At Statistics Sweden, a need for tools enabling the carrying out of large-scale field experiments on a routine basis has been identified. Rather few experiments are carried out annually, and there are no internal guidance documents available. This situation is most likely familiar to other NSIs--as well as there is no major experimentation tradition within this area.
Therefore, a project has started, which will develop an experimentation manual addressing (in contrast to the wealth of experimentation manuals available within e.g. biostatistics) issues pertinent to experiments embedded within ongoing sample surveys.
The project will aim to assist with the planning, execution, and analysis in as many as possible of the embedded experiments at Statistics Sweden. Ideally, this will bring continuous improvement already within the experiments carried out during the course of the project, and allow for much of the advice provided in the final experimentation manual to be based on first-hand experience.
Here, the current state of the project will be presented, with examples from an ongoing experiment embedded within the mail survey "Study Preferences of Secondary School Graduates."
|