Two basic methods exist for drawing probability samples to be used in telephone surveys: directory sampling (from alphabetical or street directories) and random digit dialing (RDD). RDD includes unlisted numbers, whereas directory sampling includes only listed numbers. The goal of this paper is to estimate the effect of failure to include unlisteds in the sample on survey results. Two community-wide telephone surveys were conducted for the Sacramento City Zoo (California) (N=100) and for Sacramento Department of Parks and Community Services (N=429), using the pure RDD method. Data analyses were conducted using the chi-square statistic as run on SPSS/PC+. Results showed responses to four questions exhibited significant listed-unlisted differences. There was evidence that researchers using listed RDD samples might come to different conclusions, but the evidence was not strong. Although the research presented here is clearly limited (reports on a single and fairly narrow geographic area, concerns only two surveys addressing related topics, and contains one survey with a small sample size), it suggests that survey results may depend on sampling technique. Additional inquiry is warrented for a more definitive assessment. (BAE)