Abstract
To improve traffic safety, states limit truck length and weight, and some set lower speed limits for trucks than for other vehicles. We examined the impact of truck-specific restrictions and general traffic-safety policies on fatality rates from crashes involving large trucks.
We used state-level data from 1991 to 2005 with a cross-sectional time-series model that controlled for several policy measures. We found that higher speed limits for cars and trucks contributed to higher fatality rates, but differential speed limits by vehicle type had no significant impact.
Truck-length limitations reduced fatalities in crashes involving large trucks. Our model estimates suggested that if all states had adopted a speed limit of 55 miles per hour for all vehicles in 2005, an additional 561 fatalities would have been averted.
LARGE TRUCKS ACCOUNT for less than 5% of registered vehicles in the United States and only 8% of the total miles driven, but they are disproportionately involved in passenger vehicle occupant deaths compared with other vehicle types.1 About 5000 fatalities and 120 000 injuries per year occur in large-truck crashes; 15% of these fatalities occur in large trucks, and 78% occur in the other vehicles involved.2,3
The Federal Motor Carrier Safety Administration set a goal of reducing fatalities associated with crashes involving trucks by 41% from 1996 to 2008, and the rate dropped from 2.81 fatalities per 100 million vehicle miles traveled (VMTs) in 1996 to 2.29 in 2004. The federal government regulates truck-driver behavior with required rests and time traveled per day, and the number of on-site safety reviews at trucking companies doubled from 1998 to 2004.2 The states, however, are primarily responsible for American traffic-safety policy, including licensing guidelines, speed limits, laws concerned with drinking and driving, and seat-belt requirements. Some state policies have been enacted in direct response to federal pressure, such as the adoption of 0.08 blood alcohol content (BAC) laws, but state traffic-safety policies vary considerably.
State policies often regulate driving behavior without regard to vehicle type (e.g., prohibiting drug use, limiting BAC, setting a minimum legal drinking age, and requiring the use of seat belts). These policies seek to limit injurious and fatal traffic crashes by all drivers. For truck transportation, states not only regulate speed but also limit truck weight, length, and height, and states use scales to detect violations.4 The federal government has set a standard for truck weight and length on interstate highways, but states set maximum truck lengths and weights for all other state roadways.
States have frequently modified speed limits over recent years as legislators have responded to national regulations, federal financial incentives, interest-group pressure, and constituent preferences. Although there is a spirited ongoing debate on the general impact of speed limits, truck speed limits have received scant scholarly attention. The common assumption is that higher speed limits will pose a greater danger to vehicle occupants, but research on the effect of speed limits on traffic crash fatalities has shown mixed results.5–14 In addition, studies on the impact of average truck speeds on fatalities show inconsistent results; some studies have shown that speed alone had no impact on fatalities,15 whereas others have found a significant association between higher speed limits and more fatalities.16,17
States often set lower speed limits for large trucks than for other motor vehicles. Differential speed limits by vehicle type have the potential to create 2 streams of traffic flowing at different rates.18 Studies have shown that differential speed limits resulted in observed speed differences between cars and trucks, but the actual differences were often less than the posted differences.19–21 Further, studies have found that a greater speed difference was associated with a significant increase in fatalities.15–17,22
Since 1995, states have had the freedom to set maximum speed limits. As shown in Figure 1, trucks and cars had the same maximum rural speed limits for a majority of years from 1991 to 2005. The 55-mile-per-hour (mph) speed limit was in force for both passenger vehicles and large trucks 6% of the time. Truck and car speed limits at 60 mph account for 2% of state-year observations, truck and car limits at 65 mph account for 45%, truck and car limits at 70 mph account for 15%, and truck and car limits at 75 mph account for 12%. Differential speed limits account for 20% of observations, with the largest difference at 15 mph (70 for cars and 55 for trucks) in 1% of observations, a 10-mph difference in 15%, and a 5-mph difference in 4%.
Note. Speed limits are given in miles per hour, with the speed limits for cars listed before those for trucks. The maximum rural speed limit in each state for each year was a state–year observation.
Figure 2 provides an initial assessment of the relationship between maximum rural speed limits and the traffic-fatality rate. The graph clusters the data by car speed limits; within each cluster, there is a separate bar for each truck speed limit associated with that cluster's car speed limit. Generally, states with higher speed limits have higher fatality rates. The 65-mph and 75-mph clusters have higher fatality rates than do the 55-mph and 60-mph clusters. There is no real difference between the truck speed categories in the lower-speed clusters, but the 70-mph cluster shows a strong positive relationship between higher truck speeds and much higher fatality rates.
FIGURE 2.
Mean number of fatalities in crashes involving large trucks per billion vehicle miles traveled (VMT), by rural interstate speed limits: United States, 1991–2005.
Little research has been conducted on the impact of trucking regulations on traffic safety, but we intended to fill this gap. Our main areas of interest were the maximum speed limits for trucks and passenger vehicles and the differences between the 2 speed limits. Our research focused on whether truck-specific speed limits, length restrictions, and weight limits affected US traffic-fatality rates. We used state-level data to conduct a cross-sectional time-series regression analysis of traffic fatalities from crashes involving large trucks in the United States from 1991 to 2005. We also examined the impact of a number of general traffic-safety policies on fatality rates from crashes involving large trucks.
In addition to speed limits, alcohol laws have long been at the forefront of states' traffic-safety efforts. Studies have shown that the minimum legal drinking age affects passenger vehicle safety,23,24 but this factor is invariant during our time frame. To test policies seeking to reduce fatalities related to drinking and driving, we dichotomized the presence of a 0.08 BAC state law, and we included the state's alcohol consumption level in our model. Previous research shows that states with higher levels of alcohol consumption experience higher truck-crash fatality rates25–28 and that 0.08 BAC laws reduce such fatalities.29
States also enact passenger-restraint laws to enhance traffic safety, and studies have found a significant fatality-rate reduction associated with laws requiring seat belt use in passenger vehicles.27,28,30–32 We expected this relationship to hold for fatalities from crashes involving large trucks.
Highway conditions also influence crash fatalities, and our model included funding levels for 3 categories of highway expenditures: capital, maintenance, and police and safety.27,28,33 We expected capital expenditures to increase roadway mileage and increase traffic flow, which accommodate higher speeds that could contribute to crash fatalities. Maintenance expenditures may eliminate poor road conditions or enhance safety features, thereby lowering fatality rates. We expected higher levels of law enforcement expenditures to increase compliance with safety laws and therefore decrease the fatality rate.
METHODS
We tested a cross-sectional time-series regression model of truck-crash fatalities in the United States for 1991 to 2005. The dependent variable was the number of fatalities in vehicle crashes involving a large truck per billion VMTs in the state. We obtained the dependent variable from the Fatality Analysis Reporting System, a dataset maintained by the National Highway Traffic Safety Administration.34 We dropped Hawaii from the analysis because (1) there was a low number of fatalities (5 or fewer annually in 10 of the years covered by our analysis), (2) there were few truck registrations, (3) there was no truck traffic from adjoining states, and (4) there was no variation in speed limits, truck length restrictions, or truck weight restrictions during the study period.
Given the dominance of the cross section (49 states) over the time component (15 years), we used generalized least squares regression to estimate models. The states were the unit of analysis, and the model included fixed effects for states and for each year. The state fixed effects accounted for unique circumstances in each state, and the year fixed effects accounted for changes over time in the data set, such as national factors that may have influenced traffic safety across states.
In estimating the models, we tested for random effects in the cross sections. We used the Breusch and Pagan Lagrangian multiplier test,35 and we found that χ2 = 644, with P < .001. This suggested that we reject the random-effects model and use a fixed-effects specification. Then we conducted the Wooldridge36 test for autocorrelation in the panel [H0 - no first-order autocorrelation] and found evidence of a first-order autoregressive structure (F1,48 = 4.09, with P = .048). Therefore, the model relied on a generalized least squares specification with a first-order autoregressive component.
Whereas other studies have examined the relationship between observed speeds and the fatality rate, we focused on speed limits. The only measure consistently available during the time period we studied was the maximum speed limit allowed on any state roadway. Typically, this maximum applies only to rural interstate highways, but some states allow the maximum speed on rural state highways as well. Other states (usually with large urban populations) have no differences between urban and rural speed limits. We used a variety of sources for speed limit data,37 and when sources conflicted, we confirmed all data by checking state Web sites for relevant statutes, driver's license guides, or highway patrol documents. In some cases, we contacted states to confirm the data.
We obtained data on truck length and weight restrictions from several sources, including the National Automobile Dealers Association,38 the American Trucking Association, and state agencies. Length and weight restrictions vary by roadway type; for each state we used the highest limit regularly available without a special permit. We obtained data on seat belt laws and 0.08 BAC laws from Traffic Safety Facts.39 We included a dichotomous measure for whether a state had mandatory seat belt legislation in our model. We reported state alcohol consumption in gallons of ethanol per capita as reported by the National Institute of Alcohol Abuse and Alcoholism.40
Capital, maintenance, and police and safety expenditures (listed in Highway Statistics 41) included expenditures from all levels of government. To control for inflation, we adjusted dollar values to reflect constant dollars, based on the 1982 to 1984 market basket of the Consumer Price Index for All Urban Consumers from the US Bureau of Labor Statistics.42 To reduce the effect of short-term fluctuations in budgets or the impact of a single large project, we used the 5-year average of capital expenditures.
We also included several control variables that could influence traffic safety, such as the level of risk associated with the volume of truck traffic. The number of VMTs by trucks would allow a more precise measure of truck traffic, but VMT estimates have only been available since 1994. To extend our time series and capture maximum variation in state policies, we operationalized the amount of truck traffic on a state's roadways as diesel fuel consumption in that state per VMT.41 For the observed data (1994–2005), the correlation between diesel fuel consumption and truck VMT was 0.97, indicating that the operationalization would be a viable proxy. We expected that as the diesel fuel consumption per VMT in a state increased, the truck-crash fatality rate would increase.
To capture the total amount of traffic on a state's roadways, we included the total VMT by all vehicles per capita.41 The total VMT reflected underlying trends in population, number of vehicles on the road, and distances traveled. We expected that total VMT per capita would correlate positively with truck-crash fatality rates, because higher VMT suggests more driving time, implying greater crash risk.
Per capita income (measured in constant dollars, with 1982–1984 = $100) reflects economic factors,43 but interpretations of its impact vary. Some studies suggest that citizens with higher incomes will demand greater safety,26 whereas others suggest that drivers with higher incomes will place a higher value on time, thus increasing risky driving behavior.44,45 Although risk compensation may affect individual behavior, we expected higher average income to contribute to safety in various ways, such as the purchase of newer and generally safer vehicles. Unemployment rates also influence driving habits; periods of high unemployment may suppress driving as citizens economize their leisure activities.27,28 We obtained data on income and unemployment from the US Census Bureau.43
We also included control variables for weather and population density.43 Severe cold weather may correlate with unsafe road conditions that limit traffic, thus reducing fatalities. Also, whereas precipitation may slow truck traffic, it also increases the likelihood of a driver losing control and experiencing a more severe crash. Population density is a proxy for urban areas,43 which have lower speed limits and higher levels of congestion that may reduce truck speed and induce truck drivers to seek less traveled routes or travel at less congested times, resulting in lower fatality rates.
RESULTS
Our dependent variable was the fatality rate in crashes involving large trucks (fatalities per billion VMT). Because of the high degree of collinearity among our main variables of interest, we used 4 model specifications: maximum speed limit for trucks, maximum speed limit for cars, equal maximum speed limits for both vehicle types, and the total combined speed-limit amount above 55 mph (Table 1). We relied on a dichotomous variable for whether the 2 speed limits were equivalent (1) or not (0), but in unreported models, we observed no differences if the variable was measured as an interval variable.
TABLE 1.
Cross-Sectional Time-Series Models of Fatalities per Billion VMT From Crashes Involving Large Trucks: United States, 1991–2005
| Variables | Truck Speed Limit Only, b (SD) | Truck Speed Limit With Equal Speed Limits for Trucks and Cars, b (SD) | Car Speed Limit With Equal Speed Limits for Trucks and Cars, b (SD) | Total Combined Speed-Limit Amount Above 55 mph With Equal Speed Limits for Trucks and Cars, b (SD) |
| Truck speed limit | 0.0153* (0.0067) | 0.0148* (0.0068) | … | … |
| Car speed limit | … | … | 0.0156* (0.0070) | … |
| Total combined speed limit amount above 55 mph | … | … | … | 0.0078* (0.0035) |
| Equal speed limits for cars and trucks | … | 0.0589 (0.1421) | 0.2058 (0.1453) | 0.1309 (0.1398) |
| Truck length limit | 0.0233* (0.0109) | 0.0231* (0.0109) | 0.0232* (0.0109) | 0.0232* (0.0109) |
| Truck weight limit | –0.0036 (0.0089) | –0.0036 (0.0089) | –0.0034 (0.0089) | –0.0035 (0.0089) |
| Diesel fuel consumption per VMT | 0.0222** (0.0080) | 0.0223** (0.0080) | 0.0222** (0.0080) | 0.0222** (0.0080) |
| Total VMT per capita | –0.0739* (0.0372) | –0.0753* (0.0374) | –0.0735* (0.0373) | –0.0745* (0.0374) |
| Alcohol consumption | 0.2590 (0.1939) | 0.2730 (0.1970) | 0.2821 (0.1968) | 0.2775 (0.1968) |
| 0.08 BAC law | –0.0205 (0.0530) | –0.0185 (0.0532) | –0.0182 (0.0531) | –0.0184 (0.0532) |
| Seat belt law | –0.2962** (0.0930) | −0.3000** (0.0935) | –0.2965** (0.0935) | –0.2980** (0.0935) |
| Capital expenditures per capitaa | 0.0247** (0.0064) | 0.0248** (0.0064) | 0.0248** (0.0065) | 0.0248** (0.0065) |
| Maintenance expendituresb | 0.0127 (0.0179) | 0.0122 (0.0180) | 0.0120 (0.0179) | 0.0121 (0.0179) |
| Police and safety expendituresb | –0.0362 (0.0300) | –0.0355 (0.0300) | –0.0350 (0.0300) | –0.0353 (0.0300) |
| Income per capitab | 0.1100** (0.0403) | 0.1083** (0.0405) | 0.1052** (0.0405) | 0.1067** (0.0405) |
| Unemployment | –0.0564* (0.0233) | –0.0572* (0.0234) | –0.0561* (0.0233) | –0.0567* (0.0233) |
| Temperature | 0.0029 (0.0143) | 0.0028 (0.0143) | 0.0025 (0.0143) | 0.0027 (0.0143) |
| Precipitation | 0.0012 (0.0030) | 0.0012 (0.0030) | 0.0013 (0.0030) | 0.0012 (0.0030) |
| Population density | –0.0002 (0.0018) | –0.0003 (0.0018) | –0.0003 (0.0018) | –0.0003 (0.0018) |
| Constant | −1.3900 (1.3743) | −1.3615 (1.3768) | −1.5705 (1.3989) | −0.6287 (1.3337) |
| F statistic (df) | 7.24** (31,655) | 7.01** (31,655) | 7.04** (31,655) | 7.03** (31,655) |
Note. VMT = vehicle miles traveled; BAC = blood alcohol content. State and year fixed effects were estimated but not reported. The model relied on a generalized least squares specification with a first-order autoregressive component, which fits a model when the disturbance term is first-order autoregressive. Hawaii was excluded. For state–year observations, n = 735; for State cross = sections, n = 49.
Five-year average in constant 2005 dollars.
In constant 2005 dollars.
*P < .05; **P < .01.
The first model revealed that a higher truck speed limit had a significant positive association with the truck-crash fatality rate. In the second model, when we controlled for the same speed limit for cars and trucks, the truck speed limit parameter estimates was also positive and significant. The equal-speed-limit variable did not approach significance, and the coefficient's sign was counter to expectations, with a positive value (and thus a higher fatality rate) rather than a negative one. A higher speed limit for trucks contributed to significantly higher fatality rates, but differences in speed limits between cars and trucks had no significant impact. The third model showed that car speed limits had similar positive and significant effects on the truck-crash fatality rate, and the equal-speed variable remained nonsignificant. The fourth model showed that the total combined speed-limit amount above 55 mph had a significant effect on the fatality rate; the equal-speed variable was again nonsignificant.
Two safety policies geared toward all vehicles had different effects on the fatality rate from crashes involving large trucks. The adoption of a 0.08 BAC law had no significant impact on truck-crash fatalities, and a seat belt law had a significant negative effect. Although other studies29 have found safety effects for a 0.08 BAC law, drinking and driving may be more widespread in urban areas or on rural arterials than on highways with a high truck volume. Studies have demonstrated that seat belt laws significantly reduce all traffic fatalities,27,28,30–32 and we found that seat belt laws were associated with reduced fatality rates in crashes involving large trucks.
Truck-specific regulatory policies governing maximum length and maximum weight showed mixed effects on truck-crash fatalities. A higher maximum truck length was significantly associated with a higher fatality rate, but the effect of a higher maximum truck weight was nonsignificant and not in the hypothesized direction. There are several ways to explain the null finding on maximum weight. First, a large proportion of truck traffic occurred on interstate highways, which have a consistent federal standard of 80 000 pounds in maximum weight and 65 feet in maximum overall length. Second, there were numerous exceptions to the weight limits, such as permits for manufactured homes and large, nonseparable loads. Third, studies suggest that a small fraction of trucks exceed 80 000 pounds.46 Fourth, in a 2002 study, less than 1% of fatalities involved trucks that weighed more than 100 000 pounds, and less than 1% of fatalities involved trucks longer than 80 feet in combined length.3 Finally, the low variance in the weight limit variable may have limited the results, because 34 states had a truck weight limit of 80 000 pounds in 2005.
Of the expenditure variables, only capital expenditures attained significance. Generally, states that spent more on expanding highway capacity had significantly higher truck-crash fatality rates. Higher capital expenditures may increase capacity, allowing higher speeds that contribute to more-severe crashes. On the other hand, maintenance expenditures and police and safety expenditures were not significantly related to truck-crash fatality rates. Such expenditures may focus on roadways not heavily used by large trucks or may be used for features that increase safety for cars but not trucks.
The control variables generally performed as expected. Diesel fuel consumption per VMT, as a proxy for truck travel, had a strong positive parameter estimates. As truck traffic increased, the truck-crash fatality rate also increased. The total VMT per capita variable, however, was significant and negative (contrary to expectations). In other words, the greater the mileage driven by the average state driver, the lower the state's truck-crash fatality rate. Of the remaining control variables, only income and unemployment rates attained significance; alcohol consumption, temperature, precipitation, and population density were nonsignificant.
DISCUSSION
Our results agreed with previous research that found a significant association between speed limits and fatalities,16,17 but our results contradicted studies finding a positive relationship between differential speed limits and fatalities.15–17,22 Our results suggest that states can reduce traffic fatalities from crashes involving large trucks by lowering speed limits for all drivers and that setting lower limits for trucks than for cars will not mitigate the safety effects. Overall, higher speed limits for all vehicles appeared to be a major factor in the fatality rate from crashes involving large trucks, and a speed-limit difference between cars and trucks was not a significant issue. Research has shown that differences in actual speeds are often smaller than differences in posted speeds,19–21 which may explain this finding of nonsignificance.
To add context to the interpretation of the results, Table 2 presents the estimated change in the annual number of fatalities for each state based on the fourth model in Table 1. With the parameter estimates for the total combined speed-limit amount above 55 mph (0.0078) and all other variables held constant at 2005 state levels, we predicted the number of fatalities in 2 scenarios: a state adopting a 55-mph speed limit for cars and trucks and a state adopting a 75-mph speed limit for all vehicles. We then used these values to estimate the expected change in fatalities relative to the actual 2005 fatalities. For example, in 2005, California's speed limit for cars was 70 mph and for trucks was 55 mph, and the state had 428 traffic fatalities. The model predicted that if California had adopted a 75-mph speed limit for all vehicles in 2005, the state would have had 64 more traffic fatalities than it did. Alternatively, if Texas had reduced its 2005 speed limits of 75 mph for cars and 65 mph for trucks to 55 mph for all vehicles, the model predicted that there would have been 54 fewer fatalities than the 502 observed in the state that year.
TABLE 2.
Predicted Annual Change in Truck-Crash Fatalities per State in 2005 Based on Hypothetical Adoption of 55-mph and 75-mph Speed Limits
| State | Actual Fatalities, No. | Adoption of 55-mph Speed Limit, % Change | Adoption of 75-mph Speed Limit, % Change |
| Speed Limits for Cars 65 mph, Trucks 55 mph | |||
| Illinois | 191 | −9 | +26 |
| Ohio | 177 | −9 | +26 |
| Oregon | 66 | −3 | +8 |
| Speed Limits for Cars 65 mph, Trucks 65 mph | |||
| Alaska | 5 | −1 | +1 |
| Connecticut | 17 | −5 | +5 |
| Delaware | 8 | −2 | +2 |
| Kentucky | 124 | −7 | +7 |
| Maine | 19 | −2 | +2 |
| Maryland | 60 | −9 | +9 |
| Massachusetts | 24 | −9 | +9 |
| New Hampshire | 11 | −2 | +2 |
| New Jersey | 98 | −11 | +11 |
| New York | 147 | −22 | +22 |
| Pennsylvania | 183 | −17 | +17 |
| Rhode Island | 1 | −1 | +1 |
| Vermont | 9 | −1 | +1 |
| Virginia | 112 | −12 | +12 |
| Wisconsin | 87 | −9 | +9 |
| Speed Limits for Cars 70 mph, Trucks 55 mph | |||
| California | 428 | −39 | +64 |
| Speed Limits for Cars 70 mph, Trucks 60 mph | |||
| Michigan | 111 | −16 | +16 |
| Washington | 68 | −9 | +9 |
| Speed Limits for Cars 70 mph, Trucks 65 mph | |||
| Arkansas | 116 | −6 | +4 |
| Indiana | 138 | −14 | +9 |
| Speed Limits for Cars 70 mph, Trucks 70 mph | |||
| Alabama | 122 | −14 | +5 |
| Florida | 406 | −46 | +15 |
| Georgia | 229 | −26 | +9 |
| Iowa | 73 | −7 | +3 |
| Kansas | 80 | −7 | +3 |
| Louisiana | 122 | −10 | +4 |
| Minnesota | 69 | −13 | +4 |
| Mississippi | 91 | −9 | +3 |
| Missouri | 166 | −16 | +5 |
| North Carolina | 204 | −22 | +8 |
| South Carolina | 124 | −12 | +4 |
| Tennessee | 156 | −17 | +6 |
| West Virginia | 55 | −5 | +2 |
| Speed Limits for Cars 75 mph, Trucks 65 mph | |||
| Idaho | 34 | −3 | +1 |
| Montana | 23 | −3 | +1 |
| Texas | 502 | −54 | +18 |
| Speed Limits for Cars 75 mph, Trucks 75 mph | |||
| Arizona | 97 | −18 | … |
| Colorado | 68 | −14 | … |
| Nebraska | 48 | −6 | … |
| Nevada | 54 | −6 | … |
| New Mexico | 63 | −8 | … |
| North Dakota | 17 | −2 | … |
| Oklahoma | 121 | −15 | … |
| South Dakota | 13 | −3 | … |
| Utah | 32 | −8 | … |
| Wyoming | 31 | −3 | … |
| Totals | 5200 | −561 | +362 |
Note. The predicted value is the expected reduction or increase in fatalities (compared with the actual fatalities in each state for 2005) associated with the hypothetical adoption of a 55-mph speed limit for all vehicles and a 75-mph speed limit for all vehicles. The predictions rely on model parameter estimates that specify the total combined speed limit amount above 55 mph. We used hypothetical changes in the speed limit laws compared with the actual speed limits in effect in 2005, with all other variables held constant at the actual 2005 values for each state. Hawaii was excluded from the analysis.
Overall, the model predicted that if all states had changed their actual 2005 speed limits to a 75-mph limit, 362 more fatalities would have occurred. Alternatively, if all states had dropped their 2005 speed limits to 55 mph, 561 fewer fatalities would have occurred. The potential annual total shift of 923 fatalities created by the change from 55 mph to 75 mph represents almost 18% of the actual 5200 fatalities from crashes involving large trucks in 2005, suggesting that higher speed limits have contributed to thousands of additional fatalities from truck crashes over the past decade.
States have an array of policy tools they can use to reduce the fatality rate from crashes involving large trucks. Our results suggested that truck speed limits, car speed limits, seat belt laws, and truck-length limits are significant predictors of fatality rates in crashes involving large trucks. Differential speed limits for trucks and cars did not affect safety, and truck weight limits were not significantly associated with fatality rates from crashes involving large trucks.
Human Participant Protection
No protocol approval was needed for this study.
References
- 1.Lyman S, Braver ER. Occupant deaths in large truck crashes in the US: 25 years of experience. Accid Anal Prev 2003;35:731–739 [DOI] [PubMed] [Google Scholar]
- 2.Large Truck Safety. Washington, DC: US Government Accountability Office; 2006. GAO publication 06–156 [Google Scholar]
- 3.Matteson A, Blower D, Woodrooffe J. Trucks Involved in Fatal Accidents, Factbook 2002. Ann Arbor, MI: Transportation Research Institute; 2004 [Google Scholar]
- 4.Teske P, Best S, Mintrom M. Deregulating Freight Transportation: Delivering the Goods Washington, DC: AEI Press; 1995 [Google Scholar]
- 5.Meier KJ, Morgan DR. Speed kills: a longitudinal analysis of traffic fatalities and the 55 mph speed limit. Rev Policy Res 1981;1:157–167 [Google Scholar]
- 6.Kamerud DB. Evaluating the new 65 mph speed limit. : Graham JD, ed Preventing Automobile Injury: New Findings from Evaluation Research. Dover, MA: Auburn House; 1988 [Google Scholar]
- 7.Garber S, Graham JD. The effects of the new 65 mile-per-hour speed limit on rural highway fatalities: a state-by-state analysis. Accid Anal Prev 1990;22:137–149 [DOI] [PubMed] [Google Scholar]
- 8.Baum HM, Wells JK, Lund AK. Motor vehicle crash fatalities in the second year of 65 mph speed limits. J Safety Res 1990;21:1–8 [Google Scholar]
- 9.Baum HM, Wells JK, Lund AK. The fatality consequences of the 65mph speed limits, 1989. J Safety Res 1991;22:171–177 [Google Scholar]
- 10.Chang GL, Paniati JF. Effects of 65 mph speed limit on traffic safety. J Transp Eng 1990;116:213–226 [Google Scholar]
- 11.Wagenaar AC, Streff FM, Schultz RH. Effects of the 65 mph speed limit on injury morbidity and mortality. Accid Anal Prev 1990;22:571–585 [DOI] [PubMed] [Google Scholar]
- 12.Pant PD, Adhami JA, Neihaus JC. Effects of the 65-mph speed limit on traffic accidents in Ohio. Transp Res Rec 1992;1375:53–60 [Google Scholar]
- 13.Lave C, Elias P. Did the 65 mph speed limit save lives? Accid Anal Prev 1994;26:49–62 [DOI] [PubMed] [Google Scholar]
- 14.Houston DJ. Implications of the 65-mph speed limit for traffic safety. Eval Rev 1999;23:304–315 [DOI] [PubMed] [Google Scholar]
- 15.Lave CA. Speeding, coordination, and the 55 mph limit. Am Econ Rev 1985;75:1159–1164 [Google Scholar]
- 16.Levy DT, Asch P. Speeding, coordination, and the 55 mph limit [comment]. Am Econ Rev 1989;79:913–915 [Google Scholar]
- 17.Fowles R, Loeb PD. Speeding, coordination, and the 55 mph limit [comment]. Am Econ Rev 1989;79:916–921 [Google Scholar]
- 18.Johnson SL, Pawar N. Cost-Benefit Evaluation of Large Truck–Automobile Speed Limit Differentials on Rural Interstate Highways. Fayetteville, AR: Mack-Blackwell Transportation Center; 2005 [Google Scholar]
- 19.Baum HM, Esterlitz JR, Zador P, Penny M. Different speed limits for cars and trucks: do they affect vehicle speeds? Transp Res Rec 1991;1318:3–7 [Google Scholar]
- 20.Mace DJ, Heckard R. Effect of the 65 mph Speed Limit on Travel Speeds and Related Crashes. Washington, DC: US Dept of Transportation; 1991 [Google Scholar]
- 21.Harkey DL, Mera R. Safety Impacts of Different Speed Limits on Cars and Trucks. Washington, DC: Federal Highway Administration; 1994. FHWA publication RD-93-161 [Google Scholar]
- 22.Garber NJ, Gadiraju R. Impact of differential speed limits on the speed of traffic and the rate of accidents. Transp Res Rec 1992;1375:44–52 [Google Scholar]
- 23.Saffer H, Grossman M. Beer taxes, the legal drinking age, and youth motor vehicle fatalities. J Legal Stud 1987;16:351–374 [Google Scholar]
- 24.O'Malley PM, Wagenaar AC. Effects of minimum drinking age laws on alcohol use, related behaviors and traffic crash involvement among American youth: 1976–1987. J Stud Alcohol 1991;52:478–491 [DOI] [PubMed] [Google Scholar]
- 25.Asch P, Levy DT. Does the minimum drinking age affect traffic fatalities? J Policy Anal Manage 1987;6:180–192 [Google Scholar]
- 26.Legge JS, Jr, Park J. Policies to reduce alcohol-impaired driving: evaluating elements of deterrence. Soc Sci Q 1994;75:594–606 [Google Scholar]
- 27.Houston DJ, Richardson LE, Jr, Neeley GW. Legislating traffic safety: a pooled time series analysis. Soc Sci Q 1995;76:328–345 [Google Scholar]
- 28.Houston DJ, Richardson LE, Jr, Neeley GW. Mandatory seat belt laws in the states: a study of fatal and severe occupant injuries. Eval Rev 1996;20:146–159 [DOI] [PubMed] [Google Scholar]
- 29.Tippetts AS, Voas RB, Fell JC, Nichols JL. A meta-analysis of.08 BAC laws in 19 jurisdictions in the US. Accid Anal Prev 2005;37:149–161 [DOI] [PubMed] [Google Scholar]
- 30.Chorba TL, Reinfurt D, Hulka BS. Efficacy of mandatory seat-belt use legislation. JAMA 1988;260:3593–3597 [PubMed] [Google Scholar]
- 31.Wagenaar AC, Maybee RG, Sullivan KP. Mandatory seat belt laws in eight states: a time-series evaluation. J Safety Res 1988;19:51–70 [Google Scholar]
- 32.Houston DJ, Richardson LE., Jr Traffic safety and the switch to a primary seat belt law: the California experience. Accid Anal Prev 2002;34:743–751 [DOI] [PubMed] [Google Scholar]
- 33.Zlatoper TJ. Determinants of motor vehicle deaths in the United States: a cross-sectional analysis. Accid Anal Prev 1991;23:431–436 [DOI] [PubMed] [Google Scholar]
- 34.US National Highway Traffic Safety Administration. Fatality Analysis Reporting System Web site. Available at: http://www-fars.nhtsa.dot.gov/Main/index.aspx. Accessed September 7, 2008
- 35.Breusch TS, Pagan AR. A simple test for heteroscedasticity and random parameter estimates variation. Econometrica 1979;47:1287–1294 [Google Scholar]
- 36.Wooldridge JM. Econometric Analysis of Cross Section and Panel Data Cambridge, MA: MIT Press; 2002 [Google Scholar]
- 37.Accident Facts. Chicago, IL: National Safety Council; 1990–2006 [Google Scholar]
- 38.Title and Registration Book: A Summary of Motor Vehicle Laws and Regulations. Costa Mesa, CA: National Automobile Dealers Association; 1990–2002 [Google Scholar]
- 39.Traffic Safety Facts. Washington, DC: US National Highway Traffic Safety Administration; 1990–2006 [Google Scholar]
- 40.US National Institute on Alcohol Abuse and Alcoholism Web site. Available at: http://www.niaaa.nih.gov/Resources/DatabaseResources/QuickFacts/AlcoholSales/consum03.htm. Accessed October 14, 2007
- 41.Highway Statistics. Washington, DC: US Federal Highway Administration; 1990–2006 [Google Scholar]
- 42.US Bureau of Labor Statistics. Consumer Price Index Web site. http://www.bls.gov/cpi. Accessed September 7, 2008
- 43.Statistical Abstract of the United States. Washington, DC: US Census Bureau; 1990–2006 [Google Scholar]
- 44.Peltzman S. The effects of automobile safety regulation. J Polit Econ 1975;83:677–725 [Google Scholar]
- 45.Graham JD, Garber S. Evaluating the effects of automobile safety regulation. J Policy Anal Manage 1984;3:206–224 [Google Scholar]
- 46.Trucks and Transportation. Lansing: Michigan Dept of Transportation. 1998. Available at: http://web.archive.org/web/20070310171845/http://www.michigan.gov/documents/truckinfo_16563_7.pdf. Accessed September 2, 2008