Response Styles in Telephone and Household Interviewing: A Field Experiment

Two independent groups from the Los Angeles metropolitan area were sampled and interviewed—one was sampled by a form of random digit dialing and interviewed by telephone; the other was sampled on an area probability basis and interviewed face-to-face in households. Few sociodemographic differences were found between the telephone and household samples, suggesting that the two methods sampled essentially the same population. There were differences in the quantity and quality of the data obtained, however, and the telephone sample had more missing data for family income, more acquiescence, evasiveness, and extremeness response bias on attitude questions, and more responses to checklists.