ABSTRACT
In many campaigns, the candidate simply can't afford to pay a professional research consultant to do a sample survey. Furthermore, strange things happen in campaigns when "bad" data are produced: Other people hear of the "study"— or are even told about it by members of the campaign staff— and comments are made to reporters. In campaigns in smaller communities, it may not be necessary to screen. First, we can readily determine the universe of most likely voters by examining the records of the Board of Elections. In many political surveys, the sample should be picked from the voter registration books that are usually available from the local board of elections. State laws generally require that these books be published, and frequently they are available in local libraries. In a primary contest, of course, the sample is picked only from among those voters who are registered in a single political party.
