An **explanandum** is a set of propositions that refer to the thing being explained. An **explanans** is the set of proposition that does the explaining. For example, if sales of a product have declined, an explanandum is that sales have declined. An explanans might be that a decline in marketing spend caused the decline in sales. Explanatory modeling is an explanans that makes use of statistical models for testing and providing causal explanations. Explanatory models often - have parameters with a physical meaning - arise in the context of an **experimental design** - require additional assumptions if the data come from an **observational study**. See also [[prediction versus explanation]].