Discriminant function analysis for categorization of best practices

Abstract

Many studies reveal the positive impacts of best practice use on overall project performance, resulting in a consensus of opinion in the industry that implementation of certain practices leads to improvement. Yet there have been no definitive studies reporting in a quantitative manner, the impact of best practices on different project objectives. This research seeks to suggest possible ways to improve two different aspects of project performance: cost and schedule, using suitable best practices. In order to achieve this, the study develops project performance classification models using multiple discriminant function analyses that divide project cost and schedule performance into four groups. The study will examine the best practices that discriminate the most among these four groups. These results are then summarized into a best practice uses categorization for project cost and schedule performance. The Construction Industry Institute Benchmarking and Metrics database, containing almost 1000 projects, provided the basis for development of the vii models and supplied data to perform the analyses. The study will also present a software application based on the results of discriminant function analyses that predicts project cost and schedule performance and suggests greater implementation of critical practices in order to improve project cost and schedule performance. Finally, the study will close by presenting an overall project performance improvement process, using the developed software application. Conclusions and recommendations are provided