Multilevel Synthesis

This book aims to show how the multilevel approach successfully overcomes the divisions that emerged during the rise of the social sciences— specifically, here, demography and statistics—from the seventeenth century to the present. We begin by examining how the approach connects different branches of the social sciences that conducted their observation and analysis at different levels or addressed different aspects of human complexity. Next, we describe in greater detail the statistical methods and techniques that it can use to simultaneously identify and measure the effects of the different levels examined on the characteristics studied. To introduce this work, we first provide an overview of the goal of social sciences, before discussing the various levels of aggregation at which they can operate. Social sciences start from the observation of a real-life experience and then seek to structure it according to different fields, which constitute the specific objects of study for each science. As a rule, these objects are defined independently of the scale and levels of aggregation that can be chosen to observe them. For example, the object of demography is the quantitative study of human populations, their variations, and their status, without specifying whether the level chosen is the individual, the family, the population of a town, or that of a country. Likewise, economics studies the production, distribution, and consumption of wealth, without specifying if the analysis is of an individual, a market, a firm or a nation.

From the reviews:

"This book represents a pioneer demographic synthesis of the intricacies and richness of multilevel analysis of retrospective, cross-sectional, longitudinal, or event history data. It should help readers’ better grasp of key features of multilevel regression. Specific linear, nonlinear and survival models and their corresponding mathematical and statistical equations are presented in a straightforward manner by avoiding unnecessary technical details that could sidetrack some readers. This should allow for quick comparisons by readers and help them to conceptualise and understand better multilevel analysis. This book is primarily targeted at social science scientists, students and scholars interested in population investigation. I have no doubt that individuals who are interested in learning multilevel modelling will find this book to be an invaluable asset.

… Courgeau’s volume is a neatly written book on mathematical and statistical applications in population studies that I have no hesitation recommending for advanced training and research in demography as well as for scholars unfamiliar with multilevel modelling. It firmly establishes what we have been advocating in classrooms and research settings, namely that mathematical statistics married with theoretical and empirical representation of population processes is a necessary ingredient for widening the frontiers of demography as a discipline." (Barthelemy Kuate-Defo, University of Montreal, Canada in the European Journal of Population)

"The book displays the following strengths:

1. Consistency: The author has achieved coherence and perfect homogeneity.
2. Mathematical exactness: Each of the numerous models is expressed mathematically; the consequences of each model are clearly presented, also in mathematical language. This removes all ambiguity regarding the substance of the issuesraised.
3. Usefulness: The author uses ten examples drawn from different surveys—in particular, migration surveys—to apply each method and model described. The same "concrete examples" thus recur in several chapters, treated with successive methods. The data tables are provided, as well as the statistical parameters to show how the model fits the data. In every respect, this volume offers a perfect example of what mathematical social science should be: data and questions about them; mathematical formalization; application of the model, with the statistical apparatus required to appreciate the relevance of the theory. In addition, an abundance of very clear charts and tables allows the reader, or the student, to follow the whole process in great detail.
4. Historical aspect: Demography and its methods have a history that is always present in the volume, each step in the development of methods being explained by the previous steps. The last chapter discusses future prospects, notably as regards new data sources to be established.

In its present form, this volume will serve henceforth as the reference book on multilevel analysis […]. As such, it should be found in every library on social science methodology and statistics." (Translated excerpt of a book review (in French) in Mathematics and Social Sciences, by Marc Barbut, Editor)

"On the one hand it a textbook, and on the other it constitutes an assessment of multilevel analysis which tackles one of the major problems of demography and sociology: how analysis at the individual and group levels can be integrated. Indeed one of the aims of this textbook is to demonstrate how a multilevel approach can resolve some of the contradictions that have become apparent in the evolution of demographic thought and achieve a synthesis of the different approaches employed." (Genus, Antonio Golini,Editor)

"This manuscript will yield a consummate book—the outcome of a research project that has matured over many years. In essence, it is both a textbook and an evaluation of multilevel analysis that tackles the most significant challenge of empirical sociology: how to integrate analyses at the individual and group levels.

The author’s theoretical and epistemological approach to this problem is particularly elegant. He takes us back to the foundations of sociology, then leads the discussion up to very recent examples. Rigorous in its arguments, the text manages to be accessible to a newcomer in the field yet is free of the simplifications that would turn away a specialist in social sciences.

The methodological exposition is particularly clear. Numerous examples, recurring throughout the volume, make the reading easier. The formal exposition is accessible to anyone concerned with demography, and the author examines practical difficulties very lucidly with no attempt to conceal them." (Translation excerpt from an anonymous referee’s report (in French)

"A ‘textbook [that] presents a historical panorama of the evolution of demographic thought from its seventeenth-century origins up to the present day’ … . The book’s structure and arguments are entirely appropriate – the reader is gently led towards the case for multilevel modelling in a coherent way. … There is considerable richness in this book that will reward detailed study … ." (Ron Johnston, Regional Studies, Vol. 42 (3), 2008)