Given more and more surveys are being done on-line, we need rigorous analyses to determine the nature of the digital divide and its affect on sample representativeness and sample altitudes and opinions. We used a mixed mode survey, an Internet survey with a poll survey request option, to interview registered voters in the 2006 midterm elections in New Mexico's Congressional District 1 and Colorado's Congressional District 7. We contacted over 7000 registered voters immediately following the election with o contact letter and then followed that up with 3 reminder postcards of intervals of 7 to 10 days. Because of the nature of our sample, general election voters, we know demographic and election characteristics of our sample population including their gender. age, location, and voting history. This allows us a careful comparison of our sample respondents to our sample population to determine its degree of representativeness. In addition, we also examine differences between survey mode on attitudes and behaviors as well as the overall effectiveness of our research design, which included reminder post cards and a toll free number to request a mail survey. We also compare our experience with political activists, Democratic caucus goers, in New Mexico in 2004 to determine how different populations respond to requests for Internet surveys.