This book is hard to place. As the title suggests, it’s about how to design research studies in the social sciences. However, it relies on more-or-less mathematical modeling to define how a study should proceed. The book is filled with mathematical notation – not as much to derive theorems but to give the thinking precision. It’s very abstract, perhaps too abstract for more practical needs. I’m not sure people with limited experience, like me or like undergraduates, can benefit from this work; graduate students in the social sciences probably would, though.
It honestly could use a bit of humanism in the writing. As a data person with software interests and a deep love for math – i.e., someone not trained in the social sciences – I found the mathematics a bit off-putting and pretentious. Some things could be said better with English instead of in symbolic logic. In reading this work, I wanted to know the human side of how to evaluate a study, but I often needed to reverse-translate from mathematical notation to arrive at the human component. Better writing could have helped.
Besides a mathematical orientation towards the social sciences, this book also uniquely and extensively explores the (free) statistics package R. R is wonderful, and I love how so many people have extended it to address novel situations. The authors demonstrate how R can be used to design and redesign a study so that better decisions can be made, like cost, the size of a study, and model evaluation. How they used R is simply really cool, and the expressiveness of their R package impressed me. I will certainly use their packages.
Therefore, I leave this book with mixed reviews. I read this work to try to understand how to undertake a particular project more effectively. I gathered some thoughts about this small observational study and paid close attention to their code library on casual observational designs. But this book did not hit the sweet spot for my needs in assessing practices among biomedical researchers. Perhaps I should stick more closely to biostatistics in my “native tongue” instead of statistics in the social sciences…
Research Design in the Social Sciences: Declaration, Diagnosis, and Redesign
By Graeme Blair, Alexander Coppock & Macartan Humphreys
Copyright (c) 2023
Princeton University Press
ISBN13 9780691199573
Page Count: 380
Genre: Research, Statistics
www.amazon.com