> No, they're not experiments. In an experiment, you manipulate the
> causal factor (indepedent variable) and assess the change in another
> factor. They are interesting econ data, but they're not
> experimental data.
Really? So when you apply a varied set of pricing changes over time across a chain of few hundred stores on an item or a set of items throughout your chain and see how the sales of these and substitute commodities respond, you aren't experimenting (among other things)?
All the best,
John A