ABSTRACT
The audience plays an important role in the journalistic context. The audience implications of journalistic work and output, business matters and policy making requires knowledge of public preferences and whether and how they take part in the news. Mixed-mode survey designs including Web and mail are becoming increasingly common in audience research, but it is unclear how mixing the survey modes affects the results. With the possibility of increasing response rates, is the risk of degrading the quality of answers. As audience research constitutes an important cornerstone of journalism studies, and because survey design including the Web is becoming more and more attractive, it is relevant to study the implications of merging audience data from different collection modes. This large-N survey provides a unique opportunity to analyse representative data on the matter. Altogether, 6603 respondents completed the survey, of whom 7.5 per cent chose the Web mode. Both habitual questions and opinion questions on different media were included in the study. The largest differences were found for item non-responses, whereas response distribution among respondents only differed to a small extent due to mode. There is no urgent need for caution when combining mail and Web responses in one data-set. However, thorough analyses of item non-response are necessary to control measurement errors when surveying media audiences and habits of using digital and other media in mixed-mode designed surveys.
