This study has been motivated by the lack of research in the substantial and grow1ng refurbishment sector of the United Kingdom construction industry. The reasons for inexplicably higher variability in tender bids for refurbishment work when compared to new work was investigated. This phenomenon had led to a claim by a large body of contractors that they were exposed to higher risks in such projects. The research involved the joint support of a confederation of contractors and a large number of individual contractors. The methodology adopted consisted of structured interviews with individual contractors as a means of obtaining basic information on the refurbishment industry and the particular problems that contractors have in tendering for such work. This qualitative information was augmented by quantitative information derived from measuring the bidding performance of contractors utilising two unique interactive data bases created from primary data obtained from both sets of collaborators. The first data base consisting of tender bids for a very large sample of refurbishment projects, providing global variability measures which served as indicators of the risk and uncertainty in the pr1ce prediction process. The second data base consisted of detailed tender build-ups of competing contractors for a smaller sample of refurbishment projects drawn from the maln sample; variability measures in this case isolated the main components of risk in the tender bids. The most variable main cost component in the tender was the Priceable Building Work, followed by Preliminaries, Mark-up, Dayworks and Fixed Price Allowances. The variability in Priceable Building Work and Preliminaries is founded ln poor and unsuitable tender documentation and the high reliance upon domestic sub-contractors quotations included in the tender. Variability ln mark-up has its origins ln the different perceptions of risk and in the competitive mix in tendering, in terms of specialism and size of competitors. The investigation into the confidential cost and price information within this study has also provided insights into comparative pricing methods and bidding strategies which have never previously been obtainable. It has at the same time, provided answers to many of the unresolved questions overhanging past research studies in estimating and tendering. The study concludes with a proposal for an optimal method of allocating or sharing the risk associated with the uncertain nature and unpredictable costs of refurbishment work. It also draws attention to the fact that objective statistical measurements of variability for the purposes of determining risk and uncertainty can only provide part of the information. There is a need to conduct behavioural studies on the perceptions and evaluation of the risks in estimating and tendering for construction work.