As is well-known there is no consensus on how to measure cognitive reserve. The CFAS team [AJE, PLoS Med] used weighted sums of, variously, years of education, occupation rank, and social activity. There is no consensus on the weight of each term, nor on what terms make up the sum. The most common terms are education and occupation. Although cognitive reserve has been called a latent construct, no latent variable has been derived from the terms or the indicators, preferring instead to assign weight to each term before summing all weighted terms.
We proposed here a latent variable of cognitive reserve measured by manifest indicators of years of education and occupation rank, and letting the data decide on the weights or loadings of the indicators. To ease interpretation we construct a latent class of cognitive reserve with two levels of low and high cognitive reserve, with the data assigning each person onto one of the two levels.
