|
Instructors: Prof. Dr. Martin Elff
Event type:
Seminar / exercise
Org-unit: Sociology, Politics & Economics
Displayed in timetable as:
Advanced Methods | X
Hours per week:
2
Credits:
3,0
Location:
Campus der Zeppelin Universität
Language of instruction:
German
Min. | Max. participants:
7 | 24
Priority scheme: Standard-Priorisierung
Course content:
Surveys are a central instrument for data collection in sociological and
political science research on attitudes and behaviour, the empirical analysis of
social structure, and in market research. The analysis of survey data poses a
series of challenges that are only incompletely adressed in regular
introductions to the methods of social research and statistics. These challenges
become more and more urgent in an era of declining response rates in personal
interview surveys and telefone surveys and of an increasing relevance of online
surveys:
- How can the representativeness of a survey sample be reconstructed if response
rates are highly uneven?
- How can the precision of estimates improved by taking into account
population-level information?
- How can non-responses be taken into account in such a way that it does not
bias survey sample summaries even if non-responses are unevenly distributed?
- How can uncertainty of survey sample-based estimates adequately assessed?
Beside these fundamental question also more practical problems are adressed in
the course:
- How can I make effective use of information from survey sample documentation
(such as codebooks)?
- How do I prepare data in such a way that quantitative analysis is feasible at
all?
- How do I deal with responses to filter questions and responses that are
preceded by filter questions?
- How do I construct Likert-Scales from Likert-Items and how do I assess the
quality of such scales?
- etc.
- How do I create graphical summaries of survey data that are both accessible
and informative?
Educational objective:
The aim of the course is to enable participants to independently prepare and
analyse different kinds of survey data using statistical software (in particular
using R-packages "memisc" and "survey"). The focus lies on an effective
data preparation and creation of data summaries rather than on the application
of advanced multivariate techniques.
Further information about the exams:
Assessment is either based on completing a worksheet or an analysis of survey data to answer a self-chosen research question.
Mandatory literature:
Weisberg, Herbert F.F. 2005. The Total Survey Error Approach: A Guide to the New Science of Survey Research. Chicago: University of Chicago Press.
|
