16 Transportation Research Record: Journal of the Transportation Research Board, No. 2386, Transportation Research Board of the National Academies, Washington, D.C., 2013, pp. 16–25. DOI: 10.3141/2386-03 Department of Civil and Environmental Engineering, Wayne State University, 5050 Anthony Wayne Drive, Detroit, MI 48202. Corresponding author: P. T. Savolainen, savolainen@wayne.edu. networks of rest areas. Economic issues have led to recent rest area closures in at least 14 states and several other states have down- graded the services they provide (4–9). Given these issues, it is impor- tant that states carefully weigh the potential cost savings created by such closures against prospective benefits, including reductions in fatigue-related crashes. Past research has examined various aspects of traffic safety as they relate to rest areas. An NCHRP report discussed several factors that have limited the ability of researchers to assess the safety impacts of rest areas (10). These factors include the presence of urbanized areas and major freeway interchanges, both of which tend to confound spatial crash analyses by creating crash hot spots as well as a lack of accurate traffic volume data for specific road segments (10). The NCHRP study ultimately concluded that crashes may decrease as rest area spacing is decreased up to a limit of approximately 50 mi. A Michigan study analyzed the relationship between rest area spac- ing and the rate of single-vehicle truck crashes on select freeways (11). The research showed that the majority of single-vehicle truck collisions occur between midnight and 8:00 a.m. Hazard models were developed to examine spatial trends in truck-involved crashes with respect to rest areas that were located 50 mi or more apart. Study results showed a positive relationship between rest area spacing and fatigue-related truck crashes, with a significant increase in crashes when rest area spacing exceeded 30 mi. A Minnesota study showed similar results: single-vehicle truck crash densities were found to increase during all times of the day at distances greater than 30 mi beyond a rest area (12). During nighttime hours, there was a signifi- cant increase in single-vehicle crash densities that may have been related to parking capacity issues at rest areas. Both studies attrib- uted increased crash rates to facilities that were overcrowded or had insufficient available truck parking (11, 12). A subsequent California study involved a spatial analysis aimed at determining the effects of rest area spacing on fatigue-related crashes along Interstate facilities (13). Crash data for 1995 to 2005 were obtained for all fatigue-related crashes based upon “strict” and “expanded” definitions. The study found that crashes due to fatigue generally decreased just downstream of a rest area. Similar to the studies on truck-involved crashes, the frequency of collisions due to driver fatigue significantly increased at distances of more than 30 mi from the rest areas. A recent Texas study estimated rest area safety impacts on the sim- ple basis of a before-and-after comparison of casualty rates on high- ways where rest areas had been constructed (14). The authors note that such a comparison may lead to erroneous inferences as to the safety effects of rest areas due to temporal variations, such as differences in traffic volumes and changes in infrastructure that occurred during this time period. In summary, there has been limited research as to the safety impacts of roadside rest areas to date. Furthermore, the extant research Effects of Public Rest Areas on Fatigue-Related Crashes Adam McArthur, Jonathan Kay, Peter T. Savolainen, and Timothy J. Gates Fatigue-related crashes account for 2.2% to 2.6% of all fatal crashes in the United States on an annual basis. These types of crashes are prevalent in rural areas, and public roadside rest areas were developed largely to alleviate motorist fatigue and reduce the opportunity for such crashes by providing safe parking areas for tired drivers. However, research about the safety effects of rest areas has been limited. Results are presented of a spatial analysis that investigates how a road segment’s proximity to a rest area affects the frequency of fatigue-related crashes. Poisson and nega- tive binomial models are estimated for freeways and two-lane highways to isolate the effects of proximity while controlling for other relevant factors, such as traffic volume. The results of these models indicate that the prox- imity of a road segment to the nearest rest area significantly influences crash frequencies on both freeways and two-lane highways, where traffic volumes tended to have similar effects, although the effects of proximity were slightly more pronounced on two-lane highways. The study results suggest that roadside rest areas provide a safety benefit, and the crash prediction models developed as a part of this research provide a starting point for quantifying these impacts. NHTSA estimates that approximately 83,000 police-reported crashes occur annually because of driver fatigue, including 886 fatal crashes and 37,000 injury crashes (1). These estimates suggest that fatigue is a contributory factor in 2.2% to 2.6% of all fatal crashes in the United States on an annual basis (1). In the state of Michigan, a total of 1,987 fatigue- and sleep-related crashes were reported by law enforcement in 2011 (2). These national and state figures are likely to underreport the true degree of this problem since such crashes are often not eas- ily distinguishable by an investigating officer; drivers are unlikely to admit having been drowsy or asleep at the wheel (3). Crashes involving fatigue and drowsy driving are most prevalent in rural areas and often involve drivers departing from their travel lane and result in head-on, sideswipe, and fixed-object collisions. Since these types of crashes tend to be among the most severe in terms of occupant inju- ries, reducing drowsy driving is a critical focus area for contemporary traffic safety efforts. Public rest areas were developed largely to alleviate motorist fatigue and reduce the occurrence of fatigue-related crashes by providing safe parking areas that are easily accessible to tired driv- ers 24 hours a day. However, in an economic climate of declining revenues, it has become challenging for states to maintain their