Translating Statistics to Make Decisions: A Guide for the Non-Statistician

by Victoria Cox

Examine and solve the common misconceptions and fallacies that non-statisticians bring to their interpretation of statistical results. Explore the many pitfalls that non-statisticians—and also statisticians who present statistical reports to non-statisticians—must avoid if statistical results are to be correctly used for evidence-based business decision making.

Victoria Cox, senior statistician at the United Kingdom’s Defence Science and Technology Laboratory (Dstl), distills the lessons of her long experience presenting the actionable results of complex statistical studies to users of widely varying statistical sophistication across many disciplines.

The author shows how faulty statistical reasoning often undermines the utility of statistical results even among those with advanced technical training. Translating Statistics teaches statistically naive readers enough about statistical questions, methods, models, assumptions, and statements that they will be able to extract the practical message from statistical reports and better constrain what conclusions cannot be made from the results. To non-statisticians with some statistical training, this book offers brush-ups, reminders, and tips for the proper use of statistics and solutions to common errors. To fellow statisticians, the author demonstrates how to present statistical output to non-statisticians to ensure that the statistical results are correctly understood and properly applied to real-world tasks and decisions. The book avoids algebra and proofs, but it does supply code written in R for those readers who are motivated to work out examples.

Pointing along the way to instructive examples of statistics gone awry, Translating Statistics walks readers through the typical course of a statistical study, progressing from the experimental design stage through the data collection process, exploratory data analysis, descriptive statistics, uncertainty, hypothesis testing, statistical modelling and multivariate methods, to graphs suitable for final presentation. The steady focus throughout the book is on how to turn the mathematical artefacts and specialist jargon that are second nature to statisticians into plain English for corporate customers and stakeholders. The final chapter neatly summarizes the book’s lessons and insights for accurately communicating statistical reports to the non-statisticians who commission and act on them.

Readers will:

• Recognize and avoid common errors and misconceptions that cause statistical studies to be misinterpreted and misused by non-statisticians in organizational settings.
• Gain a practical understanding of the methods, processes, capabilities, and caveats of statistical studies to improve the application of statistical data to business decisions.
• See how to code statistical solutions in R.

499 pages, Kindle Edition

Published March 10, 2017

About the author

Victoria is a principal statistician at Dstl (Defence Science and Technology Laboratory) where she provides deep technical expert knowledge across a broad number of disciplines.

Her emphasis is understanding and shaping stakeholder requirements, designing the trials and research to obtain maximum information with limited resources, and translating complex statistical jargon to answer key questions in an accessible manner for non-specialist audiences.

Victoria provides consultancy and review for Design of Experiments (DoE) and statistical modelling and analysis across numerous domains including Chemical, Biological, and Radiological (CBR), Platform Systems, Counter Terrorism and Security, and Cyber and Information Systems. She works with civil servants, military, industry, and academia; both nationally and internationally.

She is also a chartered statistician through the Royal Statistical Society, a member of the Institute of Mathematics and its Applications, has received a Chief Scientific Advisor (CSA) commendation for work during the nerve agent poisoning, and is a STEM ambassador.

Victoria obtained her Mathematics with Study in Europe degree from the University of Sheffield in 2009. The third year of which was spent studying statistics at ENSAI (l'école nationale de la statistique et de l'analyse de l'information) in Rennes, France.