Role of Statistics and Engineering Judgment in Developing Optimized Time-Cost Relationship Models

Abstract

When estimating the duration of the construction stage during the design stages of construction projects, empirical models are often used as a substitute for more accurate estimates based on detailed scheduling. These models, traditionally designated as time-cost relationships (TCR), derive from regressions relating the duration of concluded construction projects with characteristics that can be easily foreseen during the planning and design stages—both quantitative (e.g., cost, gross floor area, number of floors) and qualitative (e.g., type of contract). This paper reviews the approaches that have been adopted to develop the TCR models and discusses the complementary roles that statistics and engineering judgment can play towards their optimization and enhanced accuracy. This discussion includes the statistical and practical implications of (1) the quantitative independent variables used, (2) the qualitative independent variables used, and (3) the mathematical structure used. The paper aims at demonstrating the importance of sound statistics and highlights some of the most common mistakes. Additionally, it provides illustrative examples to show the complementary role that engineering judgment should have in the process of developing more robust, reliable, and accurate statistical-based TCR.