The role of public opinion polling in health legislation.

Two public opinion polls, in Los Angeles and Michigan, on smoking in public places contradicted findings from other major surveys. Both were commissioned by the tobacco interests during consideration of non-smoking legislation to measure opinion and influence the outcome. Analysis reveals the two polls to be remarkably similar and in violation of basic principles of survey research. Response categories lacked objective or parallel phrasing, response lists were not rotated, and the ordering of items appeared prejudicial. Both polls were effectively countered and the legislation passed.