Designing weights

Introduction

Topic: in practice, nonresponse weights should be developed keeping in mind what the dependent variable is. In situations where many variables asked among respondents are also available for the population, there are many possible nonresponse weighting models. In this week, we will focus on the question of how to develop nonresponse weights. We discuss the concept of paradata.

Literature:

• Brick, J. M. (2013). Unit nonresponse and weighting adjustments: A critical review. Journal of Official Statistics, 29(3), 329-353.

Optional

• Kreuter, F. (2013). Improving surveys with paradata: Introduction. Improving surveys with paradata: Analytic uses of process information, 1-9.

Lecture:

Introduction to paradata. Lecture about designing and constructing weights Slides

CLass Exercises

Discussion about possible weighting variables given different dependent variables that are often of interest to scientists (e.g. voting behavior, health, finances) + an exercise where you develop weights for a survey.

Class discussion weighting variables
R exercise weighting
R exercise weighting solutions sample data for R exercise
respondent data for R exercise

Take home Exercise:

Finish the weighting exercise

Before week 10

For the final assignment you will work in a group of about 4 (+- 1) on a real-life dataset. The goal is to apply all the materials covered in the course, answer a substantive research question and solve a real-life challenge covers an advanced problem in survey research. You will receive an e-mail on your @students.uu.nl account with more information about group membership, the assignment itself and your group members this week.