Statistics for social sciences II (E207)

Course Organization

Eugene Horber: Eugen.Horber@politic.unige.ch
Class Meeting Times Thursdays, 16.15-18 (Rigot)
Course Website: http://www.unige.ch/ses/sococ/heistat
Office Hours: By appointment
Prerequisites: An introductory statistics course.

Course Description

Statistical tools and statistical thinking are an essential part of the methodology of many scientific fields, including the empirical social sciences.The focus of this course will be on the application of more advanced statistical tools to typical research problems in your field(s). At the end of the course active participants should have enough knowledge to understand and criticize most publications using standard statistical tools and sufficient practical skills with statistical tools, software and data to start using them for their own research. The course is based on the following methodological and didactical principles:

• Statistics should not be taught on its own, but connected with practical methodological issues in a specific substantive field.
• Statistical skills can only be acquired through hands-on learning in a realistic setting (real data, typical research problems) using standard analysis software.
• Understanding concepts and approaches is essential, as are their underlying assumptions and pitfalls; technical/mathematical details however can often be left to the specialist.
• Understanding when, why and how to use a statistical tool in a real research situation is the key to good practice.

The course will have three main components:

• Presentation of statistical tools applied to a number of typical research problems in political
• science and related fields.
• Hands-on experience with software (namely Excel and SPSS) and real data.
• Examining methodological aspects of typical publications (books, articles) from the literature
• and whenever possible participant's research projects.

You are expected to:

• Learn to use the tools presented in each class on the course learning site
• Work the guided exercises for each course (found on the website).
• Hand in the assignments (not marked, but required to earn the credits)
• Write a research paper, based on a critical reanalysis of a published paper.

Tentative course outline

As this course is conceived around typical research problems and participant's projects, a general outline will be established after the first courses, based on both participant's research interests and methodological background. Detailed outlines for each session will be available from the course website.

Part 1 Review: Background and basics

A quick review of knowledge and skills you should already possess (…just to make sure; with some additional practical details).

- Foundations

• Statistics and empirical research in the social sciences, especially international relations
• Data: Measurement theory; data collection and data quality; data types; data structures
• Data presentation: communicating quantitative information

- Basic tools

• Descriptive statistics, data exploration
• Regression, crosstabulation
• Analysis of residuals

Part 2. Advanced statistical tools

Below you will find a provisional list of topics, subject to revision, depending upon the progress and methodological requirements of the research projects.

- Regression

• Beyond simple regression; regression assumptions and diagnosis
• Regression with categorical variables
• Logistic regression and variations
• Issues in causal modelling

- Analysis of time based data

• Time series analysis
• Longitudinal analysis; event history analysis

- Analysis of survey data

• Sampling theory
• Survey design and analysis
• Modelling categorical variables; log-linear models

- Dimensional Analysis

• Principal components, factor analysis and variations

- Classification and discrimination

• Cluster analysis
• Tree models

- Further possible themes

• Analysis of variance
• Advanced graphical methods
• Robustness: theory; bootstrap and jacknife
• Network analysis
• …