True Confidence Level of Real-World Data: implications of non-probabilistic data on decision-making and research

Abstract

Non-probabilistic data (in other words, collected without true randomization) is the most common type in real-world scenarios, but few, if any, resources are available to assess their confidence level as statistics focus on probabilistic (random) samples. This book presents a method to calculate it and discusses implications for decision-making and research. In addition to explaining the application and the theory behind it, many examples of implementation are offered, from spreadsheets to various programming languages, which are also available for download. This book proposes an approach to determine the confidence level of non-probabilistic data. It is intended to be as objective and applied as possible, without losing the theoretical coherence and fundaments. It aims to be as easy as possible to read even for those with less or no statistical background.