"Vote early and often" is an old American joke. The funny thing is, we now do that year in and year out in the form of endless public opinion polls that have reached the point of being confused with elections and with public opinion itself. Michael Wheeler's is the first book to take on the polls and the pollsters and deal forthrightly with the questions the people must ask.

For example, can polls be rigged? Do they affect voters? Do they affect policy decisions? According to the author, the answer to each of these questions is yes. To back all of his considerable digging, he makes use of direct interviews with leading pollsters like George Gallup and Louis Harris, and a host of politicians, ranging from Senators Kennedy and Humphrey to former Nixon aides Colson and Ehrlichman.

In one revealing chapter he follows a typical interviewer around for a day and gives the reader a real feeling for the ragged edges of opinion sampling. Even the pollsters admit that their record is marred by serious errors. The author shows that it is very much worse than they will admit.

Lies, Dam Lies, and Statistics is not only about political polls, even though that is its most important subject. There is a section on commercial polls, which may be even less reliable than the others, and one on polls and television, in which the emperor's clothes of the Nielsen ratings are described. There is also a chapter on policing the polls, which the public ought to find imperative in view of the rest of the book. An exposé with an alternative, this is a book whose time has come.