Modelling construction durations for public housing projects

Construction researchers and industry practitioners have recently recognized the need for developing best practice benchmark measures of construction time performance (CTP) for use by clients and contractors, in their attempts to enhance construction output. The past thirty years have seen a growing surge of international attention and discussions regarding CTP of building projects. In parallel, the demand for residential construction has been continually increasing in Hong Kong since the 1950s; while an accelerated demand has been envisaged in the next few years up to 2006. This thesis presents the findings of the three phases of a research programme which initially aimed at identifying the critical factors that affect construction durations of projects in Hong Kong. The first phase explores the effects of project-scope factors and non-project-scope factors such as productivity levels on construction duration based on a questionnaire survey and a case study of a building project. The second phase determines and compares the 10 most significant factors causing project delays as perceived by the clients, consultants and contractors, as derived from another questionnaire survey. The third phase formulates benchmark measures of standards/ norms for overall construction periods of public housing projects by modelling the durations of the primary work packages in the building process, i.e. piling, pile caps/raft, superstructure, E&M services, finishes and their respective sequential start-start lagtimes, in terms of the identified sets of critical factors (significant variables). This also enables a comparison of CTP of individual projects against world best practice. The outcomes from the first phase and second phase provide a sound basis for formulating the third phase of the study. Data from 71 case studies of standard 'Harmony' and 'New Cruciform' Type public housing blocks were obtained in the third phase, through postal questionnaires and semi-structured interviews with both the client (Hong Kong Housing Authority) and their registered building contractors in 1996. Multiple Linear Regression, together with ancillary statistical procedures is applied to analyze this project data and to derive the three sets of construction time prediction models, corresponding to three groupings of standard block types. The predictive accuracy of the regression models is evaluated using data from a further set of projects of similar types, and their relevance and practicality are verified by conducting interviews and interactions with the intended senior industry users. The test results indicate that the models can be reliable tools that could contribute significantly to the planning of measures to meet the accelerated demand for public housing units in the next few years. The potential uses and shortcomings of the models are presented and discussed. The significant variables identified through the regression models lead to recommended strategies for reducing construction durations in these projects. Current strategies and the proposed improvements adopted by the client are described and compared with the recommended strategies in the context of the findings of this study. The research methodology developed for this investigation could well be extended to similar studies and duration forecasts in other sub-sectors in Hong Kong, as well as in other countries.