The Effect of Pavement Condition Data Sampling on Project Boundary Selection

Pavement condition data are either continuously collected or sampled. In sampling it is assumed that the time and costs of data collection can be reduced and the pavement condition of shorter segments represent the conditions of larger sections. In a study sponsored by the Federal Highway Administration, pavement distress data along several miles of roads were obtained from four State Highway Agencies. Each department collects and stores distress data on a continuous basis for each 0.1 mile section along the network. The continuous data were sampled and the impacts on the accuracy of pavement decisions were analyzed. Results of the analyses are discussed herein. It is shown that, for variable pavement conditions, ten percent sampling leads to inaccurate decisions. Pavement sections in need of repair are ignored whereas healthy sections are selected for repair. The costs incurred due to inaccurate decisions could be much higher than the saving incurred by sampling.