Further Education:

June - 2010           Panel Data Analysis Seminar
Quebec Inter-university Centre for Social Statistics, McGill University, Canada
Instructor: Danielle Forest
Topics Covered:  This intensive training session deals with longitudinal (panel) data analysis. The course will cover topics in static and dynamic model specifications, as well as topics in duration data analysis. Emphasis will be placed both on theory and computer application of the methods using data from the Survey of Income and Labour Dynamics (SLID). During the course, we will gradually build an empirical specification that models four major decisions for women: (1) Working; (2) Marriage/cohabiting; (3) Fertility; and (4) Receiving social assistance. The decisions will in the end be modeled jointly and the aim is to estimate how these decisions are related and assess the importance of one decision on the other decisions.

April - 2010          Mixed Models Analysis Seminar
Quebec Inter-university Centre for Social Statistics, University of Montreal, Canada
Instructor: Danielle Forest
Topics Covered:  Cet atelier porte sur l’estimation de modèles mixtes à l’aide des progiciels statistiques SAS et STATA. Généralisation des modèles standards, les modèles mixtes permettent de modéliser la moyenne des données ainsi que leurs variances et leurs covariances. Modèles hiérarchiques (Bryk & Raudenbush, 1992), modèles avec coefficients aléatoires ou paramètres aléatoires (Rosenberg, 1983) et modèles multiniveaux (Mason et. Al., 1983) sont autant d’appellations des modèles mixtes (Goldstein, 1986). Ce type de modèle comporte à la fois une partie fixe, la moyenne, qui est la même pour toutes les données, et une partie aléatoire, la variance-covariance, qui varie selon les données, d’où l’appellation de « modèles mixtes ». Nous avons recours à ce type de modélisation en raison de la structure hiérarchique des données qui contrevient à l’une des hypothèses du modèle standard, celle de l’indépendance des observations.

February - 2010    Structural Equation Modeling Seminar
Quebec Inter-university Centre for Social Statistics, McGill University, Canada
Instructor: Rex Kline, Concordia University
Topics Covered:  This five-day seminar dealt with the principles, assumptions, strengths, limitations, and applications of structural equation modeling (SEM). Basic SEM techniques, including path analysis, confirmatory factor analysis (CFA), and full “LISREL” (structural-regression) models, were covered.

August - 2009       Brain Imaging Summer School
Quebec Bio-Imaging Network, McGill University, Montreal, Quebec, Canada.
Topics Covered:  Basics of MRI of the Brain, fMRI experimental design, analysis and interpretation, Real-time fMRI, Image quality and stability in fMRI, Modeling brain connectivity from functional imagining, Functional connectivity methods and applications.

June - 2009           Introduction to Meta-Analysis
Online course offered at www.statistics.com
Instructor: Michael Borenstein
Topics Covered:  Computing treatment effects, Computing overall effects, Constructing Forest Plots, Heterogeneity among effect sizes, Fixed-effect versus Random-Effects models, Funnel Plots, Common criticisms of meta-analysis.