微博平台上传统主流媒体科学类博文的文本特征与传播效果研究

传统主流媒体作为科学传播的重要主体，在科学事件报道和科学知识科普中扮演着重要角色。为更好地了解传统主流媒体在社交媒体上进行在线科学传播的特点和效果，本文探究了主流媒体科学类博文的文本特征及其对传播效果的影响。首先，本研究获取了九大官方主流媒体于2021年全年在微博平台上所发布的全部11万余条博文，根据科学传播相关的关键词筛选出6000余条科学类博文。基于LDA对文本数据集进行主题建模分析，归纳出29个一级主题和7个二级主题，得到主流媒体科学传播的整体主题分布情况。具体主题所囊括的意涵显示，主流媒体既对科学发现和科技创新进行及时且持续的报道，也生产分发与大众密切相关的社会民生、健康等知识普及类内容。其次，本文对抽样得到的样本数据集进行基于人工编码的内容分析，得到样本中每条博文的情感立场和引用源。最后，本文对主题、情感立场和引用源三个文本特征与博文的转发、点赞、评论三个传播效果表征指标之间的关系进行研究。结果显示，主题和情感立场对博文的三个传播指标均产生显著影响，引用源则并无显著影响。主题为社会民生类科学知识普及、持正向情感立场的博文的传播效果显著好于其他博文。大众对具有不同文本特征的科学类博文有着不同的传播积极性，与日常生活知识科普密切相关、更容易引起情感共鸣的内容能够获得更好的传播效果。

Research on the Textual Characteristics and Communication Effects of Scientific Blog Posts by Traditional Mainstream Media on Weibo Platforms

As an important subject of science communication, traditional mainstream media plays an important role in reporting scientific events and popularizing scientific knowledge. To better understand the characteristics and effects of online science communication by traditional mainstream media on social media, this paper explores the textual characteristics of science blog posts by mainstream media and their effects on communication effects. First, we obtained more than 11,000 blog posts published by nine official mainstream media on Weibo platforms in 2021, and then filtered the scientific blogs by the selected keywords related to science communication. Through processing the text data set by LDA-based topic modeling, we summarized 29 primary topics and 7 secondary topics to obtain the overall topic distribution of science communication in mainstream media. From the view of the meaning of the detailed topics, the mainstream media both provide timely and continuous coverage of scientific discoveries and innovations, and produce and distribute popularized content including social and healthy issues that is closely related to the general public. Secondly, this paper conduced a manual coding-based content analysis on the sampled sample data set to obtain the sentiment stance and citation source of each blog post in the sample. Finally, this paper investigated the relationship between the three text characteristics featuring topic, sentiment stance and citation source and the three communication effect characterization indicators of blog posts: forwards, likes and comments. The results showed that the topic and sentiment stance had significant effects on all three communication indicators of blog posts, and the citation source showed no significant effects on three indicators. The communication effect of blog posts with the topic of popularizing scientific knowledge of social and livelihood issues and holding a positive sentiment stance was significantly better than that of blog posts of other categories. The public has different degrees of motivation to disseminate scientific blog posts with different textual characteristics, and topics closely relevant to the daily life and knowledge popularization, and more emotional resonance can trigger better communication effects.