Parametric and semiparametric estimated wage equations, which correct for sample selection bias, are used to assess the returns to eduction and extent of ethnic discrimination in (Peninsular) Malaysia. In particular, this paper focuses on the level of discrimination between Malay and Chinese men and women. The Andrews-Schafgans (1998) estimator is used to estimate consistently the wage equation intercept in the semiparametric case. The results show that the Chinese-Malay offered wage difference typically is larger in absolute value among women than men. The strong discrimination favouring Chinese over Malays observed using the parametric results is negated by the semiparametric results.