Recently, small-area methods (SAM) have been used to establish a new estimation method in empirical social research. With SAM, reliable estimates can also be determined for small subareas or smaller groups or groups with low prevalence in the population.
This allows more accurate (with a smaller sampling error), small-scale population estimates, which can replace previous extensive empirical studies.
Among other things, SAM replaces or supplements the classical statistical estimation methods with Bayesian statistical methods. Thus, the advantage of SAM over traditional estimation models is that the use of external, non-survey auxiliary variables, with information about a small-scale unit, does not require the amount of empirical data that classical estimation requires. The only requirement is a clear, hierarchical structure of the data that can be geographical (but not necessarily).
The core competency of SAM is the modeling of the data, ie the application of the various auxiliary variables for the respective estimates. Depending on the object of investigation, similarities and differences of different areas as well as meaningful correlations are mapped and then used for the prediction of the characteristic values. In doing so, we use different external variables and form correlations to the applied questions. Here we are supported by the company sister infas 360, which specializes in the collection, processing and analysis of georeferenced external stock data.