Theory and Application of Longitudinal Surveys (Course code 0411)

Purpose

Allow participants to learn the various types of surveys repeated over time, while underlying the particularities of each. This is accomplished by visiting all survey steps.

Using that approach, methodological issues raised from the early steps such as sampling frame and selection, to later steps such as estimation and analysis, are examined.

Benefits to the participants

Participants will be able to apply the concepts and techniques in their work.

Target population

Employees with basic knowledge of mathematical statistics and of sampling, and who wish to increase their knowledge of the various methodological issues relative to surveys repeated over time.

Course outline

Introduction

  • Motivations
  • Basic concepts and definitions
  • Advantages
  • Disadvantages

Longitudinality and sample designs

  • Typical sample designs
  • Design parameters
  • Longitudinal units
    • Social surveys
    • Business surveys
  • Use of registers and other frames
    • Business surveys
    • Social surveys
    • Frame updating process
  • Sample selection
    • Social panels
    • Business panels
  • Conversion to panels of existing surveys
  • Questionnaire design
  • Data collection
    • Collection modes
    • Dependent interview
    • Rules concerning respondents
    • Tracing
    • Incentives
    • Electronic data reporting
  • Types and causes of nonresponse
  • Quality assurance
  • Confidentiality

Weighting and estimation

  • Basic weighting
    • Longitudinal weighting
    • Cross-sectional weighting
  • Treatment of nonresponse
    • Types of nonresponse
    • Response rate in general
    • Longitudinal response rate
    • Cross-sectional response rate
    • Creation of response homogeneity groups
  • Imputation
  • Calibration (poststratification)
    • Basic poststratification
    • Longitudinal poststratification
    • Cross-sectional poststratification
    • Integrated weight (in social surveys)
  • Descriptive longitudinal statistics
  • Variance calculation
    • Jackknife
    • Bootstrap

Longitudinal data analysis

  • Multiple regression
  • Survival analysis

Applications

  • National Population Health Survey (NPHS)
  • Survey of Labour and Income Dynamics (SLID)

Longitudinal surveys of the world

Conclusion

  • Summary
  • New needs from users

References

Notation

Prerequisite

Basic knowledge of mathematical statistics and sampling. Since some hands-on work will require handling datasets, knowledge of the SAS software program is desirable but not a necessity.

Duration: 4 days