 | |
STATISTICS
FOR THE SOCIAL SCIENCES
6-18 July, 1998
Course Director: Tamás
Rudas (Central European University, Budapest, Hungary)
Resource Persons:
Vladimir Batagelj (University of Ljubljana, Slovenia)
Anuska Ferligoj (University of Ljubljana, Slovenia)
Sándor Kabos (ELTE University, Hungary)
Nicholas Longford (DeMontfort University, Great Britain)
Course Description
Valid and reliable information regarding
the status of the society, and the changes taking place in it, is vital
for any social scientist. Statistics is the science of collecting and analyzing
data, and the application of statistical methods has proved to be the most
important tool of obtaining such information. The goal of the proposed
course is to give an overview of some of the most important statistical
methods applicable in the social sciences.
The selection of topics and the level
of presentation make this course accessible to both participants with a
social science background but with no mathematics or statistics education,
and to participants who have had statistics training concentrating on classical
chapters of statistics. Four topics will be covered and each will give
directly applicable
knowledge in the collection, analysis
and interpretation of data. Furthermore, each of the four topics will be
presented in a way that can serve as model for classes or seminars that
participants may give at their respective home universitites.
Topic 1: INTRODUCTORY STATISTICS FOR
THE SOCIAL SCIENCES
Background, Design and Planning of Studies.
Examples.
1.1. Elements of experimental
design. Observational studies.
1.2. Informativeness and selection
bias. Differences and effects.Causation and association.
1.3. Overview of survey methods;
probabilistic control.
1.4. Research design. Formulation
of substantive problems (hypotheses). Research strategy.
Execution of Studies.
1.5. Scientific (population) quantities.
Estimation. Distributions.
1.6. Random and fixed. Populations
and sets.
1.7. Measurement process.
Validity and reliability. Sources of error.
1.8. Statistical models.
Description of systematic features and of variation.
Data Analysis and Inference.
1. 9. Computing environments.
Data management and quality control. Diverse sources of information.
Meta-analysis.
1.10. The role of a statistical consultant.
Discussion, examples.
Topic 2: CATEGORICAL DATA ANALYSIS
Introduction
2.1. Contingency tables, intutitive
analysis.
2.2. Sampling schemes for categorical
data, the multinomial distribution.
2.3. Parameterizations of multivariate
discrete distributions. Odds ratios and marginal distributions.
Log-Linear Models
2.4. Conditional odds ratios and log-linear
models.
2.5. Estimation and testing of log-linear
models
2.6. Logistic regression
Other Models and Techniques
2.7. Linear models for categorical data,
association models.
2.8. Latent class models.
2.9. Graphical modeling of association.
2.10. Assessing the fit of a model.
Topic 3: THE ANALYSIS OF SOCIAL NETWORKS
Basic Notions
3.1. Historical overview, the notion
of a network, types of network data, representations of networks.
3.2. Measurement and collection of network
data, boundaries.
3.3. Connectedness, centrality and prestige
of social actors, centralization of networks, equivalences of actors.
3.4. Formalization: relations, graphs
and matrices.
Graph Theoretical Methods
3.5. Basic notions and tools from graph
theory: subgraphs, connectedness, components, invariants.
3.6. Centrality and prestige of actors,
centralization of networks.Matrix methods.
3.7. Signed graphs and balance in networks.
Blockmodeling
3.8. Equivalences of actors, blockmodels.
Indirect blockmodeling methods,
clustering approach, examples.
3.9. Direct blockmodeling methods and
local optimization, examples.
3.10. Generalized blockmodeling, pre-specified blockmodels, examples.
Topic 4: SPATIAL STATISTICS AND THE
ANALYSIS OF TIME SERIES
Basic Concepts
4.1. Correlation, autocorrelation, partial
autocorrelation. Correlated
observation units.
4.2. Random fields. Examples from econometrics,
demography, urban sociology.
Stationarity in time
and/or in space.
The Stationary Case
4.3. Computer simulation of random fields.
Numerical techniques in
spectral analysis.
4.4. Sampling in time and/or in space.
The Sampling Theorem.
4. 5.Covariance function, variogram.
Parameter estimation in geostatistics.
4.6.Parameter estimation of Markov fields.
The Non-Stationary Case
4.7. Conditional simulation of Markov
fields. MCMC methods.
4.8. Spatial segmentation.
4.9. Modelling non-stationarity.
Multiresolution analysis. Wavelets.
4.10.Case studies.
|
