Book Review: Making Sense of Data

Book Review: Making Sense of Data
Photo by Joshua Sortino on Unsplash

There are many textbooks available to provide an introduction to epidemiology and data analysis. As I was preparing for my qualifying exams for a PhD in epidemiology, I received a recommendation to explore the book Making Sense of Data from a professor at my university.

This book is comprehensive in covering the myriad of topics within epidemiology, including but not limited to reading and reporting measurements of associations and graphs, different types of biases, study designs, and causality. In a pragmatic manner, the authors effectively write a textbook that is an easy and engaging read. For example, they beautifully describe the concept of internal validity in three different dimensions: predictive, construct, and content. Elevating the discussion of internal validity to a place of critical thinking that is not solely dependent on mathematics and statistical testing, they provide many doorways through which the reader can engage with and better understand this fundamental concept. Similarly, they do an excellent job at delineating the difference between the strength of a measure of association and a significance test throughout the book. In many ways, this book could be your own independent study, as it has examples, practice problems, and solutions that are explained. 

However, while this book is an easy read that offers unique insights and explanations on many concepts, especially bias, there are also many areas of concern. he authors explicitly state that they do not use as precise a definition of “rate” as most epidemiology textbooks,1,2 and their imprecise language impedes their ability to fully describe the nuanced and important differences in various associations such as risks and rates. Moreover, while mathematically sound, the authors make bold statements that do not, in fact, aid the reader in making sense of data. For example, when discussing the difference between an exposure and disease odds ratio, the authors write, “The answer [the OR] is the same, regardless of whichever way the calculation is done; thus it becomes unnecessary to distinguish between the disease and exposure odds ratio, and we can just refer to the ‘odds ratio’ or ‘relative odds.’” There are many such opportunities to refine their prose with more precision to make their text a more accurate and reliable source. 

Overall, this book is an easy and independent read that provides a high-level overview of epidemiological techniques and public health research methods. The concepts are introduced and explained in an interactive and accessible way. Nonetheless, be wary of the exact definitions and specifics provided, as they are not always aligned with current thinking in the field. I would strongly caution against using this as the only source to study epidemiology. However, for someone new to the field, it offers an engaging and clear framework for understanding the landscape of epidemiology, and, for a more advanced learner, it is a good text to review and challenge your more nuanced understanding of the field.

Josh Yudkin

Leave a Reply