Causal-Comparative Analysis of Accuracy Enhancements

In information technology (IT) infrastructure consulting, the lack of standard project estimation models continues to lead to inaccurate estimates, inconsistent results, and financial losses. In response, the purpose of this study was to evaluate estimating models and support the standardization in the development of estimations during the planning stages. Based on the analysis of project data archives from 2 large IT organizations, a causal-comparative analysis of 3 project estimation methods (i.e. parametric based, a case based, and a mix of the first two models) was accomplished. The research questions focused on whether significant differences existed in the average absolute relative error (ARE) values of the three models and what can be done to improve the project estimation process. ARE values were computed along with appropriate t testing. The results showed that the mean ARE values were lower for the mixed method compared to the parametric and case-based methods.