联系我们

论文查询结果	返回搜索

论文编号：	14307
作者编号：	2120213484
上传时间：	2023/12/11 17:48:18
中文题目：	短视频平台算法特征对信息茧房的影响分析
英文题目：	The Impacts of Algorithmic Features on Information Cocoons in Short-Video Platforms
指导老师：	李月琳
中文关键字：	信息茧房;算法特征;短视频平台;信息素养
英文关键字：	Information cocoons; algorithm features; short-video platform; information literacy
中文摘要：	在短视频情境中，过滤和推荐算法有可能将用户困在信息茧房中，强化其已 有的信念并阻碍其批判性和创造性思维。为了探究影响信息茧房的技术和认知 因素，本研究开发了一个理论模型来检验算法特征和信息素养对信息茧房形成 的影响。为了测试理论模型，本研究进行了用户实验，选取抖音短视频社交应用 软件作为实验系统，招募了 48 位实验参与者，并进行多元线性回归分析。研究 结果表明，算法特征显著影响信息茧房，是信息茧房形成的主要原因。算法特征 主要通过以下三种机制影响信息茧房:多样性、新奇性和个性化。多样性机制表 明，当用户接触到来自不同来源的视频时，他们对熟悉来源的偏好会加剧信息茧 房。新奇性机制表明，新颖的视频会加剧信息茧房，但与用户相关的、有价值的 和有趣的新奇视频即意外发现有助于缓解信息茧房。个性化机制强调满足用户 多样化信息需求和表现出情感同理心的算法有助于缓解信息茧房，而优先考虑 用户兴趣一致性的算法反而会加剧信息茧房。 此外，信息素养对信息茧房的形成起着次要作用。一方面，信息素养直接正 向影响信息茧房。自评估信息素养的增加可能导致用户低估在短视频平台中陷 于信息茧房的客观风险和自身的过度自信，从而加剧信息茧房。另一方面，信息 素养在算法的来源多样性和新颖性对信息素养的正向影响中发挥显著的调节作 用。自评估信息素养可以增加用户的感知算法意识，减轻来源多样性和新颖性算 法特征对信息茧房的正向影响。 理论意义方面，本研究为算法引起信息茧房的原理提供了新的见解。本研究 提出了来源多样性、新颖性和兴趣一致性加剧信息茧房以及意外发现、需求一致 性和情感一致性降低信息茧房，并分析了上述影响的认知原理。此外，本研究还 发现了信息素养对信息茧房形成的双面影响，即直接的加剧作用和对算法导致 的信息茧房的调节作用。实践意义方面，本文提出短视频平台优化多样性、新奇 性和个性化算法设计，支持用户突破信息茧房的建议。此外，本研究提出数智时 代革新信息素养教育以提升用户应对算法负面影响的建议。
英文摘要：	In the context of short videos, filtering and recommendation algorithms have the potential to ensnare users in information cocoons, reinforcing their pre-existing beliefs and impeding their critical and creative thinking abilities. To address this issue, this study develops a theoretical model to examine the influence of algorithm features and information literacy on information cocoons formation. To test the theoretical model, this research conducted a user experiment using Douyin short video social application as the experimental system. A total of 48 experimental participants were recruited, and multiple linear regression analysis was conducted. The research findings indicate that the algorithm features have a significant impact on the formation of information cocoon, establishing it as the primary factor contributing to its formation. Algorithmic features mainly affect information cocoons through the following three mechanisms: diversity, newness, and personalization. The diversity mechanism suggests that users’ inclination to prefer familiar sources can exacerbate information cocoons when they encounter videos from diverse sources. The newness mechanism demonstrates that novel videos can exacerbate information cocoons, but relevant, valuable, and interesting new videos, i.e. serendipity, can break them. The personalization mechanism emphasizes that algorithms satisfying users’ diverse information needs and exhibiting affective empathy can break information cocoons, while those prioritizing consistency with user interests can make them worse. Moreover, information literacy plays a minor role. On the one hand, information literacy directly and positively affects information cocoons. In the context of short- video platform, an increase in self-reported information literacy among users may lead to a tendency to underestimate the objective risk and become overconfident in avoiding information cocoons, thereby exacerbating information cocoons. On the other hand, information literacy plays a significant moderating role in the positive impact of source diversity and novelty on information literacy. Self-reported information literacy can increase users’ awareness of algorithms, alleviating the effects of source diversity and novelty on information cocoons. In terms of theoretical implication, this study provides new insights into the principle of information cocoons caused by short-video platform algorithms. This study proposes a cognitive mechanism by which source diversity, novelty, and interest consistency intensify information cocoons, and serendipity, need consistency, and affective consistency decrease information cocoons. In addition, this study revealed a dual impact of information literacy on the formation of information cocoons, including both the direct exacerbating effect and the moderating effect on information cocoons induced by algorithms. In terms of practical implication, this paper proposes suggestions for short-video platform to optimize diversity, newness and personalized algorithm design, and support users to break through information cocoons. Moreover, this study puts forward suggestions for innovating information literacy education in the era of digital intelligence to improve users' ability to cope with the negative impact of algorithms.
查看全文：	预览  下载(下载需要进行登录)