Discriminant analysis model for predicting contractor performance in Hong Kong
thesisposted on 26.10.2010, 08:42 by Chi Ming Tam
This thesis describes the development of an operational research model for the identification of determinating variables and prediction of contractor performance in Hong Kong. The mathematical technique used is the Discriminant Analysis approach. The model is also verified with two other analyses Multiple Regression Analysis and Unidimensional Scaling Analysis. One of the aims of the research is to betray the underlying factors that influence contractor performance which are measured in the clients' point of view. The second aim is to develop an accurate model for predicting contractor performance used by clients in vetting contractors. All aspects of the model's development are described, including the quantification of the variables, data collection, analysis of the model results, verification of the model results with other models and testing the model using independent data. Further, the variables adopted in the model are compared with the actual practices in Hong Kong. The predictive model produced by the study is made up of six variables measuring the three dimensions namely the inherent characteristics of the project, the contractor's internal attributes and the external influence of the project team, including the complexity of the project, the working experience of the project leaders, the percentage of professionally qualified staff in the company, the past performance of the contractor, the origin of the company and the architect's or client's supervision and control on the quality of work and work progress. However, the developed ndels should only be used as part of an assessment process and with caution as there are other unpredictable factors which are not able to quantify and include in the model such as the changing of the ccmpany structure and straty, change in management quality, profitability and the happening of overtrading. Nevertheless, the use of the model to exclude cximpanies fran tender lists could accelerate the contractor selection process and spare niore time for clients to concentrate on more important issues.
- Architecture, Building and Civil Engineering