Estimation Biases in Construction Projects: Further Evidence

Construction projects are characterized by their unique and temporary features. Very limited data, if any, would be available and readily transferable for the analysis of subsequent projects. Hence most project analysts would depend on intuitions and gut feelings to make judgment and estimations. This experimental study is a follow-up study on investigating the possible existence of systematic estimation errors (biases in estimations) in an Indonesian context. It is focused on the two suspected types of biases; the anchoring accuracy and overconfidence biases in project time duration estimates. Two groups of estimators (experienced, n = 20 vs. non-experienced, n = 20) were involved in the study. A hypothetical project case based on an actual construction project was developed. The estimators were then individually requested to provide duration estimates (for each project activity and overall) for the project case. The estimates were then compared against the actual duration of the project. The result of suggests that anchoring bias is not statistically observable for both non-experienced and experienced estimators. This study finds that overconfidence bias is identifiable when making the range estimation of the project duration.