) | ||
| Others | 38 (27.3) | 60 (43.2) | 41 (29.5) | ||
| Type of trips | 8.58 | 0.073 | |||
| Urban | 67 (28.9) | 96 (41.4) | 69 (29.7) | ||
| Rural | 9 (15.5) | 33 (56.9) | 16 (27.6) | ||
| Both | 15 (39.5) | 14 (36.8) | 9 (23.7) | ||
| History of accident | 0.7 | 0.706 | |||
| Yes | 20 (29) | 32 (46.4) | 17 (24.6) | ||
| No | 71 (27.4) | 111 (42.9) | 77 (29.7) | ||
| Seatbelt use in accident | 4.48 | 0.106 | |||
| Yes | 13 (27.1) | 26 (54.2) | 9 (18.8) | ||
| No | 7 (33.3) | 6 (28.6) | 8 (38.1) | ||
| Driving fine for seatbelt | 5.45 | 0.066 | |||
| Yes | 16 (29.1) | 17 (30.9) | 22 (40) | ||
| No | 75 (27.5) | 126 (46.2) | 72 (26.4) |
The TPB constructs in the three groups are statistically compared in Table 2. Significant differences were observed among the three groups regarding the constructs of intention (P < 0.001), experiential attitude (P < 0.001), instrumental attitude (P = 0.004), subjective norm (P < 0.001), descriptive norm (P < 0.001), and self-efficacy (P < 0.001). The self-efficacy was the most important TPB construct in using seatbelts (Partial Eta-squared = 0.252). There was no significant difference among the three groups in terms of perceived control behavior (PBC) (P = 0.075).
Table 2.Mean (SD) of TPB Constructs Under Different Conditions of Using Seatbelt
| Constructs | Using Seatbelt for Law (N = 91), Mean (SD) | Using Seatbelt for Safety (N = 143), Mean (SD) | Not Using Seatbelt (N = 94), Mean (SD) | F | P Value | Partial Eta Squared |
|---|---|---|---|---|---|---|
| Intention | 22.71 (2.72) | 22.49 (2.57) | 18.70 (4.35) | 48.72 | < 0.001 | 0.231 |
| Experiential attitude | 10.53 (4.16) | 9.72 (3.81) | 4.81 (5.63) | 46.23 | < 0.001 | 0.221 |
| Instrumental attitude | 13.04 (7.33) | 10.49 (6.94) | 9.78 (7.02) | 5.57 | 0.004 | 0.033 |
| Subjective norm | -28.75 (15.68) | -21.08 (14.42) | -18.24 (16.22) | 11.82 | < 0.001 | 0.068 |
| Descriptive norm | -0.05 (6.30) | -4.19 (3.87) | -2.79 (4.51) | 20.41 | < 0.001 | 0.112 |
| PBC | 12.42 (6.71) | 11.84 (6.18) | 10.36 (6.39) | 2.61 | 0.075 | 0.016 |
| Self-efficacy | 31.44 (4.22) | 30.68 (4.29) | 24.77 (6.18) | 54.78 | < 0.001 | 0.252 |
Multiple regression analysis revealed that the experiential attitude (β = 0.30, t = 6.09, P < 0.001) and self-efficacy (β = 0.49, t = 9.49, P < 0.001) significantly predicted intention to use seatbelts (R2 = 0.55, F = 66.75, P < 0.001), indicating that one unit increase in experiential attitude and self-efficacy enhanced the intention score by 0.30 and 0.49, respectively.
The subjects who held more positive experiential attitudes and had higher perceived self-efficacy were more likely to have the intention to use seatbelts.
Three logistic regression models were used to assess the predictors of using seatbelts (Table 3). In the first model, the predictors of using seatbelts for enforcing the law versus non-using seatbelts (Group 1/Group 3) were examined. In this model, the constructs of experiential attitude (P < 0.001), self-efficacy (P = 0.001), and instrumental attitude (P = 0.006) had significant odds ratios. In other words, one unit increase in experiential attitude and self-efficacy enhances the likelihood of seatbelt use by 25% and 21%, respectively. Moreover, one unit increase in instrumental attitude decreases the likelihood of seatbelt use by %11 (Hosmer & Lemeshow: χ2 = 11.722, P = 0.164).
Table 3.Regression Coefficients in Logistic Regression Models Under Different Conditions of Using Seatbeltsa
| Model | Constructs | B | Exp (B) (95% CI) | Wald | P Value | R2 |
|---|---|---|---|---|---|---|
| Group 1 vs. group 3 | ||||||
| Block 0: Constant | -0.03 | 0.97 | 0.05 | 0.825 | ||
| Experiential attitude | 0.23 | 1.25 (1.11 - 1.42) | 13.28 | < 0.001 | 40% | |
| Instrumental attitude | -0.12 | 0.89 (0.81 - 0.97) | 7.41 | 0.006 | ||
| Self-efficacy | 0.19 | 1.21 (1.08 - 1.36) | 10.73 | 0.001 | ||
| Group 2 vs group 3 | ||||||
| Block 0: Constant | 0.42 | 1.52 | 9.98 | 0.002 | ||
| Intention | 0.23 | 1.26 (1.09 - 1.45) | 10.15 | 0.001 | 41% | |
| Experiential attitude | 0.23 | 1.25 (1.12 - 1.40) | 16.58 | < 0.001 | ||
| Instrumental attitude | -0.17 | 0.84 (0.78 - 0.92) | 14.35 | < 0.001 | ||
| Self-efficacy | 0.21 | 1.24 (1.11 - 1.38) | 14.98 | < 0.001 | ||
| Group 2 vs group 1 | ||||||
| Block 0: Constant | -0.45 | 0.64 | 11.36 | 0.001 | ||
| Descriptive norm | 0.15 | 1.16 (1.09 - 1.24) | 20.39 | < 0.001 | 16% |
aGroup 1: using seatbelt for law; Group 2: using seatbelt for safety; Group 3: not using seatbelt
In the second model, the predictors of using seatbelts for personal safety versus non-using seatbelts (Group 2/Group 3) were tested. In this model, the constructs of intention (P = 0.001), experiential attitude (P < 0.001), self-efficacy (P < 0.001), and instrumental attitude (P < 0.001) were the predictors of seatbelt use for personal safety. With one unit increase in intention, the likelihood of the seatbelt use for safety reasons would increase by %26. Moreover, one unit increase in experiential attitude and self-efficacy enhances the likelihood of seatbelt use for safety by 25% and 24%, respectively. In contrast, with a one unit increase in instrumental attitude, the likelihood of the seatbelt use would decrease by %16 (Hosmer & Lemeshow: χ2 = 12.20, P = 0.142).
In the third model, the predictors of using seatbelts for personal safety versus using seatbelts for enforcing laws (Group 2/Group 1) were tested. In this model, only descriptive norm (P < 0.001) had significant odds ratios as one unit increase in descriptive norm enhances the likelihood of seatbelt use for safety not enforcing laws by 16% (Hosmer & Lemeshow: χ2 = 11.89, P = 0.156).
5. Discussion
The findings of this study revealed that experiential attitudes and self-efficacy predict the intention to use seatbelts. Furthermore, instrumental attitudes also affected this variable. In individuals who used seatbelts for personal safety, the intention was the strongest predictor of behaviour, while those who used the seatbelts for enforcing laws did not intend to do so as their behaviour was under the influence of laws. Moreover, the findings showed that the descriptive norm and having a model were a construct making individuals use seatbelts for safety purposes.
Values and expectancy beliefs guide behaviours. In other words, individuals are more inclined to engage in health behaviour when they believe they can do it. In this regard, the findings of this study showed that self-efficacy predicts the intention and behaviour of using seatbelts. Moreover, in the other studies, self-efficacy was correlated with the behavior of using seatbelts (3, 19). This finding confirms with the present findings.
In several studies (3, 7, 13, 20, 21), the attitude was a predictor of the intention and behavior of seatbelts, so that a positive attitude would increase the intention and behavior of using seatbelts. This is consistent with the findings of this study.
In the present study, consistent with the findings of Simsekoglu’s and Lajunen study (7), the perceived behavioral control, which represents the effect of facilitators and inhibitors, did not predict the intention and behavior of using seatbelts. However, this finding is in contrast with the findings of some other studies (3, 13, 22). One of the reasons for this inconsistency is that all automobile companies are now obliged to equip vehicles with mechanical equipment facilitating the use of seatbelts, therefore, given the public access to these facilities, this construct does not predict the use of seatbelts.
Regarding the subjective norms, the results of the present study indicated that this construct does not predict the intention and behavior of using seatbelts. To justify this finding, it can be noted that in Iran, in urban trips, most of the front-seat occupants do not use seatbelts, In this regard, one study in Iran reported that only 34% to 44% of the front-seat occupants use seatbelts (8); therefore the important people such as parents and relatives do not recommend using seat belt. This finding contradicts with those of studies conducted on drivers (3, 7, 13, 23). This is due to the severe legal action against drivers that not using seat belts, therefore the legislator is the most important advocator for this behavior.
In individuals who use a seatbelt for safety, only a descriptive norm could predict this behavior. In other words, those who had a good practical model in their family for using seatbelts were more likely to wear seatbelts for safety purposes than those who used seatbelts for enforcing laws. One study showed that drivers whose friends and family members used seatbelts on roads, were more likely to use seatbelts (24). Moreover, in Cunill et al.’s study, individuals who did not use seatbelts reported that their friends and family members rarely used to wear seatbelts. This is in line with the findings of the present study (23).
In this study, the instrumental attitude, which consists of behavioral beliefs and outcome evaluation, has a negative relationship with using seatbelts, so that increasing instrumental attitudes reduces the behavior of using seatbelts. In this study, the participants reported negative consequences for using seatbelt, including stocked in a car during an accident, harass, harsh and annoying, limited movement, and so on. Accordingly, the present study revealed a negative relationship between the instrumental attitude and the behavior of using seatbelts. Since in the previous studies, instrumental attitudes were not examined, more studies are required in this regard. However, Gras et al. documented that some barriers to using seatbelts, including discomfort, boredom, and limited movement, do not allow the use of seatbelts (24). Furthermore, in Cunill’s study, the findings showed that most individuals who did not use seatbelts on urban roads believed that the seatbelts were not comfortable (23).
5.1. Conclusion
According to the findings of this study, individuals who use seatbelts for enforcing laws wear them with no intention. Therefore, since the intention is an important predictor of this behavior, it is believed that those who use seatbelts for safety purposes are more likely to continue the same behavior since their intention predicts the seatbelt use behavior. Moreover, since the descriptive norm, i.e., having a practical mode, is an important factor in using seatbelts for health purposes, families, especially parents, should be advised to use seatbelts in urban and rural travels so that their children would become permanent seatbelt users in the future. The findings of this study also showed that experimental and instrumental attitudes, i.e., individual perceptions about the comfort and safety of seatbelts, are important factors in using seatbelts; therefore, automobile companies should design seatbelts making passengers feel more comfortable and minimize their travel restrictions. With increasing safety, individuals’ attitudes (instrumental attitude) toward the negative consequences of using seatbelts would gradually be eliminated. In this regard, there are two implications to the finding of this study. First, in designing training programs, the emphasis should be on changing individuals’ perceptions of the negative consequences of using seatbelts, and individuals need to be more aware of their role models for their family members and children in using seatbelts. Second, automobile manufacturers need to focus on designing more comfortable and safer seatbelts.
This study had some limitations. First, given the cross-sectional design of the study, causal relationships are not decisive. Second, self-reports may lead to the overestimation of the concerned behavior and constructs. However, because of the anonymity of the questionnaires, these effects can be disregarded. Third, another limitation of this study was that the study was conducted only in one of the provinces in Iran (Bushehr province).
Finally, it is suggested to conduct further qualitative studies to extract other important components and add other constructs such as moral norms to this theory to increase its predictive power.
