Click for next page ( 2


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
1 SUMMARY Accident Modification Factors for Traffic Engineering and ITS Improvements Crash reduction factors (CRFs)--related to accident modification factors (AMFs)-- provide a quick way of estimating crash reductions associated with highway safety improvements. (The remainder of this report will use the term AMF to be consistent with other ongoing NCHRP research in this area.) AMFs are used by many states and local juris- dictions in program planning to make decisions concerning whether to implement a specific treatment and/or to quickly determine the costs and benefits of selected alternatives. AMFs are also key components of the latest safety-estimation tools and procedures, including the Interactive Highway Safety Design Model (IHSDM) and the procedures now being developed for the Highway Safety Manual (HSM). Even though accurate AMFs are critically important to states and municipalities in their attempts to achieve the greatest return on their investment when choosing among safety treatments, there is no accepted standard set of AMFs. This is because the accuracy and reliability of many published AMFs is questionable, and no AMFs exist for many impor- tant safety treatments. This lack of reliability, accuracy, and comprehensiveness has been documented in this study, in prior work, and in ongoing HSM development efforts. The sources of the problem include the lack of AMFs for newer ITS treatments and for common combinations of treatments, the fact that AMFs vary with other factors such as traffic volume, a publication bias that results in publishing only positive findings, and crash migration and spillover effects that result from some treatments but are not accounted for in the AMF. However, the major problems with existing AMFs result from the poor data and poor evaluation methods used in their development. Often, AMFs are based on sim- ple before-after studies of high-crash locations, and the results can be very biased toward overestimating accident reductions. The goals of this study were to (1) produce AMFs for high-priority treatments or treat- ment combinations where none currently exist, and/or (2) to increase the level of predictive certainty for existing AMFs in other high-priority areas. Due to funding limitations, not all possible treatment strategies could be effectively studied within this single project; therefore, the emphasis was on high-priority treatment strategies that are broadly implemented by states and local agencies and that can affect significant numbers of crash-related deaths and injuries. Reliable AMFs, at a minimum, must meet the following criteria: The AMFs are methodologically and statistically valid. The applicability of the AMF (i.e., the subset of crashes, crash locations, or crash conditions to which it is applicable) is known and documented. The AMFs reflect improvements or combinations of improvements that are of interest to DOTs. The AMFs reflect the impact of the improvement on different crash-severity and crash-type categories.

OCR for page 1
2 The AMFs reflect variability by including both the best estimate of the AMF along with some measure of variability (e.g., ranges, confidence intervals, and standard deviation). The AMFs reflect the savings in "total harm" provided by the treatments, where "total harm" is a combination of frequency and severity. The identification and development of AMFs that meet most of the above requirements involved a project effort that included four basic types of analyses: Empirical Bayes (EB) Before-After Evaluation. Before-treatment and after-treatment crash data were acquired for locations where the treatment of interest had been installed. The latest statisti- cal methodologies (i.e., EB) for conducting before-after studies were applied to produce AMFs. Reanalysis of Existing/Supplemental Data. Data from prior before-after evaluations were acquired and reanalyzed using the more rigorous EB methodology. In many cases, supplemen- tal data were acquired to enhance the evaluation. Analysis-Driven Expert Panel. A panel of knowledgeable researchers and practitioners was con- vened to review critical research studies and reach a consensus on AMFs for a given treatment. In some cases, the AMF was developed through further analysis by one of the NCHRP project teams sponsoring the expert panel meeting. Cross-Section Modeling. A new analysis was conducted in which a cross-sectional model was produced and used to derive AMFs for a specific treatment. The project team developed an initial list of 78 treatments deemed to be important in safety decisions. The treatments were divided into four groups: intersection-related, roadway- segment-related, ITS-related, and other. Additional treatments suggested by 34 state DOT users, who responded to the project survey, and further refinements by the project team expanded the list to 100 treatments. A review of the literature discovered no AMFs for 50 of these treatments. Critical literature reviews were then conducted for the remaining 50 treat- ments. Of that group, 20 were felt to have a high or medium-high level of certainty. Summary information for each of these 20 AMFs along with a knowledge matrix for all 100 treatments was published in NCHRP Research Results Digest 299 as an interim product of this research study. Each summary includes the AMF(s), the level of predictive certainty, the study methodology, a description of the sites used in the study, and supplemental comments and footnotes to describe the study results and applicability. The results of this initial effort clearly supported the need for additional research to develop new AMFs and to strengthen those with less than a medium-high level of certainty. The determination of which of the many possible AMFs should be developed or improved was based on the following: Results of the survey of state DOT users, Judgment of the level of predictive certainty of existing AMFs from the literature review, Measure of the crash-related harm that might be affected by the treatment, Whether there was existing ongoing research that might develop an AMF, and Availability of data needed in AMF development or improvement. Priority was given to conducting as many EB evaluations of the high-priority treatments as possible. Based on the funding available and on team knowledge of available data, it was decided that EB analyses would be conducted to develop AMFs for the following high- priority treatments: Installation of a traffic signal at a rural intersection; Conversion of an undivided four-lane road to three lanes including a two-way-left-turn lane (TWLTL)--a "road diet";

OCR for page 1
3 Increasing pavement friction on intersection approaches; Increasing pavement friction on roadway segments; Modification of left-turn signal phase (three combinations); Replacement of 8-in. signal head with 12-in. head; Replacement of single red signal head with dual red signal heads; and Conversion of nighttime flashing operation to stop-and-go operation. The second approach to AMF development/modification involved two analysis-driven expert panels. While earlier project discussions had noted the possibility of expert panels for AMFs related to specific focus areas (e.g., roadside crashes and pedestrian treatments), it was decided by the team and the oversight panel that a more critical need was to assist the project teams from NCHRP Projects 17-26 and 17-29 in developing AMFs needed for the Highway Safety Manual safety prediction tools these teams were developing for urban/suburban arteri- als and rural multilane highways. Working with the research teams from those two projects, this research team identified and recruited expert panel members; developed a listing of potential treatments for study; compiled and distributed copies of relevant research reports to the panel; and arranged and hosted the panels. Before each meeting, the panels prioritized the potential list of treatments in order to ensure that the most important ones were discussed, and each panel member was assigned one-third of the high-priority issues and asked to be prepared to present their opin- ions and lead the discussion. Following the meeting, the project team funded and conducted additional limited analyses and documented the findings of each expert panel in a report. The AMFs developed through consensus from a review of prior research studies or from further analysis recommended by the panels include the following: Add exclusive left-turn lane, Add exclusive right-turn lane, Prohibit right turn on red, Modify left-turn signal phase, Add intersection lighting, Add two-way left-turn lane, Change roadside slope, Add/remove on-street parking, Add segment lighting, and Reduce mean travel speed (not treatment-specific). In summary, this project has verified, modified, or developed 35 AMFs that are felt to be of high or medium-high quality. These have been documented in formats that are usable by both practitioners and researchers. These AMFs are the primary project outputs. However, the project has also documented a process that can be used with future analysis-driven expert panels and the detailed discussions of the two expert panels that were part of this effort. This material should be helpful in future efforts to develop or improve AMFs for treatments for which no AMF could be developed here. Finally, the project developed and documented a procedure for ranking needed AMF research, a procedure incorporating not only state DOT user and researcher opinions and knowledge of the quality of AMFs in the published literature, but also a method for estimating the crash-related harm that might be affected by each treatment. An approach combining these factors could also be used in more global efforts to prioritize roadway safety research needs in general.