When Congress Investigated the Scientific Believability of Polls