Classifying the aptitude of Civil Engineering Project Managers using Multivariate Discriminant Analysis

Civil Engineering Project Managers (CEPMs) plan, direct and coordinate a wide variety of construction projects, including buildings of all types of residential, commercial, and industrial structures, roads, bridges, wastewater treatment plants, schools and hospitals. CEPMs may supervise a whole civil engineering project or just part of a project and although CEPMs usually play no direct role in the actual construction of a structure, they typically schedule and coordinate all design and construction processes, including the selection, hiring and oversight of specialty trade contractors. CEPMs therefore, supervise the construction process from the conceptual development stage through to final construction whilst simultaneously meeting time, quality, economic, environmental and health and safety performance criteria as defined by the client. CEPMs evaluate and help determine appropriate construction delivery systems and the most cost-effective plan and schedule for completing the project. They divide all required construction site activities into logical steps, budgeting the time required to meet established deadline. Based upon the massive duties and responsibilities that CEPMs characteristically undertake before and throughout the whole construction cycle, this research has taken place. The predominant aim of this research is to develop a new mathematical model that can be used for classifying the aptitude (capability) of the Civil Engineering Project Manager (CEPM). Although various types of multi-variate analysis were considered, the technique known as Multivariate Discriminant Analysis (MDA) was used. This thesis reported that it can be hypothesised that five variables are the best predictor variables to classifying the CEPMs' aptitude. These variables are B v4 (Team-based bonus for timely project completion); ETv2 (specific vocational background); ETv3 (practical building knowledge and experience); PLv2 (recognition of the interdependence of the project activities) and WSv5 (An ability to ensure the health, safety and welfare of others on site). Findings revealed that the Education and Training Factor (ETv) had the highest weighting of the seven factors with 17%. This was followed by Planning Factor (16%); Quality Control Factors (15%); Management Factor (15%); Motivational Factor (14%); Personal Factor (13%); and Work Situational Factor (11%). The model performance showed that CEPMs were classified accurately by the model at an overall accuracy of 85%. Poor CEPMs were classified at 83% accuracy. The average CEPMs were classified at 88% accuracy and good CEPMs classified at 86% accuracy. Quantitative and qualitative approaches were carried out in order to determine the validity of the research findings. Within the first approach (quantitative approach), two validation procedures were investigated. The first approach was the utilisation of a procedure wherein statistical methods were used to derive values to measure the validity of the developed model using the 25% percent hold-out sample data. Then, 25% percent results were compared with the results obtained from the 75% percent sample. The second quantitative validation approach used a one-sample t-test procedure to test whether there was any difference between means of the actual and predicted classification values obtained from the 25% percent hold-out sample models. Because results for both (75% and 25%) samples are highly similar, it could be concluded that models developed were valid. The results from the second approach showed that at the 95% percent confidence interval, the p-value (sig. e tailed) associated with the t-statistics values in each of the classification models showed that there in no significant difference between the predicted and actual groups classification. Finally, a qualitative approach to validation was undertaken using the collective comments and feedback of participating construction and civil engineering professionals. The feedback and comments were generally positive and there was a general consensus that the findings were robust.