Wemple, Elizabeth

The objective of this study was to evaluate the safety impacts of increasing the speed limit from 55 mph to 60 mph on two-lane, two-way state highway road segments in Minnesota. An empirical Bayes (EB) before-after analysis was used to estimate crash modification factors (CMFs) for both segments and intersections. The segment analysis showed an 8 percent reduction in total crashes that was statistically significant, alongside a significant 15 percent increase in combined fatal, serious injury, and minor injury (KAB) injury crashes. The range of most of the segment CMFs hovered close to 1. The intersection analysis was split into two groups (all traffic control types and thru-stop control only). The aggregate CMFs for all intersections within these two groups showed, on average, between a 10% and 20% statistically significant reduction in total and injury crashes. Analysis was also performed on four subgroups (3- and 4-leg, lighting/no lighting) within the two main intersection groups. Disaggregating the intersections into further groups led to smaller sample sizes that led to higher standard errors showing a widespread range of CMFs around 1 for the individual crash types and severities. The aggregate estimated crash safety effects (for total and injury crashes) for combined segments and intersection sites showed a reduction in total crashes but an increase in the KAB injury crashes.

This evaluation determined the change in crash frequency, type or severity associated with longitudinal sinusoidal rumble strips on rural two-lane undivided Minnesota roadways constructed between 2018 and 2022. Crash modification factors (CMFs) were estimated using cross-sectional analysis to compare crash experience of locations with sinusoidal rumble strips (i.e., centerline only, centerline and shoulder, or shoulder only) compared to roads with rectangular rumble strips. The cross-sectional analysis matched sites with sinusoidal and rectangular rumble strips using matched-pair comparisons. Negative binomial (NB) or Poisson log-linear regression models were used to model the crashes at all treatment and non-treatment sites. There was a total of approximately 327 miles of treated (i.e., centerline only, centerline and shoulder, or shoulder only sinusoidal rumble strips) and approximately 302 miles of untreated (i.e., centerline only, centerline and shoulder, or shoulder only rectangular rumble strips) on rural two-lane divided roads. Overall, the results of the models indicated no significant differences in crash rates between rural two-lane undivided roads with sinusoidal rumble strips, and rural two-lane undivided roads with rectangular rumble strips.

The objective of this study was to evaluate the change in crash frequency or severity of varying lateral offset distances of high-tension cable median barriers in Minnesota. Crash and traffic data from 2016 to present were conjoined with roadway cross-sectional data and high-tension cable barrier locations to perform cross-sectional analyses for interstate segments equal to or greater than 0.05-miles long. Negative binomial regression models were estimated and used to develop crash modification factors (CMFs) for the following crash types: total crashes, target crashes, and barrier crashes. In addition to evaluating the impact of lateral offset, a naïve before-after evaluation was completed to quantify the impact of installing high-tension cable barrier.

This evaluation determined the change in crash frequency; type or severity associated with longitudinal rectangular rumble strips on rural two-lane undivided and rural four-lane divided Minnesota roadways constructed between 2012 and 2018. Crash Modification Factors (CMFs) were estimated using cross-sectional analysis to compare crash experience of locations with (i.e.; centerline only; centerline + shoulder; or shoulder only) and without rectangular rumble strips. The cross-sectional analysis matched sites with and without rumble strips using matched-pair comparisons. Negative Binomial (NB) or Poisson log-linear regression models were used to model the crashes at all treatment and non-treatment sites. There was a total of approximately 1;200 miles of treated (i.e.; centerline only; centerline + shoulder; or shoulder only) and untreated sites on rural two-lane roads and approximately 35 miles of treated (i.e.; shoulder rumble strips) and untreated sites on rural four-lane divided roads. On rural two-lane undivided roads; the CMF for centerline + shoulder rumble strips was 0.73 for all crashes; shoulder only rumble strips had a CMF of 0.68 for all crashes. For single vehicle run-off-the-road crashes on rural two-lane highways; the CMF for rumble strips on the centerline + shoulder was 0.68; the CMF for shoulder only rumble strips was 0.76. The CMF for head-on; or opposite direction sideswipe crashes on rural two-lane roads with centerline and shoulder rumble strips was 0.64. On rural four-lane divided roads; the CMF for shoulder rumble strips for all crashes was 0.66 and 0.40 for single vehicle run-off-the-road crashes.