Reducing Public Persuasion by and Contribution to Fake News: Developing a socio-cognitive model and intervention for persuasion by misinformation

本研究项目的总体目标是发展一种扩展的接种干预。这种干预将增加社会背景和社会偏见方面的培训，不仅提供一种干预策略，赋予对说服的抵抗，而且还减少重新分发错误信息的倾向。这种接种策略治疗应该：1)提高在社会环境中识别错误信息的能力；2)降低信念改变的可能性；3)降低分享错误信息的倾向。为了成功地开发这种干预措施，将测试和修订一个概念性的社会认知模型，以更好地理解在错误信息环境下认知和社会影响之间的相互作用。10月19日至3月20日：探索与错误信息相关的认知和行为之间的关系研究的第一阶段将是探索性的，将调查围绕错误信息的判断、信念和行为之间的关系。这一阶段的目的是揭示识别欺骗的能力是否与将欺骗性信息中的信息定性为虚假有关，此外，这是否会阻碍个人分享此类信息。3月20日至3月21日：共识对错误信息检测影响的实验研究研究第二阶段将采用参与者内设计，以共识为自变量，错误信息检测为因变量，研究社会共识对错误信息检测能力的影响。之前的研究使用了pre-post设计，但是为了在社交环境中测试这些影响，其中社会共识的水平会根据信息而变化，将使用Javascript（特别是jsPsych）开发一个基于行为试验的研究，并在psiTurk上运行。21 / 03 - 21 / 12：测试分享错误信息倾向的外在和内在社会线索研究的第三阶段将检验共识（外在社会线索）和归属需求（内在社会线索）对分享错误信息可能性的潜在影响。本实验将采用克拉克森等人（2013）的混合方法设计。参与者（N=200）将在在线平台多产学术上招募，并随机分配到社会认同条件（认同需求：归属感或独特性）或对照组（无身份操纵）。参与者的识别需求将按照克拉克森等人（2013）使用虚拟人格测试的方法进行操纵。随后，参与者将被要求评估他们分享一些真实和虚构新闻的可能性，并进行共识操纵。在这里，可以看到一个包含2000多个先前响应的虚拟数据库，这样就有a)共享的高共识b)共享的中等共识c)共享的低共识或d)可用信息不足。最后，参与者将根据先前对社交媒体应用的使用和满足理论的研究（如Ku， Chu & Tseng, 2013; Whiting & Williams; 2013）完成一项调查，以评估社交满足需求。12/21-08/22：错误信息社会认知模型的修正与社会接种干预的发展。以上研究结果将指导本项目最后阶段的研究。在此，我们将根据上述研究的结果对所提出的模型进行重新审视。更新后的模型将用于开发一种修订的接种方法，该方法将纳入培训，以克服错误信息过程中说服的社会影响。这种接种方法将整合社会观点，旨在减少可靠性判断、信念变化和错误信息的共享。

The overarching objective of this research project is to develop an extended inoculation intervention. This intervention will add training in social context and social biases to provide not only an intervention strategy for conferring resistance against persuasion, but also to decrease propensity to redistribute misinformation. This inoculation strategy treatment should: 1) Increase ability to identify misinformation in a social context 2) Decrease likelihood of belief change and 3) Decrease propensity to share misinformation. To successfully develop this intervention, a conceptual socio-cognitive model of persuasion by misinformation will be tested and revised to better understand the interaction between cognition and social influence in a misinformation setting. 10/19-03/20: Exploring the relation between cognition and behaviour related to misinformation The first phase of the research will be exploratory in nature and will investigate the relationship between judgements, beliefs and behaviour surrounding misinformation. The purpose of this phase is to uncover whether or not ability to recognise deception is linked to characterising information contained in deceptive messages as false, and furthermore, whether this discourages individuals from sharing such messages. 03/20-03/21: Experimental study on the influence of consensus on detection of misinformation The second phase of the research will investigate the effects of social consensus on ability to detect misinformation using a within-participant design with consensus as the independent variable and detection of misinformation as the dependent variable. Previous research has investigated this using a pre-post design, but to test these effects in a social setting, in which level of social consensus changes depending on the message, a behavioural trial-based study will be developed using Javascript (specifically jsPsych) and run on psiTurk. 03/ 21- 12/21: Testing extrinsic and intrinsic social cues on propensity to share misinformation The third phase of the research will examine the potential effects of consensus (extrinsic social cue) and need to belong (intrinsic social cue) on likelihood of sharing misinformation. This experiment will use a mixed methods design adapted from Clarkson et al (2013). Participants (N=200) will be recruited on the online platform Prolific Academic and randomly assigned to a social identification condition (identification need: belonging or uniqueness) or a control (no identification manipulation). Participants identification needs will be manipulated following the method by Clarkson et al (2013) using a fictitious personality test. Subsequently participants will be asked to rate their likelihood of sharing a number of real and fictitious news items paired with a consensus manipulation. Here, a fictitious database of 2000+ previous responses in the database will be visible, such that there is either a) High consensus in sharing b) Medium consensus in sharing c) Low consensus in sharing or d) Not enough information available. Finally, participants will complete a survey based on previous research on uses and gratifications theory applied to social media (e.g. Ku, Chu & Tseng, 2013; Whiting & Williams; 2013) to assess social gratification needs.12/21-08/22: Revision of socio-cognitive model of misinformation and development of social inoculation intervention The findings obtained in the above studies will guide the final phase of this research project. Here, the proposed model will be revisited based on findings from above studies. The updated model will be used to develop a revised inoculation method which will incorporate training to overcome social effects in the persuasion by misinformation process. This inoculation method will integrate social perspectives and will aim to reduce both reliability judgements, belief change and sharing of misinformation.