科技部專題研究計畫主持人
MOST108-2410-H-002-174-SS2
臺灣社群媒體中的假新聞、同溫層與極化現象:一種人機整合創新的研究途徑
Fake News, Echo Chamber, and Polarization in Taiwan’s Social Media: An Innovated Human-Computer Integrated Approach
科技部
Ministry of Science and Technology
張佑宗
Yu-Tzung, Chang
張佑宗
Yu-Tzung, Chang
台灣大學 政治學系
Department of Political Science, National Taiwan University
曾煥凱
助理研究員
0920417949
hktseng1983@ntu.edu.tw
計畫執行期間起:2019-08-01
計畫執行期間迄:2022-07-31
2021-11-01
2022-02-25
台灣地區具有國籍,20歲以上之公民為研究母體
1
007
01
社群媒體、民意、假新聞、群眾外包式、同溫層、兩極化、臺灣
social media, public opinion, Fake News, Echo Chamber, Polarization, Taiwan
本研究計畫旨在彌補國內現存研究對問卷研究未盡詳實之處,結合新式的研究方法對此兩因應社群媒體萌發的假新聞議題作一系統性、科學化的研究。本研究所欲回答的主要研究課題有二。首先,假新聞的類別及這些不同種類的假新聞對閱聽者所持政治立場的影響:假新聞是否強化擁有特定政治立場之閱聽者對其所持立場的堅信抑,或當新聞內容與閱聽者政治立場不同時,改變後者的政治立場。第二,社群媒體平台上同溫層的生成及內部成員的互動:成員交流的內容是否逐漸同質化(homogenize)、無視不同立場的圖文內容、又在同溫層環境已形成的狀態下,內部成員的立場是否會因暴露於不同立場圖文內容而改變? 針對上述課題,本研究欲進行施測的客體為使用社群網路表達政治意見、具備潛在投票權的本國民眾,而此樣本選取條件也決定了研究方法的設計。為最大程度地模擬在虛擬的公共場域中具政治意涵的不實圖文資訊對社群媒體閱聽者政治立場及互動行為的效應,本研究開發群眾外包式 (crowdsourcing)網路問卷平台、透過人機互動實驗來分別檢驗假新聞及同溫層現象在受試客體上的交互作用。通過投放全國性簡訊及網路廣告分派問卷配合小額獎勵機制,使本研究得以快速招募本國籍、具地域、職業、年齡、性別代表性的樣本,以問卷進行施測。而在通訊軟體上架設的特定議題組群則使本研究得以運用自動化文本分析技術歸類受測者回應所呈現的個人特質、預測其在相關議題上的立場、藉此了解在社群媒體環境中假新聞現象的本質。
This study addresses the limitations of current survey research practice on investigating “fake news” in the context of burgeoning role of social media in political discourse, marshalling novel research methods to systematically and scientifically engage such phenomenon. The goal of this study is twofold. First, we aim to uncover the typology of fake news and their effects on audience’ political orientation, in particular, if fake news tends to enhance audience’s existing political orientation or, alternatively, when their political orientation differ, would fake news alter audience’s existing political orientation. Second, does the interaction among social media users of the same partisanship tends to “homogenize” their political attitude, giving rise to the so-called “echo chamber,” even in the presence of dissenting views? Would the exposure to fake news content that is of different partisan hue alter the political orientation of members in an echo chamber? To answer these questions, this study focuses on social media users who (i) actively engage in political issues on the social media, and (ii) are Taiwanese citizens of voting age; the study also designs a new set of sampling and surveying methods to satisfy these scope conditions. We develop a crowdsourcing web-based survey platform, using interactive experiment to examine the cross-effects of fake news and echo chamber on audience’s political attitude formation. Through the distribution of nation-wide short message and web ads, augmented with instant quasi-monetary rewards (electronic coupons), we are expected to recruit sufficient respondents with their based on the imposed demographic quota (region, occupation, age, gender). The builtin analytical applications allow us to leverage automated text analysis tools for classifying key individual level features, predict respondents’ political stance on issues being presented to them, which facilitates an understanding of the essence of fake news in an era of social media.
是,已處理
除檢查有無可識別個人身份資料外,另檢查個案編號是否重複、變項及選項數值說明是否與問卷一致或未被定義,發現前述狀況時,將於確認後協助修改。
追蹤清單
下載(0)
申請(0)
遠距(0)