8Drawing Valid Conclusions From Numbers
Invalid data and poor statistical methods can lead to bad decisions! There are many ways for an engineering project manager to make mistakes, but one of the most common and most insidious is through making logical and procedural mistakes that cause us to draw erroneous and invalid conclusions from quantifiable data, and as a result, making poor data‐based decisions. As engineers, we measure things, and then we often make decisions based on those numbers. For example, we predict when our project will be done, how much it will cost when it is done, and what the technical capabilities of our product will be (e.g. how far will our new airplane be able to fly safely without refueling). And we use those data to make decisions for our project. Whenever we use numbers, however, there is a chance for error: our measurements always involve uncertainties, a particular assumption is only true under certain circumstances, we may not have collected appropriate samples, and so forth. In this chapter, I show you the most common ways that we undermine our own credibility through poor data collection, errors in logic, procedural mistakes, weak statistics, and other errors, and how you can instead use valid methods and strong statistics so as to create credible predictions for all of our project management roles and measures.
8.1 In Engineering, We Must Make Measurements
We are engineers … and, as part of our job, we routinely measure things. Qualitative ...