United States Department of Transportation
i
Developing Equivalence Tools to Control Quality of Transportation Infrastructure Asset Management Data
-
2021-10-28
[PDF-3.20 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Corporate Publisher:
-
Abstract:Transportation infrastructure monitoring requires substantial investments of time and money. New technologies are frequently developed and deployed, but the serviceability lives of transportation infrastructure assets are long, and new technologies produce challenges. One promising formulation of statistical tools for informing these decisions can be found in biostatistics, in the formulation of statistical tests developed for the acceptance of generic pharmaceuticals. Unlike statistical tests formulated to seek significant differences between methods or for variables significantly affecting outcomes, these tests are formulated to assess equivalence or noninferiority between methods, thus also holding promise for assessing if new equipment or vendors are equivalent (or noninferior) to current accepted standards. Recently, these methods have also been proposed for use in the social sciences and in quality and manufacturing engineering. In this project, statistical equivalence methods are assessed as to suitability and formulated in the context of three elements of condition assessment. This report discusses the background of equivalence testing, pavement cracking data, application scenarios using pavement roughness data, traffic-speed deflection device evaluation methods, and examples with simple MATLAB computational code for supporting the calculations. The two one-sided t-test (TOST) methodology is recommended as the easiest and most practical to apply for infrastructure asset management condition data. The biggest challenges are in making informed selections of acceptable statistical risks and data tolerances. While this study provides illustrations and some suggestions, the determinations must ultimately be driven by the limits of available technologies, the sensitivity of the agency-specific decision support systems, and the costs of collecting needed data to reduce statistical uncertainty.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
