INSURANCE INSTITUTE FOR HIGHWAY SAFETY INSTITUTE FOR HIGHWAY SAFETY 1005 N. GLEBE RD., ARLINGTON, VA 22201 TEL. 703/247-1500 FAX 703/247-1678 October 10, 2002 The Honorable John L.

INSURANCE INSTITUTEFOR HIGHWAY SAFETY

1005 N. GLEBE RD., ARLINGTON, VA 22201 TEL. 703/247-1500 FAX 703/247-1678 www.highwaysafety.org

October 10, 2002

The Honorable John L. HenshawAssistant Secretary of Labor for Occupational Safety and HealthOccupational Safety and Health Administration U.S. Department of Labor200 Constitution Avenue N.W. Washington, D.C. 20210

Petition for Mandatory Safety Belt Use While Driving for Employment Purposes

Docket No. S-776

Dear Mr. Henshaw:

The Insurance Institute for Highway Safety (IIHS) hereby petitions theOccupational Safety and Health Administration (OSHA) to require the use of safety belts in motor vehicles while on the job, pursuant tosection 6(b)(1) of the Occupational Safety and Health Act. Motorvehicle crashes are the single largest known cause of occupationalfatalities and a leading cause of work-related disability (Bureau ofLabor Statistics (BLS), 1996, 2001). Research evidence leaves no doubt that the use of safety belts by vehicle occupants can greatlyreduce the risk of death and serious injuries in crashes (Dinh-Zarr etal., 2001). There also is evidence that many vehicle occupants on thejob do not use belts (National Institute for Occupational Safety andHealth (NIOSH), 1998); in fact, recent research indicates that belt use among people at work is lower than among other vehicle occupants(Eby et al., 2002; see attachment).

In July 1990, OSHA proposed a rule (55 FR 28728) requiring safety beltuse in motor vehicles, but it never has been finalized. OSHA shouldrapidly complete this rulemaking. The number of lives lost among motor vehicle occupants on the job has increased over time, from 1,158in 1992 to 1,363 in 2000 (BLS, 2001). OSHA should act now to protectthe nation’s workers against this well-recognized occupational hazard.

Why Require Belts? Motor vehicle injuries are a major source of occupational deaths anddisability. They are the leading cause of work-related injury deaths,accounting for nearly a quarter of known occupational deaths (BLS,1996, 2001). An estimated 51,000 occupational motor vehicle injuriesresulted in lost work time during 1994, with periods of disability farexceeding the median length of time lost for other types of workplaceinjuries (BLS, 1996).

John L. Henshaw October 10, 2002 Page 2 A wide spectrum of workers are victims of motor vehicle crashes (BLS, 1996; NIOSH, 1998). In an interview survey conducted by the National Highway Traffic Safety Administration (NHTSA), 34 percent of all drivers said they sometimes drove as part of their job or business, not including commuting (Block, 2000). During 1990-92, about 35 percent of people killed in motor vehicles at work were in large trucks, and most of the remainder were in light vehicles (NIOSH, 1998). Safety belts were not used by at least two-thirds of the motor vehicle occupants who were fatally injured while working (NIOSH, 1998). Safety belts can reduce the toll of occupational motor vehicle injuries. Requiring safety belt use would be an inexpensive and effective method of reducing occupational injuries in motor vehicles (Dinh-Zarr et al., 2001). NHTSA has estimated that using belts reduces the risk of death among front-seat occupants of passenger cars by about 45 percent and the risk of moderate to critical injury by 50 percent (NHTSA, 1997). The risk reductions are estimated to be about 60 to 65 percent among occupants of pickup trucks, which many people drive on the job. The requirement for belt use should apply to all motor vehicles for which NHTSA requires belts. In 1990, OSHA proposed excluding certain classes of vehicles such as those with open cabs or those operating near waterways. However, by preventing occupants from hitting their heads during crashes, safety belts diminish the chances that occupants will be rendered unconscious; this makes it more likely that they would be able to extricate themselves if overturns or submersions occur. Existing state belt use laws are insufficient to protect workers. State laws requiring belt use undoubtedly have increased use rates among people riding in vehicles at work; however, gaps or weaknesses in these laws necessitate additional measures to protect workers. The most serious weakness is that only 19 states permit police to ticket motorists solely for failure to use belts (hereafter referred to as primary enforcement states). Belt use rates in these states are about 11 percent higher than in states with secondary enforcement, which permits citations for not using belts only if police have stopped drivers for other reasons (NHTSA, 2002). Primary enforcement states have lower motor vehicle occupant death rates after they change from secondary enforcement (Dinh-Zarr et al., 2001). Other limitations of state belt laws is that they cover only people on public roads, typically do not apply to rear-seat occupants, and in some cases impose fines as small as $5 or $10 (IIHS, 2002). A number of vehicle-related occupational deaths occur in locations other than public roads (Windau et al., 1999).

John L. Henshaw October 10, 2002 Page 3 State laws are not leading to acceptable belt use rates among people riding in vehicles while working. New evidence indicates that people driving or riding in motor vehicles at work -- specifically occupants of light vehicles such as cars, pickup trucks, and vans/minivans -- are less likely to use belts than other vehicle occupants (Eby et al., 2002). Eby and colleagues conducted an observational survey in Michigan, finding that belt use among commercial light-vehicle drivers was significantly lower, only 56 percent, compared with 71 percent in light vehicles driven for personal purposes (Eby et al., 2002). The largest belt use differences were observed among passengers, whose use rates in commercial light vehicles were 42 percent compared with 69 percent in private light vehicles (Eby et al., 2002). Michigan has primary enforcement, and belt use among commercial drivers is almost certainly lower in states with secondary enforcement. The study by Eby et al. is particularly important because NIOSH research indicated that about 65 percent of people who died in vehicle crashes on the job were occupants of light vehicles (NIOSH, 1998). Other agencies and organizations have urged employers to require belt use on the job. A number of organizations, including the federal government, have recognized the benefits of employer involvement in increasing belt use among the nation’s workers. Federal employees have been required to use belts while working since 1986 (OSHA, 1990). Federal agencies (NIOSH and NHTSA) and other organizations including the National Safety Council (NSC), Network of Employers for Traffic Safety (NETS), and Drive Smart Virginia have urged employers to require belt use on the job and have published guidelines for employers on how to improve belt use rates (Drive Smart Virginia, 2002; NETS, 2002; NHTSA, 1997; NIOSH, 1998; NSC, 2002). An OSHA belt use requirement would increase belt use rates. Behavioral change occurs when laws and regulations are enacted. Marked upsurges in belt use occurred in states following passage of safety belt laws (Williams and Lund, 1988). Past experience indicates that rules or regulations are most effective when they are well publicized and when people believe they will be enforced. Visible and consistent enforcement, including application of sanctions, boosts belt use rates because it increases the perception that undesirable consequences will occur if a belt rule is violated (Dinh-Zarr et al., 2001; Williams et al., 2000). Other research also suggests that a requirement for belt use would increase the number of drivers using safety belts while working. According to self-reports, people whose job duties include driving are almost twice as likely to use safety belts on the job if their employers require it (Block, 2000). A number of government, nonprofit, and business employers who require belt use have succeeded

John L. Henshaw October 10, 2002 Page 4 in attaining observed use rates exceeding 85 percent; these employers typically had high-visibility activities, including enforcement, to promote belt use (NHTSA, 2001; Herbel, 2002). Currently, if OSHA inspects a work site and sees an employee not using a belt, the inspector has little recourse to address this dangerous work condition. If inspectors had the authority to cite employers for workers’ nonuse of belts and could determine whether an employer had a reasonable plan to publicize and enforce the belt use requirement, then both employers and employees would have strong incentives to adhere to belt use rules. Employer Obligations Can Be Economically Feasible Requiring belts on the job would be one of the most inexpensive and cost-effective safety rules OSHA ever issued. Many employers would realize cost savings, as each occupational crash involving an injury cost employers an average of $24,000 in 1994, and medical treatment costs rise dramatically when occupants are unbelted (NHTSA and NETS, 1998). Belts already are installed in motor vehicles so, unlike the equipment required by other occupational health and safety regulations, employers would not have to pay for any new equipment. A concern voiced by some employers in response to OSHA’s 1990 proposal was the cost of training and enforcement programs. Clearly, employers must inform their employees about a belt use rule and their intentions to enforce it; however, training programs need not be part of the rule (Vernick et al., 1999). Most motorists recognize that belts are an effective way to reduce crash injuries, so more education concerning belts would have limited benefits (Cammisa et al., 2000; Williams et al., 1997). New employees should not be permitted to operate vehicles on the job until they have been informed of the rule, and existing employees should be reminded of the rule at least once a year. Based on extensive research regarding belt use and other traffic laws (Williams et al., 2000), the key determinant of compliance will be perceptions that an employer is serious about employees adhering to the rule. Workplaces with high belt use rates, such as military installations, increased use rates through a combination of observational surveys, vigorous enforcement, and incentive programs (NHTSA, 2001; Herbel, 2002). Although similar programs may not be feasible for small employers, all employers could take some steps to enforce the rule in their own work settings; for example, supervisors could correct any observed nonuse of belts and apply penalties when appropriate. Accordingly, OSHA should require employers to enforce the belt use rule without requiring specific enforcement techniques that would be inappropriate for small employers. Instead, OSHA should point to the belt use promotion and enforcement programs used by

John L. Henshaw October 10, 2002 Page 5 workplaces with high belt use rates and provide suggestions for employers on how to monitor compliance. Conclusion The Institute urges OSHA to complete its occupant protection rulemaking, which began in 1990, and mandate that employers require the use of available safety belts by employees who ride in motor vehicles as part of their jobs. Such a rule can be implemented without requiring employers to conduct expensive driver training programs or infeasible enforcement programs. This rulemaking has languished far too long -- 12 years -- and a belt use requirement should be adopted quickly to reduce deaths and injuries among workers on U.S. highways.

Sincerely, Brian O’Neill President

cc: Docket No. S-776 Attachment: Eby et al., 2002 References

Block, A.W. 2000. 1998 motor vehicle occupant safety survey: vol. 2, seat belt report. Washington, DC: National Highway Traffic Safety Administration. Available: http://www.nhtsa.dot.gov/people/injury/ research/safetysurvey/.

Bureau of Labor Statistics. 1996. Deadly highway accidents outnumber other fatal work incidents. Washington, DC: U.S. Department of Labor. Available: http://www.bls.gov/opub/ils/pdf/opbils07.pdf.

Bureau of Labor Statistics. 2001. National census of fatal occupational injuries in 2000. Washington, DC: U.S. Department of Labor. Available: http://www.bls.gov/iif/oshwc/cfoi/cfnr0007.pdf.

Cammisa, M.X.; Williams, A.F.; and Ferguson, S.F. 2000. Self-reported seat belt use in four countries: a telephone survey. Journal of Crash Prevention and Injury Control 2:103-10.

John L. Henshaw October 10, 2002 Page 6 Dinh-Zarr, T.B.; Sleet, D.A.; Shults, R.A.; Zaza, S.; Elder, R.W.; Nichols, J.J.; Thompson, R.S.; Sosin, D.M.; and the Task Force on Community Preventive Services. 2001. Reviews of evidence regarding interventions to increase the use of safety belts. American Journal of Preventive Medicine 21:4(suppl.1):48-65. Available: http://www.meddevel.com/site.mash?left=/library.exe&m1=3&m2=2&right=/library.exe&action=vil&source=now&node=TOC@@AMEPRE&site=AJPM&jcode=AMEPRE.

Drive Smart Virginia. 2002. Employer Incentive Program. Richmond, VA. Available: http://www.drivesmartva.com/navigate.htm.

Eby, D.W.; Fordyce, T.A.; and Vivoda, J.M. 2002. A comparison of safety belt use between commercial and noncommercial light-vehicle occupants. Accident Analysis and Prevention 34:285-91.

Herbel, S.B. 2002. Protecting workers from motor vehicle injury and death: an evaluation of incentive-based motivational techniques for increasing employee safety belt use. Unpublished report submitted to Drive Smart Virginia.

Insurance Institute for Highway Safety. 2002. Child restraint, belt use laws. Arlington, VA. Available: http://www.highwaysafety.org/ safety_facts/state_laws/restrain.htm.

National Highway Traffic Safety Administration. 1997. Initiative for Increasing Seat Belt Use Nationwide. Washington, DC: U.S. Department of Transportation. Available: http://www.nhtsa.dot.gov/people/injury/ airbags/presbelt/index.html.

National Highway Traffic Safety Administration. 2001. Process and outcome evaluation of the Buckle Up America initiatives. Report no DOT-HS-809-272. Washington, DC: U.S. Department of Transportation. Available: http://www.nhtsa.dot.gov/people/injury/research/BuckleUp/.

National Highway Traffic Safety Administration. 2002. Safety belt and helmet use in 2002: overall results. Report no. DOT-HS-809-500. Washington, DC: U.S. Department of Transportation. Available: http://www-nrd.nhtsa.dot.gov/pdf/nrd-30/NCSA/Rpts/2002/809-500.pdf.

National Highway Traffic Safety Administration and the Network of Employers for Traffic Safety. 1998. What do traffic crashes cost employers? Washington, DC: U.S. Department of Transportation. Available: http://www.nhtsa.dot.gov/people/outreach/employer/ WhatCost/toc.html.

National Institute for Occupational Safety and Health. 1998. NIOSH alert: preventing worker injuries and deaths from traffic-related

John L. Henshaw October 10, 2002 Page 7 motor vehicle crashes. DHHS (NIOSH) Publication no. 98-142. Atlanta, GA: Centers for Disease Control and Prevention. Available: http://www.cdc.gov/niosh/mvalert.html.

National Safety Council. 2002. Buckle-Up Safety Awareness Kit. Itasca, IL. Available: http://www.nsc.org/issues/drivsafe.htm.

Network of Employers for Traffic Safety. 2002. NETS urges employers to establish corporate safety belt policies. Washington, DC. Available: http://www.trafficsafety.org/news/52302.html.

Occupational Safety and Health Administration. 1990. Notice of Proposed Rulemaking: Occupant Protection in Motor Vehicles, 55 FR 28728. Washington, DC: U.S. Department of Labor.

Vernick, J.S.; Li, G.; Ogaitis, S.; MacKenzie, E.J.; Baker, S.P.; and Gielen, A.C. 1999. Effects of high school driver education on motor vehicle crashes, violations, and licensure. American Journal of Preventive Medicine 16:1(suppl.1):40-46. Available: http://www.meddevel.com/site.mash?left=/library.exe&m1=3&m2=2&right=/library.exe&action=vil&source=now&node=TOC@@AMEPRE&site=AJPM&jcode=AMEPRE.

Williams, A.F. and Lund, A.K. 1988. Mandatory seat belt use laws and occupant crash protection in the United States: present status and future prospects. Preventing Automobile Injury (ed. J.D. Graham), ch. 3, 51-72. Dover, MA: Auburn House Publishing Company.

Williams, A.F.; Wells, J.K.; McCartt, A.T.; and Preusser, D.F. 2000. “Buckle up NOW!”: an enforcement program to achieve high belt use. Journal of Safety Research 31:195-201.

Williams, A.F.; Wells, J.K.; and Reinfurt, D.F. 1997. Characteristics and opinions of North Carolina residents cited for not using seat belts. Journal of Traffic Medicine 25:71-76.

Windau, J.; Sygnatur, E.; and Toscano, G. 1999. Profile of work injuries incurred by young workers. Monthly Labor Review 122:3-10. Washington, DC: Bureau of Labor Statistics, U.S. Department of Labor. Available: http://www.bls.gov/opub/mlr/1999/06/art1abs.htm.

ACCIDENTANALYSIS&PREVENTION

PERGAMON Atx:ident Analysis and Prevention 34 (2002) 285-291

www.elsevier.com/locate/aap

A comparison of safety belt use between commercial andnoncommercial light-vehicle occupants

David W. Eby., TifTani A. Fordyce, Jonathon M. VivodaTrQlUportatioll Research l/lStitute, Ulliversity 01 Michigan, UMTRI-SBA, 2901 Baxter Road, AM Arbor, MI 48109-2/~ USA

Received 7 August 2{XM}; ~vcd in revised fonD 26 January 2001; accepted 28 February 2001

Abstract

The purpose of this study was to conduct an observational survey of safety belt use to determine the use rate of commercialversus noncommercial light-vehicle occupants. Observations were conducted on front-outboard vehicle occupants in eligiblecommercial and noncommercial vehicles in Michigan (i.e., passenger cars, vans/minivans, sport-utility vehicles, and pickuptrucks). Commercial vehicles that did not fit into one of the four vehicle type categories, such as tractor-trailers, buses, or heavytrucks, were not included in the survey. The study found that the restraint use rate for commercial light-vehicle occupants was55.8% statewide. The statewide safety belt use rate for commercial light-vehicles was significantly lower than the rate of 71.2% fornoncommercial light-vehicles. The safety belt use rate for commercial vehicles was also significantly different as a function ofregion, vehicle type, seating position, age group, and road type. The results provide important preliminary data about safety beltuse in commercial versus noncommercial light-vehicles and indicate that further effort is needed to promote safety belt use in thecommercial light-vehicle occupant population. The study also suggests that additional research is required in order to developeffective programs that address low safety belt use in the commercial light-vehicle occupant population. 0 2002 Elsevier ScienceLtd. A11 rights reserved.

Keywords: Safety belt; Commercial light-vehicle; Observational survey; Otx:upational safety

I. Introduction deaths from traffic-related motor vehicle crashes(NIOSH, 1998). The alert recommended that workersalways use safety belts and encouraged employers toensure that vehicles have adequate restraint systems forall occupants. Other recommendations included devel-oping company policies to require use of safety belts.

Recent data suggest that company policy requiringsafety belt use is not common. According to a recentUS motor vehicle occupant safety survey conducted atthe end of 1998,34% of respondents drove while on thejob. Only about 48% of these respondents thought thattheir company had a policy requiring safety belt use(Block, 2~). Of these people, two-thirds thought theircompany had a written policy and one-half thought thepolicy was enforced very strictly. This same survey alsofound that, at least by self-report, having a companypolicy raised safety belt use. Eighty percent of thosewho reported that their company had a policy requiringthe use of safety belts indicated that they wore theirbelts 'aU of the time' compared to 7~/o of respondents

In 1998, motor vehicle crashes were the leading causeof death in the workplace, accounting for about 24% ofUS job-related fatalities (Bureau of Labor Statistics,2<XX». A National Institute for Occupational Safety andHealth (NIOSH, 1998) analysis of fatalities in work-re-lated motor vehicle crashes for the period of 1990-1992found that 93% of the fatally injured were male, 700/0were aged 25-54 years, and 76% were the driver. Forincidents in which restraint use information was avail-able only about 33% of the workers were using occu-pant restraints. These data suggest that increased safetybelt use in commercial vehicles may significantly reducethe number of fatalities among workers who drive whileon their job. Following this line of reasoning, NIOSHreleased an alert for preventing worker injuries and

. Co~lponding author. Tel.: + 1-734-7632466; fax:9361076.

E-mail address: [email protected] (D.W. Eby).

+ 1-734-

~1-4S7S/02/S - see front matter 0 2002 Ellevier Scien«. Ltd. All rights reservedPII: SOOOI-4S7S(OI)OOO24-0

286 D. W. Eby ,t aI. / Accident Analysis and Prel¥fttion 34 (2tKJ2) 285-291

whose company did not have a policy. These resultssuggest that safety belt use in commercial vehiclesmight be high, perhaps even higher than in noncom-mercial vehicles.

On the other hand, some states, such as Michigan,allow certain types of commercial vehicles an exemp-tion from using safety belts. According to MichiganPublic Act 300, .. A commercial or United States postal

service vehicle that makes frequent stops for the pur-pose of pickup or delivery of goods or services" isexempt from mandatory safety belt use. Thus, thisexemption may result in lower safety belt use for com-mercial light-vehicle occupants when compared withnoncommercial light-vehicle occupants.

The purpose of the present study was to determinethe statewide safety belt use rate of occupants travelingin commercial versus noncommercial passenger cars,sport-utility vehicles, vans/minivans, and pickup trucksthrough a direct observation survey. No direct observa-tion study of commercial light-vehicle safety belt usehas ever been conducted. As discussed by several re-searchers, direct observational surveys of safety belt useon roadways provides a less biased source of belt useinformation than self-report surveys or police-reportedsafety belt use on crash reports (Li et aI., 1999; Eby etal., 2001), although these other sources of restraint usedata have important advantages.

2. Method

2.1. Sample design

The sample design for the present survey was closelybased upon the one used by Streff et aI. (1993). Theobjective of the sample design was to select observationsites that represent accurately front-outboard vehicleoccupants in eligible commercial and noncommerciallight-vehicles in Michigan (i.e., passenger cars, vans/minivans, sport-utility vehicles, and pickup trucks),while following federal guidelines for safety belt surveydesign (National Highway Traffic Safety Administra-tion, NHTSA, 1992, 1998). Commercial heavy-vehiclesthat did not fit into one of the four vehicle typecategories, such as tractor-trailers, buses, or heavytrucks, were excluded from the survey. Although veryimportant for understanding commercial vehicle occu-pant protection use, there are no comparable heavyvehicles in these other categories that are used fornoncommercial purposes.

All 83 Michigan counties were rank ordered by pop-ulation (United States Bureau of the Census, 1992) andthe low population counties were eliminated from thesample space. This step reduced the sample space to the28 counties in which 85% of Michigan's populationreside. The 28 counties were separated into four strata

(regions) based on historical belt use rates and vehiclemiles of travel (VMT) for each county. Wayne Countywas chosen as a separate stratum because of the dispro-portionately high VMT for Wayne County and becausewe wanted to ensure that observation sites were selectedwithin this county. Three other strata were constructedby rank ordering each county by historical belt userates and then adjusting the stratum boundaries untilthe total VMT was roughly equal within each stratum.The stratum boundaries were high belt use (greaterthan 54.00/0), medium belt use (45.0-53.00/0), low beltuse (44.9% or lower), and Wayne County (41.CJO/o beltuse).

To achieve the NHTSA required precision of lessthan 5% relative error, the minimum number of obser-vation sites for the survey (N = 56) was determinedbased on within- and between-county variances fromprevious belt use surveys and an estimated 50 vehiclesper observation period in the current survey. This min-imum number was then increased (N= 168) to get anadequate representation of belt use for each day of theweek and for all daylight hours.

Because total VMT within each stratum was roughlyequal, observation sites were evenly divided among thestrata (42 each). In addition, since an estimated 23% ofall traffic in Michigan occurs on limited-access road-ways (Federal Highway Administration, 1982), 10(24%) of the sites within each stratum were freeway exitramps, while the remaining 32 were roadway intersec-tions. Both the intersection and freeway exit ramp siteswere chosen using a method that ensured each intersec-tion and exit ramp an equal probability of selection.Direction of traffic flow to be observed and observerlocation at each site was randomly assigned.

The day of week and time of day for site observationwere quasirandomly assigned to sites in such a way thatall days of the week. and all daylight hours (7:00a.m.- 7:00 p.m.) had essentially equal probability ofselection. The sites were observed using a clusteringprocedure. That is, sites that were located spatiallyadjacent to each other were considered to be a cluster.Within each cluster, a shortest route between all of thesites was determined. An observer watched traffic at allsites in the cluster during a single day. The day in whichthe cluster was to be observed was randomly deter-mined. After taking into consideration the time re-quired to finish all sites before darkness, a randomstarting time for the day and site within the cluster wasselected. The observer visited sites following the loop ineither a clockwise or counterclockwise direction(whichever direction left them closest to the Universityat the end of the day).

The sample design was constructed so that eachobservation site was self-weighted by VMT within eachstratum. This was accomplished by selecting sites withequal probability and by setting the observation inter-

D. W. £by et Gl./ Accident Analysis and Pr'e.-tion 34 (21»2) 285-291 287

val to a constant duration (50 min) for each site. Thusthe number of vehicles observed at an observation sitereflected safety belt use by VMT; that is, the higher theVMT at a site, the greater the number of vehicles thatwould pass during the 50-min observation period.However, since all vehicles passing an observer couldnot be surveyed, observers conducted a 5-min count ofall eligible vehicles on the designated traffic leg beforebeginning safety belt observations. At the end of theobservation period, a second 5-min vehicle count wasconducted.

observed during the study. Observers worked in teamsof two, observing the same vehicles, but recording dataindependently on separate data collection forms. Teamswere rotated throughout the training to ensure thateach observer was paired with every other observer atleast eight times. Each observer pair practiced recordingsafety belt use and other data until there was aninterobserver reliability of at least 85% for all measureson drivers and front-right passengers for each pair ofobservers. An interobserver reliability of 85% was alsoobtained for vehicle type, and whether the light-vehiclewas commercial or noncommercial. During data collec-tion, unannounced site checks were conducted on atleast three occasions for each observer.

2.2. Data collection

3. Results

Use rates were calculated by determining weightedsafety belt use rates separately for each stratum (re-gion). The four use rates were then combined to deter-mine statewide use rates. The equations used forcalculating use rates, variances, and 95% confidencebands are shown in Appendix A. As shown in Table I,of the 9414 vehicle occupants observed, about 6% wereriding in eligible commercial light-vehicles. The userates for these occupants were calculated separatelyfrom the use rates for occupants traveling in eligiblenoncommercial light-vehicles. We present the results asa comparison between commercial and noncommercialoccupants of passenger cars, sport-utility vehicles, vans/minivans, and pickup trucks. Rates that are presentedas significantly different have confidence bands that donot overlap.

Table I shows the statewide use rates, confidencebands, and unweighted number of occupants for com-mercial and noncommercial light-vehicle occupants.The overall use rate of 55.8% for commcrciallight-vehi-cle occupants was significantly lower than the rate fornoncommercial light-vehicle occupants.

Also shown in Table 1 are the use rates for commer-cial and noncommercial light-vehicle occupants by re-gion. For occupants in both commercial andnoncommercial light-vehicles, safety belt use rates fol-lowed historical statewide trends (Eby et aI., 1999), withthe Wayne County region having the lowest use rates.Comparison between commercial and noncommerciallight-vehicle occupants showed that belt use was signifi-cantly lower for commercial light-vehicles for only themedium and low belt use regions in Michigan. Eventhough the commercial light-vehicle use rates for theother two regions were quite low relative to noncom-mercial light-vehicles, the confidence bands for thecommercial light-vehicle occupant use rates were verylarge. Perhaps with a larger sample of commerciallight-vehicle occupants, these rates might also differstatistically from the noncommercial rates.

Data collection for the study involved direct observa-tion of shoulder belt use, estimated age, and sex of thedriver and front-right passenger. Observers were in-structed to observe only vehicles in the lane immedi-ately adjacent to the curb for safety belt use regardlessof the number of lanes present. The vehicle type andwhether or not there were visual indications that thelight-vehicle was potentially being used for commercialpurposes was also recorded. For the purposes of thisstudy, a commercial vehicle was defined as a light-vehi-cle that was sometimes used for business purposes asindicated by the presence of company logos, commer-cial lettering, government/service license plates, orladders/tools.

Trained field staff observed shoulder belt use ofdrivers and front-right passengers traveling in passengercars, sport-utility vehicles, vans/minivans, and pickuptrucks for both commercial and noncommercial light-vehicles during daylight hours from September 2 toOctober 8, 1999. Observations were conducted when avehicle came to a stop at a traffic light or a stop sign.

All sites in the sample were visited by single observersfor a period of 1 h, with the exception of sites in andadjacent to the city of Detroit. In order to addresspotential security concerns, Detroit sites were visited bytwo-person teaIns of observers for a period of 30 min.Each observer conducted a 5-min traffic count prior tothe 25-min observation period. Because each teammember at Detroit sites recorded data for differentlanes of traffic, the total amount of time at Detroit siteswas equivalent to that at other sites.

Prior to data collection, field observers participatedin 5 days of intensive training including both classroomreview of data collection procedures and practice fieldobservations. Each observer received a training manualcontaining detailed information on field procedures forobservations, data collection forms, and administrativepolicies and procedures. After intensive review of themanual, observers conducted practice observations atseveral sites chosen to represent the types of sites andsituations that would actually be encountered in thefield. None of these practice sites were the same as sites

288 D. W. £by et aI.1 Accident Allalysis and PrePeIItion 34 (2002) 285-291

Safety belt use rates and unweighted Ns by sex areshown in Table 1. Males were much more commonlyfound in the commercial light-vehicle occupant samplethan the noncommercial sample. Males accounted for

Table IEstimated statewide occupant restraint use rates, 95% confidencebands, and unweighted Ns for oommen:ial and noncommercial light-vehicle occupants

State of SS.8 .t: 9.8 S89Michigan"

Michigan regionHigh belt 63.4.t: 12.0 190

useMedium belt 5S.2.t: 9.8 132

use"Low belt 54.9.t: 1.0 90

use"Wayne 48.8.t: 16.8 177

County

Se.xMale 54.7.t: 6.1 S34Female 68.0.t: 10.2 54

Vehicle typePassenaer 5S.4.t: 14.1 53

carSport-utility S9.3.t:27.9 17Van{ 63.7.t:6.3 194

minivan"Pickup truck 48.6.t: 7.0 325

Seating positionDriver- 56.2 .t: 8.6 493Front-right 42.1 .t: 10.1 96

paIS"

Age group16-29 48.9.t:11.8 1153O-S9" 56.9 .t: 7.8 443~up 79.8.t: 11.8 24

Road typeInte~ion" 56.9.t: S.9 39SFreeway exit 6O.6.t: 5.2 194

ramp"

Days of weekMonday" 47.2.t: 6.7 71Tuesday" 58.0 .t: 3.0 107Wednesday" 56.8 .t: 8.8 4SThursday" 58.3.t: 3.2 138Friday" 51.0.t: 8.8 142Saturday" 44.0 .t: 2.1 70Sunday" 65.9 .t: 4.3 16

71.2 * 2.2 ~

7'.7.t: ,., 2249

73.0 t; 4.5 1588

61.2 t; 3.' 1735

67.3:f: 3.6 32.13

65.1.i: 1.778.6.i: 1.9

46494171

74.9 f: 9.2 SO77

7O.7:t; 3.175.5 .i 3.6

9881344

55.3 :t; 6.2 1416

71.9%2.468.6 % 3.6

69531872

58.1 i: 2.114.8 i: 1.177.0 i: 3.4

2058S4451030

~32742

70.8 i: 1.671.6 i: l.oS

64..5 :t 2.,73.':t 1.074.1 %: 1.669.8 :t 2.373..5 :t 1.273.7:t 0.777.1 :t 0.7

15931310631

161917081147817

over 9()O/0 of the commercial light-vehicle occupant sam-ple whereas males represented only about 53% of thenoncommercial light-vehicle occupants. Use ratesfor both commercial and noncommercial light-vehicle occupants showed that females used safetybelts at a higher rate than males, in agreement withfindings conducted in Michigan (Eby et al., 2000) andnationally (NHTSA, 1998). However, comparisons be-tween commercial and noncommercial light-vehicle userates by sex showed that they did not differ signifi-cantly.

The analysis of use rates by the type of vehicle isshown in Table I. As found in several other statewidesurveys in Michigan, safety belt use was significantlylower for pickup truck occupants than occupants inother light-vehicle types (Eby et aI., 1999). Consideringonly commercial vehicles, we found that more thanone-half of occupants were in pickup trucks. In non-commercial vehicles, on the other hand, nearly 600/0 ofoccupants were traveling in passenger cars. Althoughbelt use was lower for commercial light-vehicle occu-pants in every vehicle type studied, there were nosignificant differences in safety belt use rates for any ofthe vehicle types except vans and minivans, wherecommercial light-vehicle occupants used safety belts ata lower rate than occupants in noncommercial vans andminivans.

As shown in Table 1, safety belt use was higher forthe driver than for the passenger for both commercialand noncommercial light-vehicle occupants. Higher beltuse for drivers than for passengers is a common findingin the literature (Williams et aI., 1987; Williams andWells, 1992; NHTSA, 1997; Eby et aI., 2«XX». In bothcommercial and noncommercial light-vehicles, front-right passengers were present about one-third of thetime. Comparing the belt use rates between commercialand noncommercial light-vehicle occupants by seatingposition, we found that use was significantly lower incommercial light-vehicles for both the driver andpassenger.

As shown in Table 1, belt use increased with age forboth commercial and noncommercial light-vehicle oc-cupants. Increasing belt use with age is a well estab-lished trend in Michigan (Eby et aI., 2«XX» and in otherstates (Williams et al., 1987; Ulmer et aI., 1994). Use ofsafety belts differed between commercial and noncom-merciallight-vehicle occupants for only the 30-59-year-old age group, with use significantly lower forcommercial light-vehicle occupants.

Safety belt use was also analyzed as a function of thetype of roadway (Table I). For both commercial andnoncommercial light-vehicles, belt use was higher fortraffic exiting freeways than for traffic on local road-ways. Again, this finding is consistent with several otherdirect-observation studies of safety belt use (Fockler

D. W. Eby et oJ. / Accident Analysis and Prevention 34 (2002) 285-291 289

and Cooper, 1990; Chatterjee et al., 1991; Eby et al.,2000). Safety belt use for both types of roadways wassignificantly lower for commercial light-vehicleoccupants.

Finally, safety belt use was analyzed by day of weekand is shown in Table I. There were no consistenttrends in use by day of week for either commercial ornoncommercial light-vehicle occupants except that beltuse was highest on Sunday. Comparison between com-mercial and noncommercial light-vehicles showed thatbelt use was significantly lower for commerciallight-ve-hicle occupants each day of the week.

4. Discussion

These findings suggest that traveling in a commerciallight-vehicle results in lowered safety belt use. It is alsopossible that some demographic data not collected inthis study but related to low safety belt use, such aseducational level (Wagenaar et al., 1987) or level ofsensation seeking (Clement and Jonah, 1984; Jonah etal., 1997), may account for the lower rates for commer-cial vehicle occupants in eligible vehicle types.

Although belt use was lower for commercialvehicle occupants in every vehicle type studied, the onlysignificant difference in safety belt use rates for any ofthe vehicle types was found in vans and minivans.Occupants of commercial vans and minivans usedsafety belts at a lower rate than occupants in non-commercial vans and minivans. A possible explanationfor this finding may be the differing demographiccharacteristics of van/minivan occupants dependingupon whether the vehicle is used for commercialpurposes.

These results suggest that further efforts are neededto increase safety belt use in commercial light-vehicles.Since drivers are about twice as likely to wear a safetybelt while on the job if their company has a policyrequiring belt use (Block, 2000), efforts should be madeto encourage companies to adopt a mandatory safetybelt use policy. Companies should also be encouragedto enforce and strengthen existing safety belt use poli-cies. It may, however, be difficult to reach those seg-ments of the commercial light-vehicle drivingpopulation that are least likely to use safety belts.Anecdotally, we found that the large majority of com-mercial nonusers were in construction-related busi-nesses. The majority of construction companies inMichigan are small, privately owned businesses, whichwould make them difficult to reach. In addition, it ispossible that a 'culture' not conducive to safety belt usemay exist in this population.

This study provides some important preliminary dataabout safety belt use in commercial versus noncommer-cial light-vehicles through direct observation of use.Even with the low number of commercial vehiclesfound in the statewide sample, the significantly lowerbelt use found for occupants of these vehicles warrantsattention. Further research is needed to understandwhy belt use is low in the commercial light-vehicleoccupant population so that effective programs can bedeveloped.

The study was designed to determine safety belt userate trends in Michigan for commercial light-vehicleoccupants and to compare these rates to noncommer-cial light-vehicle occupant rates collected at the sametimes and places. Only passenger cars, sport-utilityvehicles, vans/minivans, and pickup trucks were in-cluded in the study. Thus, the study results do notinclude safety belt use information for occupants inheavy vehicle types such as large trucks, tractor-trailers,buses, or recreational vehicles. Statistically comparingcommercial and noncommercial light-vehicle occupantsafety belt use revealed that commercial light-vehicleoccupant safety belt use was significantly lower thanbelt use for occupants in noncommercial light-vehicles.The 55.8% use rate shows that a significant portion ofMichigan's commercial light-vehicle occupants are notusing safety belts. One possibility for this low rate ofsafety belt use in commercial light-vehicles is that themandatory safety belt use law allows certain commer-cial vehicle occupants an exemption, if the vehiclemakes frequent stops. While the business of the com-mercial vehicle was not collected in this study, anecdo-tal reports from the observers indicated thatonly a small percentage of commercial vehicles ob-served were pickup/delivery or postal service vehicles.According to our observers, around three-fourths ofcommercial light-vehicles were in construction-relatedbusinesses.

Safety belt use rates for commercial light-vehiclesvaried by Michigan region, occupant sex, vehicle type,seating position, age group, road type, and day of weekin the same ways that it varies for noncommerciallight-vehicles in Michigan, both in the current studyand historically (Eby et al., 2000). In fact, the onlydifference between commercial and noncommercial oc-cupant safety belt use rates was that use was signifi-cantly lower for commercial light-vehicle occupants inmost cases; in those cases where there was no signifi-cant difference between rates, low sample sizes proba-bly accounted for the lack of statistical difference.

Acknowledgements

This project was sponsored by the Michigan Office ofHighway Safety Planning (OHSP) through contractnumber OP-OO-OI. Hans Joksch assisted in the sam-pling design and data analysis. Lisa Molnar and LidiaKostyniuk provided valuable comments on an earlier

29() D.W. Eby et at. / Accident Analysis and Pre~tion 34 (2002) 285-291

draft of this manuscript. The opinions, findings, andconclusions expressed here are those of the authors andnot necessarily those of OHSP.

Cochran (1977, Sectionshown in Eq. (3):

1.8). The resulting formula is

Appendix A

(4)

The Wayne County stratum (number 4) variance wasmultiplied by 0.88 to account for the similar weightingthat was done to estimate overall belt use. The 95%confidence bands were calculated using Eq. (5):

95% confidence band = 'all:!: 1.96 x JV~ce, (5)

where, is the belt use of interest. This formula is usedfor the calculation of confidence bands for each stratumand for the overall belt use estimate.

This safety belt survey was designed to estimate beltuse for the state of Michigan based on VMT. Theself-weighting-by-VMT scheme employed in the designis limited by the number of vehicles for which anobserver can accurately record infonnation. To correctfor this limitation, the vehicle count infonnation wasused to weight the observed traffic volumes so theywould more accurately reflect VMT.

This weighting was done by first adding each of thetwo 5-min counts and then multiplying this number by5 so that it would represent a 50-min duration. Theresulting number was the estimated number of vehiclespassing the site if all eligible vehicles had been includedin the survey during the observation period at that site.The estimated count then was divided by the actualvehicle count for each vehicle type to obtain a VMTweighting factor for that site and vehicle type. Thisweighting factor was multiplied by the actual vehiclecounts at the site, yielding a weighted N for the numberof total drivers and passengers and total number ofbelted drivers and belted passengers for each vehicletype.

The overall estimate of belt use per VMT in Michi-gan was detennined by first calculating the belt use ratewithin each stratum for observed vehicle occupants inall vehicle types using Eq. (1):

r, = Total number of belted occupants, weighted (1). Total number of occupants, weighted' ,-,

where'i refers to the belt use rate within any of the fourstrata. The totals are the sums across all 42 sites withinthe stratum after weighting, and occupants refers toonly front-outboard occupants. The overall estimate ofbelt use was computed by averaging the belt use ratesfor each stratum. However, comparing total VMTamong the strata, one finds that the Wayne Countystratum is only 88% as large as the total VMT for theother three strata. In order to accurately representsafety belt use for Michigan by VMT, the WayneCounty stratum was multiplied by 0.88 during theaveraging to correct its lower total VMT. The overallbelt use rates were determined using Eq. (2):

" +'2 +'3 + (0:88'4) f?)

References

raIl = 3.88' '-J

where, is the belt use rate for a certain vehicle typewithin a stratum and '4 the Wayne County stratum.

The variances for the belt use estimates were calcu-lated using an equation derived from Eq. (11.30) of

Block, A., 2000. 1998 Motor Vehicle Occupant Safety Survey, vol. 2.US Department of Transportation, Washington, DC.

Bureau of Labor Statistics, 2000. Census of Fatal OccupationalInjuries Summary 1998. (http:ffstats.bls.govfnews.releasefcfoi.nws.htm). Accessed on May 4, 2000.

Chatterjee, A., Evans, G., Richards, S.H., Hafford, H., 1991. Safetybelt use in Tennessee. ITE Journal January, 61, 23-27.

Clement, R., Jonah, B.A., 1984. Field dependence, sensation seekingand driving behaviour. Personality and Individual Differences 5,87 -93.

Cochran, W.W., 1977. Sampling Techniques, third ed. Wiley, NewYork, NY.

Eby, D.W., Kostyniuk, L.P., Vivoda, J.M., 2001. Restraint usepatterns for older child passengers in Michigan. Accident Analysisand Prevention 33, 235-242.

Eby, D.W., Molnar, L.J., Olk, M.L., 2000. Trends in driver andfront-right passenger safety belt use in Michigan: 1984 to 1998.Accident Analysis and Prevention 32, 837-843.

Eby, D.W., Vivoda, J.M., Fordyce, T.A., 1999. Direct Observation ofSafety Belt Use in Michigan: Fall 1999. (Report No. UMTRI-99-

291D. W. £by el aI. / AcciMnl AItalysis /DId PrerIeIIJio1I 34 (2002) 285-291

1998. Preventing Worker Injuries and Deaths from Traffic-Re-lated Motor Vehicle Crashes. (Report No. 98-142). US Depart-ment of Health and Hwoan Scrvi~, Washington, DC.

Streff, F.M., Eby, D.W., Molnar, LJ., JokJCb, H.C., W~, R.R.,1993. Di~ Observation of Safety Belt and Moto~ycIc HdmetUse in Michigan: Fall 1993. (Report No. UMTRI-93-44). Univer-sity of Michigan Transportation Rcsca~h Institute, Ann Arbor,MI.

Ulmer, R.O., Preusser, C.W., Preusser, D.F., 1994. Evaluation ofCalifornia's Safety Belt Law CbanF to Primary Enfon:ement.(Report No. DOT -H5-808-205). US Department of Transporta-tion, Washington, DC.

United States Bureau of the Census, 1992. 1990 Census of Populationand Housina (from University of Michipn UM-U Library 00-pher-computcr datafilc).

Wagenaar, A.C., Molnar, LJ., Businski, K.L., 1987. Direct Observa-tion of Safety Belt Use in Michigan: December 1986. (Report No.UMTRI-87-O3). University of Michigan Transportation Rcsean:hInstitute, Ann Arbor, MI.

Williams, A.F., Wd]s, J.K., 1992. Safety belt use in New Hampshire.Journal of Safety Rcsca~h 23, 147-ISO.

Williams, A.F., Wells, J.K., Lund, A.K., 1987. Shoulder belt use infour states with belt use laWi. Alx:ident Analysis and Prevention19, 2SI-2~.

33). University of Michigan Transportation Research Institute,Ann Arbor, MI.

Federal Hiahway Administration, 1982. Hiahway Statistics 1982. USDepartment of Transportation, Wasbinat°n, DC.

FockJer, S.K., Cooper, D.S., 1990. Situational cbaraeteristics ofsafety belt use. A~t Analysis and Prevention 22, 109-118.

Jonah, B.A., Thiessen, R., Au-Yeung, E., Vincent, A., 1997. Sensa-tion Seeking and Risky Driving. Paper preKnted at Risk TwnaBehavior and Traffic Safety Symposium, Cape Cod, MA.

Li, L., Kim, K., Nitz, L., 1999. Prulictors of safety belt use amonscrash-involved drivers and front seat passengers: adjusting forover-reporting. Accident Analysis and Prevention 31, 631-638.

National Hiahway Traffic Safety Administration (NHTSA), 1992.Guidelines for state oblervational surveys of safety belt andmotorcycle helmet use. Federal Resister 57 (125), 28899-28904.

National Hiahway Traffic Safety Administration (NHTSA), 1997.National occupant protection use survey: controlled intersectiondetailed results. Research Note, February. US Department ofTransportation, Wasbinat°n, DC.

National Hiahway Traffic Safety Administration (NHTSA), 1998.Uniform Criteria for State Observational Surveys of Seat BeltUse. (Docket No. NHTSA-98-4280). US Department of Trans-portation, Washington, DC.

Natio.nal Institute for Occupational Safety and Heahb (NIOSH>,