Evaluation and improvement of craftsman questionnaire in measuring construction productivity

This dissertation is mainly concerned with evaluating and improving the Craftsman Questionnaire. The historical background and current state-of-art of the Craftsman Questionnaire are first reviewed. Its strengths and weakness are thoroughly discussed. Three main drawbacks in current practice are identified. The are: unproven validity, lack of application of inferential statistics, and time consumption in data collection and analysis. To examine the validity, a field investigation is conducted. The validity of Craftsman Questionnaire is empirically verified by comparing its results with the results of Work Sampling which is a proved method. To explore the feasibility of applying inferential statistics to the Craftsman Questionnaire, random and repeated sampling from a census of past data is simulated in a computer. Comparisons among the population and the samples given evidences prove that it is feasible to use inferential statistics in the Craftsman Questionnaire. To speed up the process of data collection and analysis, a new system called Craftsman Questionnaire Sampling which combines both merits of Craftsman Questionnaire and Work Sampling is developed and tested at a large nuclear power plant construction site. This dissertation represents three initial attempts and significant contributions for the construction profession. They are: (1) It is the first formal demonstration of the validity of Craftsman Questionnaire and can clear up the public's suspicion while using Craftsman Questionnaire to collect information for construction management in planning the work improvement program. (2) It is the first attempt to apply inferential statistics to the study of Craftsman Questionnaire. It upgrades the investigated quality of Craftsman Questionnaire in its numerical estimates from the level of descriptive statistics to the level of inferential statistics. It enables the users of Craftsman Questionnaire to measure the precision of their estimates from a sample to a population. (3) It is an initial introduction of Craftsman Questionnaire Sampling which not only saves a great amount of craftsman's time to participate in the survey, but also facilitates the data analysis so that the information related to the administrative constraints which hinder the craftsman's performance can be promptly identified and removed.