We show why considering a number of education-dependent covariates in the wage equation decreases coefficient of education in the wage equation. We use a meta-analysis of results for Portugal to show, empirically, that this is the case. The coefficient decreases when we use covariates that can be considered post education decisions; it is independent of the sample size, tenure and the fact of using hourly or monthly wages. At this stage the use of the simple specification of the Mincer equation for the study of total returns to education continues to hold our support.