An Analysis of Estimating Errors on Government Contracts.
Abstract
This thesis considers what measures should be used to evaluate estimating error of government estimates on competitively bid construction contracts. A data set from the U.S. Army Corps of Engineers was subjected to analysis using two measures of accuracy. One objective of this research is to develop mathematical models and find cumulative distributions which may describe estimating practices within the Corps of Engineers. Other objectives are to offer large scale adjustments to compensate for any error present and to give a baseline for future studied and/or attempts to improve estimating accuracy. It is found that there were significant estimating errors in the Corps of Engineer estimates. The variability of these errors is more than expected. The principal conclusions are: to retain as a measure of error, the standard percentage form, i.e., ERROR = ((OBSERVED - PREDICTED)/PREDICTED)*100; and large scale adjustments are not useful. Baselines were established for further research and study of estimating errors. Keywords: Estimating errors; Regression analysis; U.S. Army Corps of Enginneers Data; Cumulative distribution function.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 22, 1985
- Accession Number
- ADA157617
Entities
People
- D. L. Diedrich