커뮤니티 브랜드에 대한 외부의 위협으로 인해 촉발된 온라인 커뮤니티에서의 소비자들간의 다방향적(multi-directional) 의사소통 과정에서는 다양한 사회적 요인들이 커뮤니티 및 커뮤니티 브랜드 평가에 영향을 미칠 수 있다. 본 연구에서는 이러한 영향을 구체적으로 알아보기 위해 사회정체성 이론의 게시자 전형성과 (열성회원 vs. 신입회원) 댓글의 성향의 (공격적 vs. 정중한) 상호작용을 중심으로 도출된 예측을 가상의 온라인 커뮤니티를 배경으로 한 실험을 통해 검정하였다. 외부의 위협에 대해 열성회원이 공격적인 댓글을 게시했을 경우 커뮤니티 독자들이 1) 댓글을 호의적으로 평가하였으며 2) 커뮤니티 및 브랜드에 대해 긍정적인 태도를 보였으며, 3) 더 낮은 집단 수치심을 보였으나, 4) 열성회원이 공격적인 댓글을 게시하던 혹은 정중한 댓글을 게시하던 커뮤니티 독자들이 느끼는 집단 자긍심에는 차이가 없는 것으로 나타났다. 이에 반해 신입회원이 정중한 댓글을 게시했을 경우에 커뮤니티 독자들이 1) 댓글을 호의적으로 평가하였으며, 2) 커뮤니티 및 브랜드에 대해 긍정적인 태도를 보였으며, 3) 더 낮은 집단 수치심을 보이고, 4) 더 높은 집단 자긍심을 가치고 있는 것으로 나타났다. 또한 매개된 조절효과 검정 결과 집단 자긍심과 집단 수치심은 게시자 전형성의 조절효과를 댓글 평가, 커뮤니티 빛 브랜드 태도에 부분적으로 매개하는 것으로 나타났다. 본 연구는 커뮤니티 내 의사소통과정에서의 집단 정체성의 영향을 이해하는 새로운 시각을 제시함으로써 마케팅 실무자들이 보다 효과적인 온라인 마케팅 커뮤니케이션 전락을 수립하는데 도움이 되는 많은 시사점을 제공하고 있다.

Many communications that consumers participate and observe in online communities show multi-directional characteristics of a complicated social network. When community members encounter an outgroup member`s negative posting on the community brand, they often generate very active multi-directional intra-community discussions by placing replies against the negative posting. The current research attempts to examine such multi-directional communications with a theoretical framework based on Social Identity Theory (Tajfel 1982 Turner 1987). Self-categorization process of Social Identity Theory describes the underlying cognitive process of group identification, and provides an interesting perspective for understanding intra-community communication process. While the majority of theories on persuasion often emphasizes intra-or inter-individual processes, self-categorization theory provides an alternative framework of persuasion based on the shift in the locus of self-perception from an individual level to a collective level. Such sift in the locus of the self provides a theoretical basis that can explain why and how characteristics of group membership may influence intra-community communications. The community members would perceive the negative posting as a threat to the core value of their online community. Such identity threat often increases the perceived uncertainty of social life in the online community, and community members would develop various behavioral strategies to manage the impact of the threat. Further, community members would also form shared expectations on how other community members should response to the threat. We propose that group prototypicality, which describes a community member`s representativeness or centrality in the community, plays an important role in shaping expectations on a community member. Facing the external threat, community members would expect an active member with high group prototypicality should post very aggressive replies that can damage the credibility of the negative posting, and lessen other community members` perceived uncertainty. On the other hand, community members would expect a new member with low group prototypicality should post rather civil replies that can help community members to perceive their community to be more agreeable with the higher level social norms of "Netizenship" behavior. In other words, community members do not expect all the community members should respond to an identity threat in the same way, but rather develop multi-faceted collective strategies to protect their community and the community brand. Depending on the level prototypicality of a community member, community members expect different responses toward the negative posting, and use them strategically to protect positive images of the community and/or manage social uncertainties, Therefore, when a community member`s reply is consistent with community members` strategic expectations, they would evaluate it more persuasive and perceive the community and the community brand more favorably than when his/her reply is inconsistent with the expectations. We further propose that the collective emotions of shame and guilt will mediate the moderating effect of group prototypicality to the dependent variables: message persuasiveness, community attitude, and brand attitude. We tested the proposed predictions in an imaginary online community of a notebook computer and found the following results. First, when the writer of a reply was an active community member with high group prototypicality, community readers evaluate an aggressive reply more favorable than a civil one. They also developed more positive attitudes toward the community and the community brand when the active community member posts an aggressive reply. Second, when the writer of a reply was a new community member with low group prototypicality, community readers evaluate a civil reply more favorable than an aggressive one. They also developed more positive attitude toward the community and the community brand when the new community member posts a civil reply. Third, the interaction between group prototypicality and message aggressiveness influences community readers` collective shame and collective pride. Community readers experience higher level of collective shame when the active member posts a civil reply and when the new member posts an aggressive reply against the threatening posting. On the other hand, community readers experience higher level of collective pride when the new member posts a civil reply than an aggressive one. For the active member`s postings, community readers showed no significant difference between aggressive and civil replies. Fourth, a series of regression analyses were conducted to test the mediated moderating effect (Muller, Judd. and Yzerbyt 2005) of collective emotions, and the results suggest that collective shame and collective pride partially mediated the moderating effect of group prototypicality to the dependent variables. The current research provides interesting theoretical and managerial implications. In our experiment, we observed a significant interaction effect between group prototypicality and message characteristics. This finding suggests that multi-faceted social interaction between community members create a high level social network in an online community. and generate a group based source effect. In addition, marketers often attempt to provide continuous influx of positive information through online communities and curtail negative information as much as possible. However, such marketing practice often backfired and consumers were suspicious about the authenticity of the positive information. Our results imply that marketers, instead of providing positive information all the time, may use some types of negative information as a trigger that increase the group cohesion of an online community and build strong brand communities and brand loyalty.