Methodological Guidelines for Social Listening Research: A Case Study of Social Media Users’ Dialogue about PM 2.5

Abstract

        The rapid expansion of social media use has generated immense volumes of user‑generated content that reflect opinions, experiences, and emotions in everyday life. In conjunction with the development of social listening tools, such data allow researchers to access and interpret public attitudes toward diverse issues in near real time, based on naturally occurring communication. This article aims to propose a methodological guidedance for applying a social listening tools to conduct research, informed by a case study of Facebook dialogue regarding PM 2.5 research. The proposed framework encompasses six stages: 1) defining research objectives and questions 2) specifying the time frame and platform scope and selecting appropriate social listening tools 3) constructing keyword sets 4) data collection 5) data cleaning and preparation and 6) data analysis. Particular emphasis is placed on the data cleaning and preparation stage, which operates both at the stage of collecting data and at the stage of cleaning data after data files were exported through the condition of excluded accounts and remove some content types for instance the advertising posts, spam, promotional and commercial links, and off‑topic comments, supplemented by qualitative screening prior to coding. The study highlights both the potential of social listening data to capture public environmental discourse and important limitations related to population representativeness, keyword and platform selection biases, and dependence on commercial social listening services. It is recommended that social listening–based research be combined with other research designs to mitigate these limitations and enhance the quality of PM 2.5‑related studies using social media data, while offering a methodological reference that can be adapted to other environmental issues.