Question

5. Explain why, in a difference-in-differences framework, it is not sufficient to label the treatment effect as the change in outcomes (post-minus-pre) among treated units.

Answer 1

A difference in difference framework works across two different classes of treated and non treated units, so the the simple calculation of treatment effect (post minus pre) is not sufficient here. For example, there are two states : State A and State B, and both introduce a policy to boost farmers' income. In each state the policy is applied on a group of a farmers (Group 1 in state A and Group 2 in state B). The control groups in each state are group 3 and group 4). A simple treatment effect in each state A would have been change in farmers income of group 1 minus change in farmers income group 3. The treatment effect in state B would have been change in farmers income of group 2 minus change in farmers income group 4. But in difference-in-differences framework we also account for the changes arising due to hetrogeneity of states. Accounting for this hetrogenenity, the difference-in-differences framework treatment effect equals: Treatment effect in state A minus treatment effect in state B.

Thus while treatment effect normally involves single difference, difference-in-differences framework involves double difference to account for additional hetrogeneity.