10.9 Looking ahead

The relationships in this unit were allowed to differ across groups, but every observation was still a separate individual, observed once.

A great deal of economic data is not like that. The same states are observed year after year; the same households are revisited; the same firms report annually. Such data carry information that a single cross-section cannot, because a unit observed twice can be compared with itself.

That comparison is the closest this book will come to the thought experiment of Section 6.1 — the same student, observed twice, differing only in attendance. It cannot be arranged in a cross-section at all. It can sometimes be arranged when the same unit is observed more than once, and doing so removes every difference between units that stays fixed over time, whether or not it was ever measured.

That is the subject of the next unit, and it is the first genuinely new answer to the causal problem since Unit 8.