Trust in advisors: Origins, effects, and implications for risk communication

It is now well recognised that trust in sources of advice is important for effective risk communication. To maintain some influence over peoples opinions and risk-taking behaviours, government departments, medical bodies and other agencies must endeavour to maintain the trust that people have in them. When trust is lost, so is influence. As an example of this, commentators have pointed to effects of poor advice during the bovine spongiform encephalopathy (mad cow) crisis. The government assured people that it was safe to eat beef but later this behaviour was later linked to a brain disease that is always fatal. This reduced trust in the government as a source of advice about risk for many people. As a result, there appears to be a reluctance to take government advice on other matters. For example, the number of measles cases has increased in many parts of the country: fewer people are having their children vaccinated against the disease because they do not believe the governments assurances that the MMR vaccine is safe. Recently, Onora ONeill, in her 2002 Reith lectures, has argued that we should be cautious about accepting peoples claims that they have lost trust in some agency as evidence that they have actually lost their trust in it. For example, she suggests that many people who say they no longer trust supermarket food still shop in supermarkets rather than elsewhere. For her, it is important to make a distinction between stated trust and actual trust. Our aim is to find out more about trust by answering a number of questions about it while keeping ONeills distincion in mind. What determines trust? How should it be measured? Does it have similar effects on peoples estimates about levels of risk and their risk-related behaviours? How valid is the view that it is easier to destroy than to create trust? Is trust in advisors modified in a rational way? If not, what determines how it is modified? Does trust (or lack of it) transfer across domains? Answers to these questions should be relevant to many of the concerns of those responsible for the development of effective risk communication strategies in government, business and non-profit-making organizations. This is a proposal for systematic studies to answer these questions. In simulated situations, people will be required to assess levels of risk associated with various hazards on the basis of information received from advisors. Advice will be attributed to different sources, such as government departments and consumer organizations. Its quality will be under our control. We shall be able to measure actual trust in an advisor (in terms of the degree to which peoples own risk judgments depend on the risk estimate provided by that advisor) and stated trust in that advisor (in terms of a trust rating). We shall be able to determine how these measures depend on quality of previous advice from that source and on other variables, such as the degree to which judges perceive their values to be similar to those of the advisor

如今学界已达成共识：对信息来源的信任，对于开展有效的风险沟通（risk communication）至关重要。为维持对公众意见与风险决策行为的影响力，政府部门、医疗机构及其他相关机构必须竭力维系公众对其的信任。一旦信任崩塌，影响力亦随之消散。

对此，评论者曾以牛海绵状脑病（bovine spongiform encephalopathy，俗称疯牛病）危机期间的不良建议影响为例展开说明：当时政府向民众保证食用牛肉是安全的，但后续该行为被证实与一种致命性脑部疾病存在关联，这导致许多民众对政府作为风险咨询来源的信任大幅下降。受此影响，民众似乎不愿再听从政府在其他事务上的建议——例如，该国多个地区的麻疹病例数有所上升，由于民众不再相信政府关于麻腮风三联疫苗（MMR）安全的保证，越来越少的家长愿意为子女接种该疫苗。

近期，奥诺拉·奥尼尔（Onora O'Neill）在其2002年的瑞斯讲座（Reith Lectures）中提出：我们不应仅凭民众声称自己不再信任某一机构，就认定他们确实丧失了对该机构的信任，对此需持谨慎态度。例如，她指出，许多声称不再信任超市食品的民众，依然选择在超市购物而非其他渠道。在她看来，明确区分宣称信任（stated trust）与实际信任（actual trust）十分关键。

本研究旨在结合奥尼尔的上述区分，通过解答一系列与信任相关的问题，进一步深化对信任的认知。具体而言，我们将探讨以下问题：信任的决定因素是什么？应如何对信任进行量化测量？信任对民众的风险评估与风险相关行为是否具有相似的影响？“信任易毁难立”这一观点的合理性如何？对咨询者的信任是否会以理性方式发生变化？若非如此，那么信任的变化又由何种因素决定？信任（或信任缺失）是否会在不同领域间发生迁移？对这些问题的解答，将有助于政府、企业及非营利组织中负责制定有效风险沟通策略的人员解决诸多相关关切。

本提案旨在通过系统性研究解答上述问题。研究将设置模拟场景，要求参与者根据咨询者提供的信息，评估各类危险源对应的风险等级。咨询者的身份将被设定为不同机构，例如政府部门与消费者组织，且我们可操控建议的质量。我们将能够同时测量参与者对咨询者的实际信任（以参与者自身的风险判断是否依赖该咨询者提供的风险评估值为衡量标准）与宣称信任（以信任评分作为衡量标准）。此外，我们还可探究这两类衡量指标如何受该咨询者过往建议的质量，以及其他变量（例如参与者感知到的自身与咨询者价值观的契合程度）的影响。