Exercise 2 solution

The claim has the form of the ecological fallacy. The Ecological fallacy page, as Chapter 10 reports it, describes it as deducing inferences about individuals from inferences about the group they belong to. The data are about regions (an average income per region); the conclusion is about each person ("a degree makes each person richer"). Robinson's 1950 case is the warning: the page reports a state-level correlation of -0.46 (after a 2011 correction) while the individual-level correlation was +0.12, so the group figure had the wrong sign for individuals.

The claim also falls under correlation taken as cause, from the Correlation does not imply causation page. The three standard alternatives from that page:
- A hidden third variable: rich regions may have more graduates and higher incomes because of the kinds of employers located there. The graduates and the incomes are both products of that.
- Reverse causation: where incomes are high, more families can afford university, so incomes would be raising graduate numbers.
- Coincidence: unlikely with many regions, but the page lists it.

A good answer picks one and explains it in a sentence or two. For example: a region with a large financial sector has many high-paying jobs and many graduates drawn to them, so the region average rises without the degree itself adding anything to each worker.

What would be needed: data on individuals, comparing people with and without degrees while taking account of background, region and occupation (my own suggestion of method, from general knowledge) (check). The claim should also say what its sample was, which is the selection bias question. A good answer stays modest: it does not claim a degree has no effect, only that this evidence does not show it.