A frame is a list of all eligible respondents. For a large frame, i.e., a large population, we can ignore the distinction between sampling with replacement and sampling without replacement. Many surveys use the census as a sampling frame. Any bias in this frame?