Using data from a 1987 telephone survey of over 1,000 Illinois adult residents, this research looked into factors causing male under-representation in surveys using the last-birthday method of respondent selection. Differences by sex in rates of refusal and non-contact (after respondent selection) were evaluated for relative contribution to gender bias. Accuracy of informant's response to the birthday question was gauged by checking against birth dates of adults rostered during the interview. Other factors, such as the proportion of informants who do not know the last birthday, or who misreport from a desire to be the one interviewed (or not interviewed), were evaluated for relative effect on sample representation of males and females.