Course 10: Sampling, Weighting, and, Estimation

Lecturers: Dr. Tarek Al Baghal, Dr. Alexandru Cernat

Date: 20-24 August 2018
Time: 11:00-13:00, 14:00-18:00

Short Course Description:

The course will provide an overview of those aspects of survey design, implementation, and analysis that are unique to longitudinal surveys or that have distinct features in the longitudinal context. The course will specifically cover:

  • A review of the advantages and limitations of longitudinal surveys and an outline of some of the uses to which longitudinal surveys are put.
  • Key aspects of longitudinal survey design such as the sampling design, interval between waves, and data collection modes.
  • Important aspects of designing a questionnaire and measurement for longitudinal studies, particularly for capturing micro-level change.
  • The impact of non-response and attrition in a panel, and adjustments such as weighting and imputation given that missing data patterns differ between waves.
  • An introduction to important analysis considerations and methods when using a longitudinal survey design.
  • In addition, each morning time will be available in a computer lab for exercises for practical application of these methods and an opportunity for more interaction with the instructors.

Course Prerequisites:

  • participants should have basic knowledge of survey methodology from a cross-sectional perspective, in particular with respect to survey design, instrument development, and survey implementation
  • basic knowledge of statistics and statistical modelling (i.e., regression) and of a statistical software (e.g. Stata, SPSS, SAS, R).

Target Group:

Participants will find the course useful if:

  • they are interested in the uses and importance of longitudinal surveys
  • they are involved in planning or conducting a longitudinal survey
  • they are interested in aspects of longitudinal data collection
  • they are in the processes involved in measurement of phenomena over time
  • they want to use appropriate methods to analyze longitudinal data.

Course and Learning Objectives:

By the end of the course participants will:

  • be familiar with the central design issues of longitudinal surveys;
  • understand different strategies on how to collect longitudinal data;
  • be able to design questions that meet research objectives for longitudinal surveys and minimize error
  • be able to identify and implement features that should help to prevent panel attrition;
  • be able to study the nature of non-response and attrition in a longitudinal survey;
  • understand generally the methods used in weighting and imputation in a panel survey.

Organizational Structure of the Course:

The summer school courses are structured around four hours of classroom instruction and 2 hours of practical sessions per day. Course instructors shall be available for group and/or individual support for these two hours during the practical sessions. In these meetings, practical applications will be given for the students to work through, with the opportunity to ask instructors questions one-on-one about the course or their own projects. The practical application will relate to topics in longitudinal survey design covered in the lectures, including the impact of question wording on measurement, the identification of attrition, the development of weights, and how these can impact analytic findings.

Software and Hardware Requirements:

None. This course takes place in a computer lab. GESIS will provide participants with access to the statistical software packages used in the course (i.e., Stata, R) and various other statistical software packages (e.g., SPSS, Mplus).