With limited information processing capacity, people often rely on heuristics, or rule of thumb, to make decisions. In most situations, these heuristics are useful, however, it is possible to result in systematic biases. One of the biases is framing effect, which refers to the phenomenon that the framing message significantly affects how decision maker infer meaning and hence understand the situation. When a specific attribute is framed in positive or negative terms and result in different decision outome, the attribute framing effect occurs. Although a large amount of studies on framing effect has been cumulated, related works on debiasing the framing effect is limited. Based on past literatures, this study proposed a comprehensive framework to suggest and investigate the effect of debiasing strategies, which were developed in terms of sources of bias including individual difference, imcomplete external information and insufficient cognitive effort of decision makers. Four experitmts were conducted in this study. First, the subjective knowledge, objective knowledge and need for cognition are considered as the possible moderator of attribute framing effect. The second experiment aimed to understand that whether the completeness of decision problem and amounts of attributes affect the phemenon of attribute framing effect. The debiasing effect of warning and elaboration were examind in the third and fourth experiments respectively. The results suggested that both subjective and objective knowledge as well as participants¡¦ need for cognition did not moderate the attribute framing effect. Specifically, the attribute framing effect is observed in all groups regardless of the individual differences. Second, the attribute framing effect disappeared when subjects were provided with positive and negative messages simultaneously. Third, attribute framing effect occurred for subjects in one attribute, three and five attribute conditions. That is, one attribute is sufficient for the framing effect to be observed. Moreover, there is an inverted U relationship between subjects¡¦ attitude and the amount of negative attributes. The framing effect was weakened but is not eliminated when the participants were provided with warnings. In addition, weak warning can prevent subjects of high level need for cognition from framing effect, whereas strong warning can eliminate subjects¡¦ framing effect successfully for group of low level need for cognition. Finally, elaboration is the most effective debiasing strategy in this study to eliminate the framing effect.

The internet has served as the virtual world since the beginning of the digital era, and it has provided consumers the valuable source of information and become a fundamental basis of e-commerce by passing the limit of time and distance of offline stores. It is hard to imagine our life without the internet. Because consumers store and access their private and financial information on the internet, information security is even more important than ever. Although many studies demonstrate the importance of information security to consumers, researchers have paid little attention to consumers\' inference processing underlying their perceptions of information security. We investigate how consumers infer and evaluate online information security based on consumer inference making process and metacognitive experience. We argue that consumers\' perceived security could be enhanced by simply increasing complexity, even if that increased complexity is meaningless. It is because consumers have a belief that security is achieved by sacrificing convenience or increasing complexity. We demonstrated that consumers evaluated a website more secure when asked to enter redundant information in Chapter 1. Chapter 2 suggested that disfluency and difficulty of retrieval could increase perceived security because metacognitive experience makes consumers misattribute their feeling of difficulty to technical difficulty. We found that the positive effect of disfluency was held when a product was not security-related. In Chapter 3, we focused on how to improve the accuracy of security judgments. We found that perceived security enhanced by meaningless complexity would be adjusted by asking specific dimensions of security (Confidentiality, Integrity, and Availability), and the positive impact of a disfluency effect could be debiased by providing participants the true source of their subjective difficulty. Furthermore, we demonstrated that consumers\' interpretation about accessibility experience varied depending on what kind of naïve theory was activated. Through a series of experiments, we demonstrated our arguments were valid and these results provided useful insights and implications about consumers\' inference processing and perception of information security. / Ph. D.

How can empirical science, and psychology in particular, be harnessed to avoid or eliminate unwanted biases? The body of work herein explores this question across twelve experiments. The first approach we consider is placing the onus on the individual to root out any already existing bias within him or herself. Chapter 3, for example, presents experiments that assess whether people (viz., jurors during voir dire) can accurately "self-diagnose" when they are irreparably biased by negative pretrial publicity. (The answer is a resounding no). A second approach is to try and avoid letting bias enter the courtroom in the first place. Chapter 4, for example, provides an experimental test of an institutional solution known as blind expertise, wherein certain biases of an expert witness are avoided by having an intermediary pick the expert, and then having the expert render an opinion before knowing which litigant made the request. In Chapter 7, we consider a third approach to handling bias, one that concedes it will exist in the courtroom. Namely, instruct jurors on the existence of bias, so that they can try to weigh it properly. To this end we test a recently enacted New Jersey instruction on eyewitness testimony. We find that jurors do not become more sensitive to low versus high evidence quality, but instead they discount the eyewitness testimony across the board. Across this inquiry, we deploy several novel tactics; in Chapter 5, for instance, we explore how continuous response measurement (CRM) can provide unique insights into the study of reasoning, and in particular how jurors parse trial evidence. We end in chapter 8 with a more general discussion of how behavioral science can be applied across law and policy.