CAIRN-INT.INFO : International Edition

1As its title indicates, this collective volume presents a selection of papers from an international conference on the advanced methods for the statistical analysis of longitudinal data known as sequence analysis (SA). In the population sciences, SA is based on the coding of life histories and biographical data (work, conjugal, health) using an alphabet that codes states or events over time. It allows researchers to capture individual trajectories in holistic fashion, uncover characteristics common to groups of trajectories, identify atypical and deviant trajectories, and compare types of recurrence, according to characteristics such as gender, birth cohort, and region.

2The book is aimed at social science researchers with a solid background in statistics who wish to quantitatively analyse trajectories using data collected at the level of individuals, institutions, or companies. It consists of 15 articles, each of which can be read independently of the others. They include theoretical contributions as well as presentations of innovative methods and original applications. The introductory chapter offers a helpful methodological framing as well as a selection of essential references for understanding and applying this family of methods.

3While it stresses the intrinsically exploratory scope of sequence analyses as they have been used to date, the book’s objective is nevertheless to show how combining these methods with probabilistic approaches to longitudinal analysis expands the field of possibilities for analysing the relevant types of data. It is precisely because SA captures trajectories in holistic fashion that it cannot be used on its own to uncover the dynamic processes at work within them.

4The book’s main contribution is its presentation of recent research to this end. Part I (two chapters) offers a theoretical presentation of the links, within population science, between SA, duration models (including survival analyses), multilevel modelling, and network analysis. This presentation is followed by three illustrative case studies combining SA and probabilistic modelling. Part II (three chapters) focuses on the combination of SA with survival analyses, from the simple juxtaposition of descriptive and inferential approaches to the integration of the past trajectory at each time point in duration modelling. Part III (two chapters) presents the theoretical and empirical contributions of network analysis. Part IV (three chapters) presents a set of methodological approaches to analysing the processes at work in the construction of trajectories: the decomposition of trajectories into phases (multiphase trajectories), qualitative comparative analysis of divergences between sequences and bifurcations (critical turning points), and finally the analysis of multidimensional sequence data, combining SA and hidden Markov models. Part V (three chapters) returns to the primary use of SA. It is dedicated to advanced methods for clustering sequences: creating groups of trajectories described by continuous variables (Gaussian mixture model); allowing trajectories to belong to more than one cluster at the same time, to varying degrees (fuzzy clustering); and recalculating dissimilarities between sequences on the basis of their decomposition into binary sequences. The sixth and final part (two chapters) offers synthetic measures of sequence quality, based either on the association of each state (or letter of the alphabet) with the notion of success or failure, or on a scale for the quality of transitions between states, rather than the quality of the states themselves.

5The strength of this collection lies in its presentation of such a diverse set of methods and applications. However, this richness also creates a difficulty: that of the sheer number of concepts and statistical analysis methods that the reader must have at least some familiarity with in order to understand the issues, problems, and technical solutions addressed. However, the index points to some of the key concepts and offers a path to ways for the reader to learn more about them, through the bibliography for the corresponding chapter(s). Finally, with its focus on statistical sequence analyses, the book does not directly address either questions around the construction of data (collection, coding, imputation of missing values) or the available software.

6Nonetheless, this fully open access volume represents a significant contribution to this field of methodological research, which has been receiving increasing attention in France in the last decade.

Uploaded on on 22/07/2020
Distribution électronique pour I.N.E.D © I.N.E.D. Tous droits réservés pour tous pays. Il est interdit, sauf accord préalable et écrit de l’éditeur, de reproduire (notamment par photocopie) partiellement ou totalement le présent article, de le stocker dans une banque de données ou de le communiquer au public sous quelque forme et de quelque manière que ce soit.
Loading... Please wait