This study examines the results of three instruments developed to measure sexual harassment. Two instruments were used in the Dutch national victimization survey: an oral interview and a written questionnaire. Three issues will be discussed: (1) do both instruments produce the same victimization rates? (2) are the type of incidents mentioned similar? and (3) is the relationship between the answers on both instruments linked to socio-demographic characteristics? Results indicate that the written questionnaire produces higher victimization rates but that generally less serious incidents are mentioned. Secondly, information on the relation of victimization with educational level is presented. For this purpose, the results of the victimization survey (written questionnaire) are compared with another Dutch study measuring sexual harassment. As is found in the literature, a relation between victimization and educational level is found in the national victimization survey, but, in contrast to many results reported in the literature, no such relation was found in the Leiden study. The implications of these results are discussed.