Distribution of Bids for Buildings and Roads Contracts

This paper describes the analysis of data relating to bids for 535 contracts to determine the distribution of these bids. The findings are that bids for building contracts can be treated as samples from a normal distribution and that for roads, although the assumption is statistically less valid, it is adequate for practical purposes. The tests for normality used were the studentized range and the Anderson-Darling statistic. Comments on the uses of these distributions in testing for unrealistic bids and for predicting the lowest bid are given.