Introduction

The application stories selected for this chapter are intended to demonstrate the broad usefulness of our methodology, across a range of models and situations. The confidence distribution machinery allows the practitioner to compute and exhibit a full (epistemic) probability distribution for the focus parameter in question (without having to start with a prior).

We start with a reasonably simple story about golf putting, with data from top tournaments. A model built for the purpose does a better job than with logistic regression, allowing inference for various relevant parameters. We then turn to an analysis of parameters related to a population of bowhead whales, from abundance to growth and mortality rates. Next we offer a different analysis of data used by C. Sims in his acceptance speech when receiving his 2011 ‘Nobel Prize of Economics’. Sims used a Bayesian setup with noninformative priors, whereas we reach rather different conclusions for the main parameter of interest via confidence analysis (using the same data and the same model).

Jumping from prewar US economics to sports, we demonstrate next that the Olympic 1000-metre speedskating event is unfair and needs a different design. The Olympic unfairness parameter is of the size of 0.15 second, more than enough in that sport for medals to change necks. Turning to Norwegian income distribution we investigate aspects of the Gini index relevant for understanding how the so-called Nordic model works in Scandinavian societies. Then we provide an optimal inference analysis of the odds ratio across a range of 2×2 tables, for a certain application that has involved not merely losses of lives and billions of dollars and legal mass action, but also statistical controversy, as it has not been clear how to give proper meta-analysis of such tables when there are many zeroes. Our final application story pertains to understanding patterns of Google Scholar citations, of interest in the growing industry of bibliometry, citographics and culturomics.