Cologne, Germany

Sampling and Weighting in Survey Statistics

online course
when 22 August 2022 - 23 August 2022
language English
duration 1 week
fee EUR 200

Methods to analyze data taught in introductory statistics and econometric courses often rely on the assumption that the data are collected in a simple random sampling process. However, in practice rather complex sampling techniques are often used, such as stratification or clustering. In these cases, a weight has to be assigned to the sampled units to account for the sampling design. The course will cover methods to select random samples and weight the sampled units to infer from the single sample to the population. We will discuss the several steps in a weighting process including 1) obtaining the design weights to account for the random sample selection, 2) adjusting the design weights to compensate for nonresponse, and 3) adjusting the weights such that the sample estimates agree with known population totals. The focus of the course is on the intuition of the underlying theory and the statistical methods so that the participants can assess the appropriateness of the methods in their application by themselves. For the course, no prior knowledge on survey statistics is required, but the participants are expected to be comfortable with statistics and to have some experience with data analysis. We will do exercises in R that apply the techniques learned in the lecture. Participants will get the most out of the class if they have prior experience with R.

Course leader

Anne Konrad is a postdoc at Trier University, Germany. She is involved in a research project with the Federal Statistical Office of Germany.

Target group

Participants will find the course useful if:
- you have experience conducting surveys and/or analyzing survey data but have no experience with survey sampling and weighting.
- you plan your own survey and have to weight the collected data, or you are analyzing survey data.

Prerequisites
- Introductory course in statistics. No prior knowledge of sampling theory is assumed.
- Prior knowledge of R is required.
- Basic understanding of survey methodology and how to handle survey data is helpful but not necessary.

Course aim

By the end of the course participants will:
- know the most commonly used sampling designs including element sampling and multistage sampling (simple random sampling, stratified sampling, sampling proportional-to-size, cluster sampling, and related designs).
- know to compute design weights, how to compensate for nonresponse, and utilize external auxiliary information into the weighting process.
- understand how the sampling design affects the analysis of survey data.
- be able to assess the advantages and disadvantages of the different sampling designs and weighting methods.
know how to apply the discussed methods in R.

Prerequisites:
- Introductory course in statistics. No prior knowledge of sampling theory is assumed.
- Prior knowledge of R is required.
- Basic understanding of survey methodology and how to handle survey data is helpful but not necessary.

Credits info

Certificate of attendance issued upon completion.

Fee info

EUR 200: Student/PhD student rate.
The rates include the tuition fee and the course materials.
EUR 300: Academic/non-profit rate.
The rates include the tuition fee and the course materials.

Scholarships

Scholarships are available from the German Academic Exchange Service (DAAD) and the European Survey Research Association (ESRA), see https://www.gesis.org/en/gesis-training/what-we-offer/summer-school-in-survey-methodology/scholarships.