Legal Requirements for statistical modelling

Funding: Land Steiermark


Led by Iris Eisenberger, REASON traces to which extent political decision makers used statistical models to support measures in connection with the COVID-19 crisis in Austria. Based on these insights, REASON examines which legal requirements for statistical models can be derived from the constitutional principles of democracy, the rule of law, and the liberal state. The results will be used to formulate legal policy proposals that successfully combine the innovative potential of statistical modelling with the requirements of the Austrian constitution.

Statistical models represent aspects of reality in a simplified mathematical equation. This enables observing complex relationships are making forecasts. Among others, pandemic control is an application field for statistical models. However, the quality and validity of these models always depends on the data they are built on.

If these data, assumptions and methods as well as possible downfalls of the model in use are not transparent, democratically legitimized bodies may not be able to make informed decisions: the line between undeniable facts and politically relevant judgement is already blurred before the model’s documentation is presented to the political decisionmakers.

We therefore ask which criteria statistical models must fulfil to ensure that they comply with constitutional requirements. To answer this, REASON first focuses on the analysis of press conferences given by the Austrian federal government in the beginning of the Coronavirus Crisis. The analysis aims to show, if and how political decisionmakers emphasized the information deficit that was present at the beginning of the crisis. In another step, experts that have been directly involved in the decision-making process, e.g. as members of the Coronavirus-Taskforce, are interviewed. By this, we aim to analyze which role statistical models really played in the decision-making process.

Parallel to this we will conduct a thorough analysis of relevant legislation. By this we will identify already existing regulations regarding statistical models and eventual gaps within the legislation.


Download Interim Report (PDF)


Involved Institutions

University of Graz

Technical University of Graz