Do Citizen Narratives Matter? Explaining Digital Government Responsiveness through Online Complaints

• In Event: Accountable Government

Abstract

Do citizen narratives influence how governments respond in digital governance systems? Governments increasingly rely on online complaint platforms to gather information and signal responsiveness, yet substantial variation exists in the depth and substance of official replies. Why do some complaints receive detailed, problem-solving responses while others generate minimal engagement? This study explains variation in digital government responsiveness by examining the narrative structure of citizen complaints. Treating online grievances as strategic acts of storytelling rather than simple reports of dissatisfaction, we argue that how citizens construct their claims shapes how governments react.

We draw on the Narrative Policy Framework (NPF) to theorize how micro-level narratives influence administrative behavior within the policy process. Although the NPF has been applied primarily to macro- and meso-level policy debates, its core elements—character, setting, plot, and moral—also structure everyday citizen–state interactions. We extend the framework to the micro level by operationalizing these elements in individual complaint texts. Character captures the identification of victims and vulnerable populations and the attribution of responsibility to specific actors. Setting refers to the policy domain and institutional context of the grievance. Plot reflects causal logic and emotional dynamics, including urgency, injustice, or blame. Moral captures explicit or implicit calls for government intervention. We theorize that these narrative components shape how administrators interpret problems and determine how substantively to respond. 

On the other hand, we conceptualize digital government responsiveness as response quality, capturing the substantive depth of administrative engagement rather than mere acknowledgment. Official replies are coded along four binary dimensions: (1) Problem Explanation, (2) Direct Communication, (3) Specific Information, and (4) Problem Resolution. These dimensions are analyzed individually and as a composite index of responsiveness.

Empirically, we analyze more than 100,000 citizen grievances from a public dataset of submissions to China’s Local Leader Message Board (LLMB). To measure narrative elements at scale, we combine large language model (LLM)–based text classification with systematic human validation. The LLM classifies narrative components across the corpus, while trained coders iteratively validate and refine outputs to ensure conceptual fidelity and measurement reliability. This hybrid approach integrates theory-driven narrative coding with computational scalability, advancing methods for studying large-scale textual data in the policy process.

Our analysis tests whether variation in narrative elements predicts variation in response quality across policy domains and administrative contexts. We find that complaints featuring clearly defined victims, vulnerable groups, emotionally salient plots, and explicit policy morals are more likely to elicit substantively engaged responses, whereas narratives assigning blame or invoking sensitive contexts shape responsiveness in more complex ways. By demonstrating that micro-level storytelling systematically influences bureaucratic behavior, this study advances the NPF and contributes to scholarship on digital governance and the policy process, showing that online complaints are structured narratives that shape how governments interpret problems and deliver solutions.

Author

• Xinyi Wang, The Hong Kong University of Science and Technology
  Presenting Author