Analyzing the relationship between trajectories and predictors
Matthias Studer, doctoral student at the Department of Economics, University of Geneva, a member of LIVES, will speak on Tuesday, January 24th, 2012, at 12:30 at the premises of the NCCR LIVES, Vidy Building, Room 531.
- Meeting "Methods and Research": dispersion analysis of sequences of states, trajectories and links between variables
- Organized by Prof. André Berchtold, senior lecturer and fellow at PRN LIVE, a joint seminar between the Faculty of Social and Political Sciences (SSP) of the University of Lausanne and the Swiss Centre of Expertise in the Social Sciences (FORS)
The sequence analysis has become one of the preferred methods to analyze the trajectories. It allows to study in a holistic recurring patterns and taking into account the multiplicity of possible states. Technically, this is based on a measure of distance between trajectories, which allows to compare them. In practice, these distances are often used to construct a typology of trajectories and trajectories identify types.
Beyond the descriptive approach, the focus is typically to identify factors that influence the construction of the path. To do this, it is customary to relate the types obtained with other factors of interest, such as gender, using logistic regression or tests of association. However, focusing on the types of trajectories, we lose information, which can lead to misleading conclusions. We present a set of methods that can analyze the relationship between the sequences on the one hand and one or more explanatory factors on the other. Originally used in ecology, these methods rely on the definition of a measure of dispersion of the sequences and on a generalization of the principles of analysis of variance (ANOVA) for all types of dissimilarities. Conceptually, these methods allow a paradigm shift. Rather than relying on the search for models of trajectories, we consider that they are inserted in multiple contexts that influence - in their own way - the construction of a trajectory. These methods thus complement traditional sequence analysis, primarily exploratory, with a confirmatory approach.
Matthias Studer holds a DAS in sociology from the University of Geneva. He is currently a doctoral student in the Department of Economics, University of Geneva. Inside NCCR LIVES, he participates to IP6 and IP14.