Primary Author: Mr. Thiago Henriqe Rezende Medeiros and Mr. Rafael Gonalves Monteiro
Audience Focus: Intermediate
Application Type: Research
Venue: 2014 AACE International Annual Meeting, New Orleans, LA, USA
Abstract: Using averages to set parameters of interest is a quite common practice in project management, ranging from simple trade off studies to complex economic and financial models. In this context, the expectation is that whenever arithmetic means values are used as processes inputs, the outputs will be an average of the possible results. In fact, this does not occur due to the so called “Flaw of Averages”. This flaw is often one of the causes of errors on estimated values. In order to evaluate these kinds of problems, this article presents two examples to illustrate the significant impact that this flaw can have in estimating key parameters. For both examples, the authors demonstrate how the pragmatic use of probability distributions can diminish errors and avoid misjudgments. Finally, a survey was performed in Brazilian companies aiming to bring out the use of statistical tools in project environments, to put light on the main concerns for those who do not use such tools and also to support proposals of improvement in performing statistical analyses.
Audience Focus: Intermediate
Application Type: Research
Venue: 2014 AACE International Annual Meeting, New Orleans, LA, USA
Abstract: Using averages to set parameters of interest is a quite common practice in project management, ranging from simple trade off studies to complex economic and financial models. In this context, the expectation is that whenever arithmetic means values are used as processes inputs, the outputs will be an average of the possible results. In fact, this does not occur due to the so called “Flaw of Averages”. This flaw is often one of the causes of errors on estimated values. In order to evaluate these kinds of problems, this article presents two examples to illustrate the significant impact that this flaw can have in estimating key parameters. For both examples, the authors demonstrate how the pragmatic use of probability distributions can diminish errors and avoid misjudgments. Finally, a survey was performed in Brazilian companies aiming to bring out the use of statistical tools in project environments, to put light on the main concerns for those who do not use such tools and also to support proposals of improvement in performing statistical analyses.