The relationship between income and education, with occupation held constant, can be ascertained indirectly from incompletely cross-tabulated data, making use of formulas involved in the analysis of covariance. The between-occupation component of the difference in income level between categories of educational attainment is substantial. Occupation, therefore, is an important intervening variable in the translation of educational advantage into income advantage. The method may be useful in a variety of problems, familiar to demographers and social statisticians, whose analysis must be approached in a roundabout manner owing to the limited detail of published tabulations.