Is it more effective to evoke negative emotions in social advertisements than positive emotions? This study compared positive and negative message framing strategies in social marketing advertisements that aimed to encourage a reduction in meat consumption. This project explored how each strategy influences consumers' attitudes toward the recommended behaviour and investigated the role of emotional and attentional responses to each message framing strategy. The purpose of this research was to determine whether negatively framed messages are more effective than positively framed messages in influencing consumers' attitudes, emotions, and attention. The motivation of the study was to provide formative research for the design of social marketing interventions to effectively influence consumers' attitudes towards advertised causes with the use of message framing, and to advance theoretical understanding of how consumers respond to social marketing interventions. Furthermore, this research attempted to resolve differences between results obtained in previous framing research in the social marketing context. This study uniquely proposed the use of cutting-edge consumer neuroscience techniques to develop a clearer understanding of consumers' emotional and attentional responses to social marketing advertisements. The results were presented from a mixed-method approach, which combined quantitative and qualitative research methods. An experiment was conducted by using two social marketing print advertisements aimed at encouraging a reduction in meat consumption, by highlighting the impact of consuming meat products on animal welfare. Respondents involved in the experiment viewed an advertisement that was either positively framed or negatively framed. The research applied self-reporting methods, as well as consumer neuroscience methods, including facial coding, galvanic skin response (GSR), and eye-tracking, to explore the proposed research framework. The combination of these methods allowed the collection of data on attitudinal, emotional, and attentional responses. The results of this research demonstrated that negatively framed advertisements are more effective in changing consumers' attitudes towards reducing meat consumption than positively framed advertisements. Thus, messages aimed at encouraging a reduction in consumption should highlight the negative consequences of participating in certain behaviours. Neither emotion nor attention were found to mediate the relationship between message framing and attitude. However, positively framed advertisements elicit significantly higher levels of emotional valence; and negatively framed advertisements elicit significantly higher levels of disgust and attention. Social marketers should, therefore, leverage these feelings of disgust; and they should implement negative framing strategies to increase the attention paid to an advertisement. However, educational social marketing interventions should be considered, in combination with negative message framing, to effectively influence consumers' attitudes towards social issues. These findings have provided research for better developing message framing strategies for the communication of sustainable consumption. Furthermore, these strategies contributed to the existing social marketing literature by addressing the lack of information on marketing efforts aimed at reducing meat consumption. This research also filled important gaps in the literature regarding positive versus negative message framing strategies, and social marketing interventions can now be implemented with an increased understanding of how consumers respond to different message framing strategies.

本研究旨在探討網路信任問題，先從網路人類學研究方法開始，找出網路言論可信度的前置變數。根據質性研究結果，發現網站特質、發言人特質、言論內容為可能影響網路言論可信度之變數。並依據此結果進行初步的量化問卷調查，在380份有效樣本中找出明確的影響因素，再同時進行兩個實驗研究，以確認各主要變數間的關係。 研究發現，言論內容提供正面資訊或負面資訊、以及言論內容與發言網友之間熟識程度的交互作用將影響消費者對網路言論的信任程度；而產品特質之差異會干擾上述變數對網路言論信任的影響程度。 / In this research, we try to explore the determinants of online information trust and the relationships between information trust and product attitude. In Study 1, we explored the online information trust by a qualitative research method, Netnography, to understand and shape the basic rationale of the determinants of online information trust. Then we employed an empirical survey on consumers to examine the proposed determinants of online information trust. We ranked the relative trust levels of the determinants, and explored the moderating differences existing between experience goods and credence goods. In the last study, two experiments using 2 (message statements: positive vs. negative) × 2 (perceived social relationships: strong vs. weak) between group factorial design were conducted to examine the causal relations among determinants, online information trust, attitude, and purchase intention. The moderating effects of product categories were examined, too.