Survey Data Analysis

Professor Cautrès is an extremely good and very passionate teacher. I never learned statistical methods this well before. — participant from Japan

The range of survey resources is constantly expanding and facilitating the systematic analysis of public opinion and socio-political attitudes around the globe. This course surveys multivariate methods for the study of public opinion data. It combines the study of sophisticated statistical techniques with the discussion of practical problems you will encounter when conducting your own survey analyses, and it provides you with the statistical skills required to tackle these problems.

This course complements Interviews and Survey Design, which focuses on the steps that precede the analysis of survey data, i.e., the design of surveys and the collection of survey data. As these two courses run consecutively, participants who are interested in taking both courses can combine them.


Dates

This one-week course was offered in 2013.


Instructor

Bruno Cautrès (picture), Sciences Po Paris


Detailed Description

One of the most important recent transformations in empirical social science research has been the expanding range of survey resources. These resources dramatically facilitate the systematic analysis of public opinion and socio-political attitudes around the globe. The globalization of the study of public opinion over successive decades as well as recent developments in statistical methods have stimulated the social sciences in such areas as the quantitative and comparative analysis of political values, social cleavages, and attitudes to democracy and democratization.

This course addresses two main and complementary methodological issues. First, we ask how we can use survey data sets? What are the designs of these studies and how can we use quantitative techniques to analyze the quality of surveys data. How can we analyze their reliability and how do we deal with missing data and errors? The second issue addresses the question of how the major multivariate statistical techniques (i.e., uni- and bivariate analysis, multiple linear regression, logistic regression, factor analysis, attitudes scales construction) can be used to analyze survey data and to test the relationships between variables? The course teaches these multivariate methods in simple and accessible terms.

The course consists of lectures, exercises, and assignments. For hands-on training, we analyze various data sets, e.g., the World Values Studies, European Social Survey, and the Asian Barometer (a study of public opinion in different Asian countries, including Singapore), and test for the relations between people's values, attitudes, and behavior that can be found in these data.

Complementing Interviews and Survey Design, this course combines the learning of fundamental statistical techniques used to analyze surveys and the learning of the practical problems encountered when using them.


Prerequisites

There are no formal prerequisites. However, we strongly encourage participants to combine this course with Interviews and Survey Design. In addition, a basic background in quantitative methods guarantees that participants get the most out of this course.


Requirements

Participants are expected to bring a WiFi-enabled laptop computer. Access to data, temporary licenses for the course software, and installation support will be provided by the Methods School.


Core Readings

Harkness, Janet A., Fons J.R. Van De Vijver, Peter P. Mohler, eds. 2003. Cross-cultural Survey Methods. Hoboken, NJ: John Wiley and Sons.

Gujarati, Damodar N., and Dawn C. Porter. 2009. Basic Econometrics. 5th edition. New York, NY: McGraw-Hill.