Then they estimated the relationship between people's
neighborhood scores and their performance on cognitive tests
over two years, factoring in issues like age, gender, education and wealth, that might
influence people's cognitive scores independently of
neighborhood characteristics.