Surveys like these do often have implicit bias, because only a small portion of people would be willing to take them when offered. For instance, phone surveys dramatically overrepresent Republicans because almost all the people who choose to complete them are quite old. Additionally, people with more extreme views are more likely to complete political surveys than people with moderate views, especially those who aren't very interested in politics. These are all biases that are difficult to account for.