Evaluation of factors affecting time and cost performance in Hong Kong building projects

Time and cost are usually critical to construction clients. Given the many contributory factors, improved quantitative models of time and cost may help clients to predict project outcomes at the outset, and also at different stages of the project life span. These can also help to compare deviations in significant contributory factors, and to suggest corrective actions. Multiple linear regression (MLR) and artificial neural networks (ANN) were applied in developing such quantitative models in a research project based in Hong Kong. A comparative study indicated that ANN had better prediction capabilities than MLR by itself. Significant factors identified through quantitative models developed, indicated that time over‐run levels were mainly governed by non‐procurement related factors (e.g. project characteristics and client/client representative characteristics), while cost over‐run levels were significantly influenced by both procurement and non‐procurement related factors (e.g. project characteristics, client/client representative characteristics and contractual payment modalities). A parallel approach yielded interesting comparisons of the variations of mean time and cost over‐runs, when comparing groups of projects using different procurement sub‐systems, from the Hong Kong sample.