NC Local Safety Partnership Evaluation Methods
Workshop Roadmap Program Background and Overview Crash Data Identifying Potential Treatment Locations Preparing Collision Diagrams Selecting Interventions Evaluation Methods Program Implementation and Discussion 2
Module Objective ■ Understand why evaluation is important ■ Learn about the two basic kinds of evaluation methods ■ Learn how to perform a simple before- after analysis 3
How well has a treatment, project, or group of projects reduced crash frequency or severity? Safety Effectiveness Evaluation Basics Evaluation 4
Safety Effectiveness Evaluation Basics Why do safety effectiveness evaluations? ■ Prove investment effectiveness ■ Demonstrate program value to decision makers (accountability) ■ Contribute new scientific knowledge ■ Improve decisions ■ Optimize future safety investments
Results in Winston-Salem MetricBeforeAfterChange Targeted Crashes8,5024,078-51% Total Crashes18,65915,129-18% Injuries11,0587,578-31% Property Damage$52.3 mil.$41.9 mil.-19% ■ Result of 858 Low Cost Treatments in Winston-Salem over past 25 years 6
■ Goal – Measure true effect of a countermeasure ■ We want to be sure that the observed change is due to the countermeasure alone ■ What other factors could cause the change? Safety Effectiveness Evaluation Basics 7
■ What other factors could cause the change? ■ Other “treatments” at the same time (e.g., primary seat-belt law at the same time as adding a protected left-turn phase to intersections) ■ Changes in AADT ■ Regression to the mean ■ Underlying trends in crashes (e.g., economy- related changes) ■ Others ■ So how do we control for/discount these other “causes”? 8
Two Basic Evaluation Study Designs 1. Before-after studies 2. Cross-sectional studies Choice of method is affected by: Nature of treatment Site type Available data 9
Before-After Study ■ Examines crash data before and after the treatment is installed ■ Types of before-after studies ■ Simple before-after study ■ Does not account for certain biases ■ Before-after study with reference/comparison groups ■ Accounts for changes in volumes and other factors 10
Simple Before-After Study Estimated average expected crash frequency without treatment Measured Is this assumption realistic? TREATMENT ASSUME these are the crashes WITHOUT TREATMENT 11
Before-After Study with Reference/Comparison Group Expected average crash frequency without treatment Measured Expected average crash count without treatment TREATMENT 12
Before-After Study with Reference/Comparison Group Estimated average expected crash frequency without treatment Measured 13
Cross-Sectional Study ■ Compare crash data for sites with and without treatment over same time period 14
Cross-Sectional Study ■ Why do a cross-sectional study? ■ Treatment installation dates unknown ■ Volumes and crash counts in before period unknown 15
Evaluation Study Type Selection Guide Evaluation Method Treatment SitesNontreatment Sites Before Data After Data Before Data After Data Simple Before-AfterXX Before-After Using Reference/Comparison Group XXXX Cross-Sectional StudyXX 16
EXAMPLE SIMPLE BEFORE- AFTER EVALUATION Module 6 – Evaluation Methods 17
Example Simple Before-After Evaluation ■ Traffic Signal Installation in Wake Co. ■ NCDOT Safety Evaluation Unit conducted before-after evaluation ■ Intersection of SR 1004 (East Garner Road) and SR 2555 (Auburn - Knightdale Road) ■ Before: Two-way STOP-controlled ■ After: Traffic signal (actuated) 18
Example Simple Before-After Evaluation 19
Example Simple Before-After Evaluation ■ Signal installed in October 2003 ■ Before period: Nov 1998 – Sept 2003 ■ After period: Dec 2003 – Sept
Example Simple Before-After Evaluation BeforeAfter Percent Reduction (-) Percent Increase (+) Total crashes % Target Crashes345-85% Fatal injury Crashes00N/A Class A injury Crashes20-100% Class B injury Crashes70-100% Class C Injury Crashes94-56% Total Injury Crashes184-78% Total Severity Index % Target Crash Severity Index % Volume8,8009,3006% 21
Example Simple Before-After Evaluation 22
Example Simple Before-After Evaluation 23
EXAMPLES FROM WINSTON- SALEM EVALUATIONS Module 6 – Evaluation Methods 24
Before and After Collision Diagrams 25
26
Summary 27
Graph of Poisson Distribution 28
29
30
31
32
Take Away Messages ■ Evaluation helps justify investments and improves decision making ■ Pros and cons of different types of evaluation ■ Steps for conducting simple before/after evaluation 33