Large-scale sample surveys are not designed to produce reliable estimates for small population domains, such as specific geographical areas or population groups. This is because area sample sizes may be “too small” to compute reliable direct estimates (obtained via the use of the sample data only). Therefore, small area estimation methods that borrow strength information from auxiliary data, such as the Census or administrative data, can be used to produce reliable estimates. This course will cover widely adopted small area estimation methods based on the direct and model-based estimation approaches. It will be structured in three parts. The first part will be about the introduction to the small area estimation problem and the use of direct estimators to produce small area estimates. In the second part, we will introduce the unit-level approach based on the Battese, Harter, and Fuller model, assuming that auxiliary information is available at the unit level. The third part will focus on the area-level approach based on the Fay-Herriot model, which is useful when the auxiliary information is available at the area level only. The course will also focus on the evaluation of the small area estimates via diagnostics tools and simulation studies. Practical applications and examples will be provided in each part of the course.

 

Practical applications and examples will be carried out in R. We will show some common R packages available to users. Applications will be based on both real data and simulation studies.

 

Organizational structure of the course

Lectures, exercises, and tutoring sessions will be held daily throughout the week. In the morning, we will focus on lectures and practical examples, whereas the afternoons will be dedicated to labs and workshops.

By the end of the course, you will:

Software and hardware requirements

You need to bring a laptop to the course with R (https://cran.r-project.org/) and R Studio (https://posit.co/download/rstudio-desktop/) installed to successfully participate in this course. Both programs are free and open source.

 

You should have the ability to download packages. Internet access is provided by GESIS.

 

Please install R, R Studio, and the following R packages and their dependencies prior to the start of the course:

 

packages <- c("car", "dplyr", "emdi", "ggplot2", "lme4", "influence.ME",

             "lavaan", "msae", "nlme", "patchwork", "readxl", "rsae",

             "sae", "sampling", "sf", "spdep", "survey", "tidyverse")

 

Also, you will need the package 'saery', which is not available on CRAN anymore, but it is on the archive. You can find it here: https://cran.r-project.org/src/contrib/Archive/saery/.

 

“For an introduction or refresher in R programming, you might consider enrolling in GESIS's two-day onsite course, “ Introduction to R for Data Analysis ” held in the first week of the Summer School in Cologne, or the four-day online workshop, “ Introduction to R ” offered in May.