-
PDF
- Split View
-
Views
-
Cite
Cite
Emily Eyles, Beyond Multiple Linear Regression: Applied Generalized Linear Models and Multilevel Models in R, Journal of the Royal Statistical Society Series A: Statistics in Society, Volume 185, Issue 2, April 2022, Pages 726–727, https://doi.org/10.1111/rssa.12773
Close - Share Icon Share
There are many textbooks covering basic statistics and multiple linear regression, but there are far fewer which address intermediate concepts such as generalised linear models accessibly. This book is written for intermediate students, particularly undergraduates, but would work for graduate students, or researchers who want to expand or reinforce their knowledge of more sophisticated statistical topics.
The first seven chapters are a helpful review of concepts, such as multiple linear regression, Poisson and logistic regression, and a unifying theory of linear models. The first chapter is a valuable review of the basic concepts a student may learn from a previous regression course. There are four chapters on multilevel models. The multilevel modelling chapters are particularly thorough, and include topics such as data with more than two levels, and applying the linear model concept to multilevel data.
The book uses real case studies with real data, available from GitHub (https://github.com/proback/BeyondMLR). These cases are drawn from a variety of fields, and provide helpful substantive context to concepts. The examples are described with both mathematical equations and R code, which is helpful for understanding the conceptual level and the practical level.
The exercises are also thoughtfully structured, with three exercise types per chapter: conceptual exercises, guided exercises and open-ended exercises. Conceptual exercises help a reader think through problems and data, and deepen theoretical knowledge. Guided exercises are fairly explicit, step-by-step practice exercises using R, with hints. Open-ended exercises are similar to guided ones, but without the step-by-step question structure, and are useful for building confidence in conducting independent analyses.
The only stumbling block is that the solutions manual is incredibly difficult to obtain. I was unable to obtain it via application to the publisher’s sales website: I did not receive any response to my application. While it is a course textbook, and therefore the solutions manual should not be readily available, independent learners should have some simpler way of accessing it.
Overall, a very well-structured book, with all concepts thoroughly described, featuring multiple real world examples. I would highly recommend this for course instructors, students and researchers who require an overview of intermediate to advanced regression concepts.