Household surveys are increasingly moving toward self-administered modes of data collection. To maintain a probability sample of the population, researchers must use probability methods to select adults within households. However, very little experimental methodological work has been conducted on within-household selection in mail surveys. In this study, we experimentally examine four methods—the next-birthday method, the last-birthday method, selection of the youngest adult in the household, and selection of the oldest adult in the household—in two mail surveys of Nebraska residents (n = 2,498, AAPOR RR1 36.3 percent, and n = 947, AAPOR RR1 31.6 percent). To evaluate how accurately respondents were selected from among all adults in the household, we also included a household roster in the questionnaire for one of the surveys. We evaluated response rates, the completed sample composition resulting from the different within-household selection methods, and the accuracy of within-household selection. The analyses indicate that key demographics differed little across the selection methods, and that all of the within-household selection methods tend to underrepresent key demographic groups such as Hispanics and persons with lower levels of education. Rates of selection accuracy were low among the four selection methods analyzed, and the rates were similar across all four methods.