This project contains code for chosen parts of the doctoral thesis titled Model-Based Prediction and Estimation Using Incomplete Survey Data, submitted in partial fulfillment of the requirements for the degree Doctor rerum politicarum (Dr. rer. pol.) to the Faculty IV at Trier University by Anna-Lena Wölwer. The thesis is freely available for download under the following DOI: 10.25353/ubtr-xxxx-25a6-5f2c.
The doctoral thesis was submitted in June 2022, defended in December 2022, and published in January 2023.
- MMFH_example: Shows a commented example of generating data according to a MMFH model (using MMFH_gen_dat_m3.R) and estimating its parameters using algorithm MMFH_fitting.R. The presented example can be used as the basis for a model-based Monte Carlo simulation study similar to the ones shown in Chapter 6. It is available as qmd, html, pdf, and docx.
- MMFH_gen_dat_m3.R: Code for generating the different data parts used in model-based Monte Carlo simulation studies of multivariate Fay-Herriot Models without and with randomly missing direct estimates. The code is written for 3 dependent variables.
- MMFH_fitting.R: Code for fitting a multivariate Fay-Herriot model under missing direct estimates (MMFH). For fitting, we use a Fisher-Scoring algorithm, either with ML or REML.
- XXXX: Under construction