Multilevel Modeling in Plain Language

Karen Robson is Assistant Professor in the Odette School of Business at the University of Windsor. She holds a BSc(h) in Psychology from Queen’s University, an MA in Psychology, an MBA and PhD from Simon Fraser University. Karen’s research investigates consumer innovativeness, including how consumers repurpose or use market offerings in ways not intended by the manufacturer, the intellectual property law implications of this practice, as well as the gamification of consumer experiences. A recipient of the Joseph-Armand Bombardier Doctoral Scholarship, her work has appeared in journals such as Information & Management, European Journal of Marketing, MIS Quarterly Executive, Business Horizons, Journal of Marketing Education, Journal of Advertising Research, and Journal of Advertising, among others. She is a member of the Editorial Advisory Board of the Sage Express Case Collection. David Pevalin is Professor in the School of Health and Human Sciences and Dean of Postgraduate Research and Education at the University of Essex. He previously served in the Merchant Navy, the City of London Police and the Royal Hong Kong Police. He studied part time at the University of Hong Kong before graduate studies at the University of Calgary, Canada. He returned to the UK in 1999 as Senior Research Officer at the Institute for Social and Economic Research at the University of Essex and joined his current School in 2003 after obtaining his PhD. He co-authored (with Karen Robson) The Stata Survival Manual (Open University Press), co-edited (with David Rose) The Researcher’s Guide to the National Statistics Socio-economic Classification (Sage), and authored research reports for the Department of Work and Pensions and the Health Development Agency. He has published papers in the Journal of Health and Social Behavior, British Journal of Sociology, Lancet, Public Health, and Housing Studies.

I started to read the book with vivid interest because of the subject that too often does not find enough space in books which provide an overview of the most used statistical methods  leaving out those who are somewhat a little bit more elaborate. After a while I found that I had read many pages, as a story, in a short time, and, rethinking to the title of the book, I remembered there was a part saying "…. In plain language". This is really genuine. The Authors do really introduce the subject in a very friendly way, propose examples which facilitate the reader to better  understand and explain the output of Stata.  I suggest the book both to students and instructors who want a specific text on this subject. On the one hand, students will be not afraid of formula, considering that the book is centred on the understanding of the subjects, on the other hand, instructors will benefit in reviewing the path of the multilevel analysis very quickly. It is a book for those who have some knowledge of statistic but I think that this aspect is definitely clear to the reader. The book is really complete in all the phases of a multilevel analysis, the "plain approach" helps the reader to grasp the idea,  follow the Stata commands and outputs and, finally, to interpret the findings. I think that the Authors were very skillful in preparing this book and added a very useful resource, in particular, for those who use Stata for their analysis.

• Chapter 1: What Is Multilevel Modeling and Why Should I Use It?Mixing levels of analysisTheoretical reasons for multilevel modelingWhat are the advantages of using multilevel models?Statistical reasons for multilevel modelingAssumptions of OLSSoftwareHow this book is organizedChapter 2: Random Intercept Models: When intercepts varyA review of single-level regressionNesting structures in our dataGetting starting with random intercept modelsWhat do our findings mean so far?Changing the grouping to schoolsAdding Level 1 explanatory variablesAdding Level 2 explanatory variablesGroup mean centringInteractionsModel fitWhat about R-squared?R-squared?A further assumption and a short note on random and fixed effectsChapter 3: Random Coefficient Models: When intercepts and coefficients varyGetting started with random coefficient modelsTrying a different random coefficientShrinkageFanning in and fanning outExamining the variancesA dichotomous variable as a random coefficientMore than one random coefficientA note on parsimony and fitting a model with multiple random coefficientsA model with one random and one fixed coefficientAdding Level 2 variablesResidual diagnosticsFirst steps in model-buildingSome tasters of further extensions to our basic modelsWhere to next?Chapter 4: Communicating Results to a Wider AudienceCreating journal-formatted tablesThe fixed part of the modelThe importance of the null modelCentring variablesStata commands to make table-making easierWhat do you talk about?Models with random coefficientsWhat about graphs?Cross-level interactionsParting words