Methodological Bias in Public Opinion Surveys

摘要: Statistically designed sample surveys have enabled pollsters to gauge public opinion on a wide range of issues. In such surveys, selection data collection technique is generally based four criteria: (1) cost; (2) completion time; (3) response rate; and (4) bias. Typically, more weight placed the first three factors and, as result, adequate attention has not been given latter consideration. The study described in this paper looks at one type bias-that which results from use specific method. More specifically, research uses controlled experimental design order determine whether responses polling are influenced by method used collect data. Three methods investigated: mail questionnaire; telephone interview; personal interview.