Public opinion cue or persuasive argument? : The effect of user comments on personal opinion

다국어 초록 (Multilingual Abstract)

This study explores how and why different types of public opinion cues affect personal opinion. More specifically, after an exploration of the results of 16 different studies on comment effects, this study identified two main issues in previous research. Firstly, previous research did not identify the persuasive effect of comments, which links to the related problem of most of the extant research using long comments which contained argumentation (i.e., argumentative comments). Secondly, previous research did not consider alternate possibilities, such as the direct effect of comments, or the effect of the opinion distribution of comments on perceived public opinion via personal opinion. To address these issues, the present research used a design that separated public opinion cues to three conditions (i.e., argumentative comments, non-argumentative comments and poll results) in a two-wave online survey experiment. The results of the study showed that the effect of comment distribution affected personal opinion via perceived public opinion, but was moderated by comment type, or comment argumentativeness. Although argumentative comments were more effective on personal opinion compared to non-argumentative comments, the effect was only found for majority-opposing conditions. The study also found that an alternative path model, in which comment distribution affects perceived public opinion through personal opinion, was also significant, and that this effect was virtually indistinguishable from the original model. Furthermore, when compared with poll results, there was no significant difference with non-argumentative comments, but argumentative comments were significantly more effective than poll results on the issue. The implications of these results are discussed in further detail.

국문 초록 (Abstract)

이 연구는 다양한 유형의 여론 단서가 개인의 의견에 어떻게, 그리고 왜 영향을 주는지 탐색한다. 좀 더 구체적으로 말하지면, 댓글의 효과에 대한 16개의 각자 다른 연구 결과를 살펴본 후, 이전 연구에서 크게 두 가지 연구문제를 도출해냈다. 첫째, 대부분의 이전 연구는 주장 및 논증을 포함 댓글을 사용하였다. 이는 기존 연구가 댓글의 설득적인 영향을 살펴보지 않았다는 단점과 연결된다. 둘째, 기존의 댓글 효과 연구는 댓글의 직접적인 효과나 댓글이 개인 의견에 먼저 영향을 주어 이 변화된 의견이 여론추정에 영향을 미친다는 등의 대안 가능성을 고려하지 않았다. 이러한 문제점들을 해결하기 위해, 이 연구는 두 가지 차수로 나뉜 온라인 설문조사 연구에서 여론 단서 조건(긴 댓글, 짧은 댓글, 여론조사 결과)을 세 가지로 나눠 이와 같은 문제에 대해 연구하였다. 연구 결과에 따르면, 댓글에서의 의견분포는 여론추정을 통해 개인 의견에 영향을 미쳤지만, 이는 댓글의 형식에 조절되는 것으로 나타났다. 또한, 긴 댓글이 짧은 댓글보다 더 큰 영향을 미쳤지만, 반대다수 조건에서만 이러한 결과 나왔다. 마지막으로, 짧은 댓글, 긴 댓글, 그리고 여론 조사 결과가 개인 의견에 미치는 영향을 비교해보니 짧은 댓글과 여론 조사 결과 간에는 유의미한 차이가 발견되지 않았지만, 긴 댓글과 여론 조사 결과를 비교해보니 긴 댓글 쪽이 더 유의미하게 큰 것으로 나타났다. 이 연구 결과의 함의를 밑에 기술한다.

목차 (Table of Contents)

• I. Introduction 1
• II. Literature Review 5
• Comments as Cues of Public Opinion 5
• Findings of Previous Studies on the Effect of Comments 6
• Table 1 8
• Table 2 11
• Issues in Previous Studies on the Effect of Comment Distribution on Personal Opinion via Perceived Public Opinion 17
• Public Opinion Cues vs. Persuasive Arguments. 17
• Majority Social influence vs. Social Projection 22
• Comments and Poll Results as Information about Public Opinion 25
• III. Method 28
• Study Overview and Design 28
• Participants and Procedure of the Experiment 28
• Stimulus Materials and Manipulation of Independent Variables 30
• Comment Distribution 31
• Recall Test 31
• Comment Argumentativeness and Type of Public Opinion Cues 32
• Measures 32
• Personal Opinion 32
• Personal Opinion Change 33
• Perceived Public Opinion 33
• Perceived Argument Quality 33
• Control Variables 34
• IV. Results 35
• Preliminary Analyses 35
• Manipulation Checks 35
• Hypothesis Testing 36
• V. Discussion 52
• Research Summary 52
• Theoretical Implications 54
• Practical Implications 58
• Limitations and Suggestions for Future Research 59
• VI. Conclusion 62
• References 64
• Appendix 79
• Appendix A. Article on the 2nd Airport at Jeju Island 79
• Appendix B. Experimental Stimuli 80
• Appendix C. Questionnaires 86
• Questionnaire (First Wave) 86
• Questionnaire (Second Wave, Comment Condition) 91
• Questionnaire (Second Wave, Poll Condition) 98
• Questionnaire (Second Wave, Control Condition) 104
• Abstract (in Korean) 109