Statistical analysis is one of the key parts of engineering studies and management by researchers and highway/ other agencies. It is common to use regression analysis to develop a relationship between a response and one or more regressor variables. However, in many instances, correlation between different variables are suggested and used for the prediction of an event without proper evaluation/interpretation of the developed models. This may lead to a gross error, particularly in a time when many senior engineers are retiring with limited skilled successors. The issue is further complicated by the fact that engineers receive limited exposure to statistics during their undergraduate programs. This results in learning on the job which can result in missing some key aspects to statistical analysis. For example, the coefficient of determination (R2) is used by many people to evaluate the adequacy of the developed model which can also be meaningless in some circumstances. Points of intersection of two or more trend lines are sometimes taken as the optimum value of various criteria without considering practical or engineering aspects. This paper has focused on some important aspects of statistical analysis and showed through various pavement engineering/ management models how a model and resulting conclusion can be misleading. In short, proper evaluation of the causal relationship among variables and interpretation of the model are important to support its usefulness as well as to draw meaningful conclusions from it. Finally, a simple and concise approach is presented to aid researchers and practitioners with regression modeling.
Abstract