During the last 30 years, a large literature has accumulated showing how survey results differ depending upon whether the data are collected in face-to-face interviews, via telephone interviews, or via computer-assisted self-administration. Professor Krosnick will review this literature and offer some general conclusions about the advantages and disadvantages of each mode. In addition, the presentation will highlight ways in which measurement error differs systematically between the various modes. Thus, the idea of mixing modes in a single survey or in a time series means that results may well differ as a result. The talk will offer suggestions regarding ways to manage and minimize data distortions associated with mode mixing.