The reason why the attempts to improve the problems of election poll reporting have not led to satisfactory results is that (1) the media has low awareness of the limits of representativeness inherent in the method of opinion polls and (2) that they do not meet the expectations of voters, consumers of articles. Therefore, this study aimed to raise awareness of the media in this sense of the problem and to find and suggest specific ways to solve the problem from the expectations and wishes of the audience. To this end, an online survey was conducted targeting voters over the age of 19 who had the right to vote in the 20th presidential election (N = 1,304). As a result of the study, the bias of the survey respondents was confirmed, which can explain what causes the inconsistency in predicting election results with surveys. As a result of comparing the telephone interview method and the ARS method respondent groups, there was no difference in representativeness. The audience's evaluation of the research company and the media was less than 60 points out of 100 points, It is necessary to reduce the difference among the research companies and to strengthen analysis and interpretation when reporting the survey results by the media to improve the evaluation scores. In the process of future policy and institutional improvement, the needs of the audience must be treated as an important consideration, and this study is meaningful in that it provided necessary information in this process.