The Numbers Don’t Mean What You’ve Been Told They Mean: Data Storytelling Without Creating Fiction
We live in an age of “Big Data” and evidence-based decision making, inundated with stories based on numbers. It can be difficult to understand and interpret the figures by looking at highly condensed summaries that have gone through quite a bit of processing and through many hands. One such example is the use of standardized effect sizes and syntheses of meta-analyses in education research, such as in John Hattie’s Visible Learning books. This session offers illustrations of why the use of this approach can easily lead to fictional narratives detached from reality. It presents simple approaches to go back to the original data context and extract actionable insights. Demonstrating what questions to ask when presented with statistical results and understanding the limits and pitfalls of such analyses, the goal is to enhance one’s critical thinking skills with a better understanding of statistics, data science, and the art of making sense and truthful stories out of numbers and figures.
Pierre-Jérôme Bergeron is a private statistical consultant and an adjunct professor at the Department of Mathematics and Statistics at the University of Ottawa. He has over a decade of experience in academic, government and industry work, applying and developing new statistical methods in a variety of fields such as public health and medicine, finance, sports analytics and literature; translating data into actionable insights and meaningful stories. He holds a PhD in statistics from McGill University