ADJUST FOR WINDOWS: Extrapolation of micro data for representative samples

Project: Research

Project participants


A central problem when working with surveys and samples in general is the representativeness for the underlying population. Due to many errors, 'distortions' can occur even in random samples, so that new weights can improve representativeness. With the optimization approach of minimal loss of information under the restrictions of the aggregate data to be achieved, the program package Adjust developed by Prof. Dr. Joachim Merz was improved and extended in the reporting period. This efficient package calculates new weighting or extrapolation factors with which the individual data of the micro units can be reweighted. Adjust is already used internationally in numerous statistical offices and institutions. In order to implement the previous stand-alone program flexibly as a module in other applications, the C++ version was ported to Java in the reporting period. Thus, ADJUST will be available to all interested students as one of many modules for microsimulation within MICSIM.