01152430

Component

https://doi.org/10.3141/2165-06

http://scholar.google.com...Lord&publication_year=2010

Regression to the mean (RTM) in before-and-after speed data is a purely statistical phenomenon that makes random variation in repeated speed measurements from multiple time points before and after the introduction of an engineering treatment look like a genuine speed change brought about by the engineering treatment. This study shows that an observational before-and-after speed data analysis cannot collect speed measurements without measurement error and cannot be free from RTM bias. To obtain accurate estimates of the magnitude of the mean speed change brought about by an engineering treatment, RTM bias needs to be reduced. This study first uses a graphical method to illustrate the RTM phenomenon and then uses numerical examples (with aggregated speed data) to show how to reduce RTM bias in before-and-after speed data analysis. The numerical examples show that the estimated magnitude of the mean speed change that results from the introduction of an engineering treatment or the amount of uncertainty (measured by the estimated standard error and confidence interval) associated with the mean speed change can be misleading if RTM is not taken into account. The paper concludes with suggestions for more rigorous statistical methods, preferably suited for use with disaggregate speed data, that may help to reduce RTM bias in future speed data analysis.

01220491

10-1209

English

pp 52-58

2010

Transportation Research Record: Journal of the Transportation Research Board

9780309142977

Print

Figures (3) ; References (30) ; Tables (2)

Before and after studies; Bias (Statistics); Highway operations; Regression analysis; Speed data; Traffic engineering; Traffic safety; Traffic speed; Uncertainty

Engineering treatments

Data and Information Technology; Highways; Operations and Traffic Management; Safety and Human Factors; I73: Traffic Control

TRIS, TRB, ATRI

Jan 25 2010 10:32AM

• Combined Effect of Traffic and Geometrics on Rear-End Collision Risk: Driving Simulator Study
• Empirical Investigation of Continuous Logit for Departure Time Choice with Bayesian Methods
• Examination of Methods to Estimate Crash Counts by Collision Type
• Gaussian Processes for Short-Term Traffic Volume Forecasting
• Getis–Ord Spatial Statistics to Identify Hot Spots by Using Incident Management Data
• Integration of Traffic Simulation into Design Visualization: Case Study
• Leverage of Spiral Graph for Transportation System Data Visualization
• Modeling Injury Severity of Multiple Occupants of Vehicles: Copula-Based Multivariate Approach
• Multiple Applications of Multivariate Adaptive Regression Splines Technique to Predict Rear-End Crashes at Unsignalized Intersections
• Spatial Correlation in Multilevel Crash Frequency Models: Effects of Different Neighboring Structures