More positive or more extreme? A meta-analysis of mode differences in response choice

Some researchers have argued that respondents give more extreme answers to questions involving response scales over the telephone than in other modes of data collection, but others have argued that telephone respondents give more positive answers. We conducted a meta-analysis based on 18 experimental comparisons between telephone interviews and another mode of data collection. Our analysis showed that telephone respondents are significantly more likely than respondents in other modes to give extremely positive answers (for example, the highest satisfaction ratings in a customer satisfaction survey) but are not more likely to give extremely negative responses. This tendency to give highly positive ratings appears to be related to the presence of an interviewer, and it may reflect respondents' reluctance to express bad news, a tendency some social psychologists have dubbed the MUM effect.