ABSTRACT
The purpose of a sample is to generalize about a defined population by inferring from a much smaller group of elements under consideration of the research question. The sample should be a scaled down model of the statistical population that reflects all with regard to the research question relevant characteristics 'true to scale'. It looks into sampling procedures and their impact on the sampling error and the cost of research. Provided a sample can be considered representative, sample parameters allow projections on respective parameters of the statistical population with certain likelihood and without systematic error. So research that is based on today's sample from today's supposed total population tries to gain knowledge about future behaviour and developments. Instead of working with a simple random sample, the sampling error can be reduced using a stratified sample. The financial controller will come and highlight that the smaller the sample size the lower the project costs.
