Multivariate Statistical Analysis and Comparative Crossnational Surveys Data - Bruno Cautrès

This course aims to give the participant an overview and practical experience of the major theoretical and practical issues that are faced by the statistical analysis of cross-national comparative survey analysis. This course combines the learning of fundamental points about the statistical techniques used to analyse such surveys (without going into too complex mathematical things) and the learning of the practical problems encountered when using them. The course is an introductory level one even if methods taught are multivariate ones. It is designed to help social sciences students that have difficulties into entering in the multivariate techniques world and to make them crossing over these first difficulties during the summer school. Because the course is introductory, it does not require any other previous skills than basic descriptive statistics and hypotheses testing. The originality of the course is to present in most simple formats the multivariate methods and to concentrate on their use for cross-national comparative analysis. The course proposes a kind of a panorama of the most used techniques such as linear regression analysis (and ANCOVA), logistic regression, latent class and loglinear models; factor analysis and correspondence analysis. Typically, students attending the course may wish to use cross-national data sets (ESS, EVS, Eurobarometers, ISSP, macro-comparative data bases) for their research and may have as a key problem the question of the homogeneity/heterogeneity of the statistical relationship across their countries cases. The “black box” explanation of country effect can be investigated through multivariate technique that test for this homogeneity/heterogeneity hypotheses. More generally, the statistical methods that will be used in this course are relevant for “multiple groups” analysis, when groups are countries, but could be any groups such as gender, ethnic, regional groups. The course will start from the questions raised by the fact that in a cross-national survey framework there is rarely complete equivalence of data. We then will meet different families of methods: from the simple correlational and regression perspective (linear regression, ANCOVA, logistic regression) to the more advanced techniques such as methods for data reduction (factor analysis, correspondence analysis) and through methods of comparative analysis of cross-classified data (loglinear models). The students will have the opportunity to use the European Social Survey as the basis for conducting their own analysis and will also be asked to complete some readings during their own time. They will be asked to organize in small groups (3 or 4 students), to work together and to deliver a final group work. This final group work is an original work, done during the summer school. Short Bio Bruno Cautrès is attached to the CEVIPOF- Centre de recherches politiques de Sciences Po (Paris) at the Fondation Nationale des Sciences Politiques in Paris. He is a senior CNRS research fellow with interests in voting behaviour, comparative survey research and quantitative techniques. He is involved in European Social Survey, European Values Studies, International Social Survey Programme and European elections studies. His main research interest is on elections studies, survey research and on the changing political values and orientations due to EU integration and globalization.

---