Recent research indicates that survey data collected using standard random-digit dialing (RDD) or list-assisted telephone interview methods may encounter problems that bring into question the generalizability of results. One of the major problems facing survey methodologists and social scientists employing this data is the penetration and wide use of cellular opposed to landline telephones as a main source of communication. Initial research shows that there are important differences in a number of respondent characteristics between those answering surveys over cellular and landline telephones. It is therefore likely that these respondents also differ on a number of behavioral and social aspects as well. To dismiss these factors would undermine an array of research involving a number of social issues. This paper examines the comparative mode effect and ownership differences of these modes on self-reported voting behavior. There are theoretical reasons to believe cellular telephone respondents and those owning will differ in vote report. Understanding the potential impact of these populations impact survey methodologists and political scientists alike when developing theory on such behaviors. A dual frame listassisted survey collected data from respondents on either landline or cellular telephones. Two aspects are examined, whether mode of response itself is predictive of response, and whether different telephone ownership patterns are related to voting behavior. Question order - a well-known correlate to response distribution - of retrospective and prospective voting behavior questions was alternated to determine differential sensitivity between modes. Logistic regression shows that cellular telephone response is not a significant predictor to self-report of voting, although question order does have some impact as expected. More importantly, however, ownership of only a cellular telephone is a significant predictor of reported voting behavior. The model incorporates significant correlates from the political science literature, which are significant indicators. Implications and further research needed are examined.