Information Cocoon Effect And The News Consumption Behaviour Modes From The Perspective Of Structuration Theory

With the rise of mobile internet,mobile reading has become one of the mainstream news reading methods,but the accompanying algorithmic personalization also deepens people's concern about the effect of information cocoon.Information cocoon effect is also referred to as "echo chamber effect" or "selective exposure",which refers to people's(actively or passively)biased acceptance of information.In previous studies of similar phenomena,the researchers' data were mainly derived from readers ' selfreporting or laboratory tests,but with the advent of the big data era,the behavioral data generated by people on the Internet made the study of information cocoon more likely especially in terms of data integrity and reliability.Therefore,this paper uses the real behavior data generated by a mobile news application to re-study the information cocoon effect in the information age.By using the research methods in the field of computational social science,this paper makes a dynamic analysis of people's real-world news consumption behavior,and guarantees the reliability and authenticity of the research data,revealing the complexity of people's daily life in their behavior level.At the same time,the rise of computational social science research paradigm has brought about the re-discussion and understanding of interpretive sociological theory and microscopic perspective.Therefore,this paper reunderstands and uses Giddens' structuration theory through the reappearance of the real situation of the audience using the news application,and takes this as the perspective,considering the role of the news consumer as an actor and the algorithm of the news application as the structure in the practice of information acquisition.By emphasizing the interaction between the two in the time dimension and the resulting daily practice(daily behavior flow),this paper analyzes the information flow of readers' in the process of consuming news at the level of microscopic dynamic behavior,and obtains two behavior patterns that conform to the characteristics of the information cocoon effect.This paper also analyzes the influencing factors that affect the occurrence probability of these two modes.In the emphasis on the daily practice,structuration theory also emphasizes the "critical situations" for the destruction of this daily practice.This article corresponds this theory to the process of consuming news,and finds that the reader's daily reading content can be interrupted by sudden trending events.Therefore,there is a competitive relationship between the content in the audience's daily practice and the sudden trending events at the attention level.Through the analysis of this process,this paper explores the microscopic selection process of people in the information cocoon.Therefore,this paper is divided into two parts.The first part discusses the two reading modes presented by the information flow selected by the audience,and explores the differences between different groups,as well as the changes of these two behavior patterns with time.The second part of the research design explores how the information cocoon affects the reader's participation in the discussion of trending events,and how this participation can reverse affect people's news reading behavior patterns,to show the microscopic changes of people in information selection.The data used in this article is generated from January to March,2017,the news review data of a news application users in China.For the behavior pattern of the audience in information cocoon,this paper takes the news theme sequence that the news application users continuously consumed on the time dimension as the research object,and uses the Relational Event Model to analyze it.From the individual level,the reader's reading of different news types constitutes a continuous sequence of discrete events in time,and the individual's news reading behavior is on the one hand the path dependence effect of past reading habits,on the other hand,it is influenced by the latest recommendation.The model results show that it is easier for users to exhibit two modes of behavior,continuous and walking type,in the process of hard news consumption.And with the change of time,different news reader's behavior pattern will also change in different degrees.On the other hand,this paper analyzes the attention distribution between daily reading practice content and trending events,and uses the survival model to explore the reason why audience in different information cocoons choosing to participate or not in the discussion of trending events.The results show that when trending events erupt,soft news readers will be more likely to participate in the discussion of these events,while hard news readers are less likely to participate.At the same time,once participate in discussing trending events,these events can help them transform between different news types(hard and soft news),compared to non-trending events.Therefore,combining sociological theory as a research perspective,this paper analyzes the specific process of information cocoon effect through two models at the microscopic dynamic level,and presents the microscopic state and performance of audience information selection.This study not only helps to enhance the understanding of the information cocoon effect,but also provides a practical path for computational social science researchers to combine traditional theories with data mining methods.