Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Although the Heckman approach has often been used in empirical analysis, the marginal effects, necessary to interpret the effect of the regressors on the dependent variable, appeared to be overlooked. Using the Heckman approach, general expressions are derived for calculating the conditional and unconditional marginal effects. Based on a sample of Brazilian women, the conditional and unconditional return to education are calculated for the logarithm of earnings equation estimated by Heckman's procedure, comparing them to the marginal effect of education obtained without correcting for selectivity bias. The same analysis is carried out for a discrete variable 'black'.