Short communicationThe validity of self-reported seatbelt use in a country where levels of use are low
Introduction
Many studies have demonstrated the effectiveness of seat belts in reducing injury severity. Elvik and Vaa (2004), for example, estimated that use of seat belts reduces the probability of being killed by 40–50% for drivers and front seat passengers and by about 25% for the passengers sitting in back seats. World Health Organization (WHO, 2004) reported that injury reduction effects of seat belts in different types of car crashes varied from struck side with 27%, non-struck side with 39%, frontal with 43%, rear with 49%, and roll-over with 77%. These figures clearly underline the importance of seat belt usage for decreasing the number of injuries and fatalities everywhere. However, there are considerable differences among countries in seat belt usage rates.
Compared to the countries with high seat belt use rates such Western and Northern European countries and the U.S., for instance, seat belt use rates are considerably low in Turkey. In one observation study conducted by General Directorate of Police, 16% of the car drivers and 18% of the front seat passengers were using a seat belt on city roads in Ankara while the usage rate was much higher (71%) among car drivers on intercity roads (T.C. Emniyet Genel Müdürlüğü, 1999).
In general, national and local estimates of safety belt use are typically based on one of the following two assessment strategies: direct observations of seat belt use or self-report surveys. As compared to self-report surveys, it is well-known that direct observation provides a valid index of seat belt use rate. The observations can, on the other hand, be mostly conducted during day light hours (at least in bad lighting conditions outside urban areas); second, the observation of seat belt use among rear seat passengers is difficult; third, the vehicle type and weather conditions (rain, fog or snow) influences the measurement error in observations; and fourth, subjective assessment of driver's age and gender is prone to random and systematic errors.
Waller and Barry (1969) were the first ones who compared self-reported seat belt use with observed seat belt use among same participants. It was found that only 77% of drivers, who reported ‘always’ wearing seatbelts when driving in town, actually used a seatbelt when their behavior was observed. In other words, drivers over reported seatbelt use by 23 percent point. Fhaner and Hane (1973) examined the relationship between observed and self-reported seatbelt use in a sample of 105 Swedish drivers. Respondents, who reported using seat belts 60–100% of the time, over reported their seat belt use only by 14%. Stulginskas et al. (1985) assessed the validity of self reported seatbelt use in a sample of 106 Canadian car occupants and found over-reporting rate of 24% among the respondents. Streff and Wagenaar (1989) conducted one of the most comprehensive validity studies about seat belt observation and self reports. Roadside interviews were conducted among 42% (n = 1869) of the observed sample. Over reporting rate of the seat belt use was only 6% among participants who said they ‘always’ wore seatbelts. It should be noted, however, that in Streff and Wagenaar's (1989) study the participants knew that they had been observed.
Later, Dee (1998) reported a mismatch of 10% between observed use rates and self-reported use rates while Robertson (1992) reported a mismatch as high as 24% between observed use rates and self-reported use rates. Parada et al. (2001) found that Hispanic (n = 338) and white/non-Hispanic (n = 126) drivers over reported seat belt use by 27% and 21%, respectively. However, the correspondence between self-report and observation measures of seat belt use has remained mainly unexplored in the most parts of the world and in particularly in low belt use countries (e.g., Turkey). As low belt use countries are more likely to be targets of seatbelt use interventions, the validity of the self-report data is important for assessing the efficacy of large scale interventions among low belt use road users groups in low belt use countries.
The aim of the present study is, therefore, to compare the self-reported and observed seatbelt use in samples of drivers who were recruited from populations of two cities (Ankara and Afyon) with differently low rates of seat belt use in a low belt use country (Turkey).
Section snippets
Participants
Nine hundred and ninety drivers were randomly recruited from shopping centers, car parks, and suitable locations located in Afyon (n = 301) and Ankara (n = 658) cities of Turkey. Participants were part of a larger international study investigating factors influencing seatbelt and child seat usage and speeding (Middle East Technical University, Johns Hopkins International Injury Research Unit, Road Safety 10). Ninety percent of sample was male. Drivers from Afyon were older (mean = 38.4, SD = 11.8) than
Results
Participants from Afyon (n = 301) and Ankara (n = 658) reported 39% (n = 117) and 45% (n = 306) of ‘always’ seat-belt usage rate, respectively. No significant differences in the self-reported usage rate were found between drivers from Afyon and Ankara (39 vs. 45%, χ2 = 2.87, p = NS). However, a significant difference in the observed usage rate was found between Afyon and Ankara (19 vs. 32%, χ2 = 14.23, p < .01). Only 47% in Afyon and 70% in Ankara of those, who reported using ‘always’ a seat-belt, actually
Discussion
The results of the present study revealed that response bias within low belt use populations in low belt use countries may be much greater than suggested by observational data and data from high belt use countries. Over reporting was found between 53 and 30 percent change in two cities of a low belt use country. These rates are clearly higher than the over reporting figures ranging from 6 to 27 percent change in Western countries. The results are in line with Parada et al.’s (2001) suggestion
Acknowledgement
This study is supported by Bloomberg Philanthropies.
References (12)
Reconsidering the effects of seat belt laws and their enforcement status
Accident Analysis and Prevention
(1998)- et al.
Seat belts: the importance of situational factors
Accident Analysis and Prevention
(1973) Seat belt usage rates: a test of Peltzman's hypothesis
Accident Analysis and Prevention
(1986)- et al.
The validity of self-reported seatbelt use: Hispanic and non-Hispanic drivers in El Paso
Accident Analysis and Prevention
(2001) - et al.
Committing driving violations: an observational study comparing city, town and village
Journal of Safety Research
(2009) - et al.
Are there really shortcuts? Estimating seat belt use with self-report measures
Accident Analysis and Prevention
(1989)
Cited by (27)
The prevalence of seatbelt and mobile phone use among drivers in Riyadh, Saudi Arabia: An observational study
2018, Journal of Safety ResearchEvaluation of the five-year Bloomberg Philanthropies Global Road Safety Program in the Russian Federation
2017, Public HealthCitation Excerpt :Seat belt use among drivers and front seat passengers was constantly higher than rear seat passengers; this difference is in keeping with the existing literature20 and suggests that there is a need for more targeted interventions focused toward rear seat passengers. In line with the existing literature, the self-reported rate of seat belt use was higher than the observed rate in the initial rounds.21–25 The self-reported rate remained unchanged in Lipetskaya and increased by only 7 percentage points in Ivanovskya, whereas the observed seat belt wearing rate increased significantly in both the regions during the same period.
A comparison of observed and self-reported helmet use and associated factors among motorcyclists in Hyderabad city, India
2017, Public HealthCitation Excerpt :For instance, Bachani et al.28 found self-reported helmet use in Cambodia to be 1.5 times and 8.7 times higher than observed helmet use among motorcycle drivers and passengers, respectively. The difference between observed and self-reported use has been extensively studied for seatbelt use, suggesting that compared to self-reports, observational studies provide a more correct estimate of seatbelt use.23,32–36 Our study reaches a similar conclusion in the estimation of helmet use.
Evaluation of a five-year Bloomberg Global Road Safety Program in Turkey
2017, Public HealthSeatbelt wearing rates in middle income countries: A cross-country analysis
2014, Accident Analysis and PreventionCitation Excerpt :Previous literature has demonstrated the advantage of observational studies on seatbelt use to assess its prevalence (Bendak, 2005; Iribhogbe and Osime, 2008; Routley et al., 2007). However, little data is routinely collected on seatbelt use in low and middle income countries, and self-reported seatbelt use is generally considered to be a low-validity metric, especially in low-prevalence settings (Hyder et al., 2012; Ozkan et al., 2012). For this reason, observational studies on seatbelt use have been proposed for assessing prevalence at national and local level.
Predictors of intrinsic motivation behind seatbelt use in a country where current use is low
2013, InjuryCitation Excerpt :This study is based on data from the Bloomberg Philanthropies Global Road Safety Programme in Turkey.42–44 Roadside interviews for seatbelt use were conducted in October, similar to previous studies in Turkey.45 Drivers were approached and randomly recruited from car parks, shopping centers, and other suitable locations (i.e., places where researchers could observe drivers before they parked their vehicles) in the cities of Ankara and Afyon to participate in roadside interviews.