A disposição para revelar informações pessoais a sistemas de recomendação: um estudo experimental

• Ano de defesa: 2017
• Autor(a) principal: Oliveira, Bruna Miyuki Kasuya de
• Orientador(a): Sanchez, Otávio Próspero
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Palavras-chave em Português: Preocupações de privacidade na internet; Paradoxo de privacidade-personalização; Sistema de recomendação; Reatância psicológica; Efeito bumerangue
• Palavras-chave em Inglês: Internet privacy concerns; Privacy-personalization paradox; Recommendation system; Psychological reactance; Boomerang effect
• Link de acesso: https://hdl.handle.net/10438/18712

Information privacy on internet is one of the biggest concerns that arise with web 2.0. However, it is increasingly common for companies that use Recommendation Systems (RS) the request and manage of personal data aiming to guarantee personalized services or products to the users. However, consumers often face a privacy-personalization paradox because they need to provide information, but fear how companies will use it. Incoherent use of such data can give to the individual the feeling that their freedom is being curtailed, causing reactions differently than the system’s intention. It is a boomerang effect, understood as an opposed response to the threat of its freedom on the web. Considering that the IS literature insufficiently explores the effects of the perception of intrusion on the willingness to disclose information, especially through the theory of psychological reactance – where the boomerang effect comes from – the objective of this research is to verify how the users' perception of the intrusion of the Recommendation System may affect your willingness to disclose your information. Two experiments were conducted in the United States and Brazil, with valid samples of 213 and 237 participants, respectively. A prototype of an Experimental Recommendation System (ERS) was developed on the Qualtrics platform. The techniques used for data analysis were the analysis of one-way variance (one-way ANOVA) and covariance analysis (ANCOVA). Among the results, the boomerang effect of RS was demonstrated, because the higher the level of SR intrusion, the less is the willingness to disclose its information. It was verified the existence of only two levels of intrusion perceived by the user. The impact of Internet privacy concerns on the relationship between perception of intrusion and willingness to disclose information was verified, as well as the behavioral indifference between the two samples. Based on the results, RS developers and companies that use them are expected to avoid future boomerang effects in their recommendations, which would scare away a potential customer.