Statistical Challenges presented by the SPS-2 Experiment and Data

Abstract

The challenges presented by the Specific Pavement Studies-2 (SPS-2) experiment and data for statistical analyses are presented in this paper. Experiment-related issues included (i) wide variation in traffic levels and ages among the fourteen test sites; and (ii) unbalanced distribution of test sites among climatic zones and subgrade types. The challenges presented by the data included (i) outlier issues, (ii) quantification of performance, and (iii) the lack of measurable distresses due to the “young” age of test sections. In this paper the importance of selecting appropriate data analysis methods for obtaining reliable results is illustrated. Though most of the methods that were applied for the analysis are wellestablished, the choice of the methods (magnitude- versus frequency-based) was driven by the extent and occurrence of distresses. Based on the current status of the data, frequency-based methods such as Linear Discriminant Analysis (LDA) and Binary Logistic Regression (BLR) lend themselves well to explaining trends associated with occurrence of cracking; whereas a magnitude based method like analysis of variance (ANOVA) is appropriate for evaluating joint faulting and change in IRI.