Presents the methodology adopted and the data sources employed, an extension of the Blinder-Oaxaca decomposition, which is the prevailing approach in the empirical work on earnings gaps. Three limitations are noted: (1) the relationship between characteristics and earnings is not necessarily linear; (2) Blinder-Oaxaca is informative only about the average earnings gap decomposition, providing no clues about the distribution of the differences in pay; and (3) Blinder-Oaxaca fails to restrict its comparison to comparable individuals, which is likely to substantially upwardly bias the estimators for unexplained differences in pay. The extension of the Blinder-Oaxaca decomposition approach uses a nonparametric matching approach, in an attempt to explore the extent to which gender and ethnic earnings gaps can be attributed to differences in observable characteristics. This alternativeapproach addresses the traditional Blinder-Oaxaca question notonly for averages but alsofor the distributionof earnings, emphasizing the role of gender and ethnic differences inthe common supportof the distribution of observable human capitalcharacteristics.