Research

This page offers an overview of my research journey so far. As a Master’s student, I’m still learning and growing, but each project reflects my dedication and curiosity. I warmly welcome any discussions, feedback, or opportunities to collaborate—feel free to reach out!

Click the sections below for more information. (Last updated: October 2025)

A. Peer-reviewed Journal Papers

A.1. Under Review

From Functional Glitches to Ethical Lapses: Exploring Chinese Users’ Forgiveness for GenAI Chatbots’ Moral Errors

Junheng Qian, Zhi-jin Zhong

As generative artificial intelligence (GenAI) chatbots proliferate, their open-ended outputs increasingly risk producing moral errors, such as gender stereotypes. Existing algorithmic audits have mainly focused on identifying harmful content in GenAI, while overlooking how users perceive and respond to such failures. Likewise, service recovery research addressed chatbots’ functional breakdowns, but paid limited attention to their moral violations. To bridge these gaps, this study draws on Interpersonal Forgiveness Theory (IFT) to investigate: (a) how moral versus functional errors differently influence users’ perceptions and forgiveness intentions; (b) how four traditional recovery strategies (inoculation, anthropomorphism, symbolic apology, and utilitarian apology) perform in promoting forgiveness for moral failures; and (c) the psychological mechanisms that explain the effectiveness of these strategies. Across three online experiments (N=871), we find that moral errors elicit significantly stronger negative reactions than functional ones, confirming the effectiveness of our new value-based classification of service failure. Among four traditional service recovery strategies, only symbolic apology proves effective in moral service failure context. Its perceived sincerity activates all three forgiveness pathways proposed in IFT, significantly enhancing forgiveness intentions. Theoretical contributions and practical applications were systematically discussed.

# Human-AI Communication

# Forgive

# Moral Error

# Service Failure

# GenAI

News Is Becoming Similar? A Diachronic Text Mining Exploration of News Convergence in Chinese Media

Junheng Qian, Zhi-jin Zhong

The convergence of news has long been a central concern in media studies, yet empirical evidence and precise measurements remain limited. China’s unique media system, which operates simultaneously in domestic and global contexts under distinct communication strategies, offers a rare opportunity to advance this line of inquiry. This study draws on news posts (N = 607,871) from four leading Chinese outlets over the past decade and applies a text-mining approach to examine Chinese news convergence across the two contexts and its relationship with audience news engagement. The results show a significant increase in discourse similarity among Chinese media in the global context but not in the domestic context. No upward trend in thematic structural similarity is observed in either context. Moreover, the associations between convergence and audience engagement are context-dependent: in the domestic context, discourse similarity is negatively associated with engagement, whereas in the global context, both discourse and thematic similarity are positively associated with engagement. These results challenge conventional views of news convergence, highlighting its dynamic nature across temporal and contextual dimensions. Further implications for theory and practice are discussed.

# News Convergence

# Text-Mining

# News Engagement

# Social Media

Engagement and Conflict within Narratives: A Computational Study of Narrative Effects in the Context of International Communication

Junheng Qian, Keying Zhong, Zhi-jin Zhong

In international communication, narrative—conveying messages through storytelling—is widely recognized as an effective communicative approach. However, this recognition often entails an overly optimistic view of its efficacy, overlooking the potential tensions and conflicts embedded in narrative practice. To address this gap, this study conducts computational analyses of 2,698 narrative posts and 144,224 associated comments from four Chinese international media outlets to examine three major narrative strategies on social media: value orientation (globalism vs. nationalism), narrative perspective (individual vs. collective), and the use of iconic symbols. Using a multi-level model of narrative effects, the analysis evaluates both the positive and negative outcomes of these strategies. Results indicate that nationalistic and collective narratives correlate positively with oppositional stances and negative emotions in audience comments, whereas iconic-symbolic narratives are linked to higher audience attention, more positive emotions, and fewer oppositional responses than non-iconic ones. Overall, these findings advance understanding of digital narrative dynamics, highlight the variability and dynamism of narrative effects, and offer theoretical insights for narrative practice in international communication.

# Digital Narrative

# International Communication

# Narrative Effects

# Computational Method

A.2. Publication

The Communication Effects and News Convergence of Multiple Actors in International Communication: A Diachronic Study Based on Text Mining. Journal of Journalism and Communication Studies*, 32 (04), 91-108+128.[Source]

Zhi-jin Zhong, Junheng Qian**. (2025).

Despite growing interest in China’s multidimensional international communication strategy, empirical analysis of diverse media actors remains limited. This study examines 492,832 tweets from 60 central, provincial, and municipal-level international communication accounts to assess communication effectiveness (average, cumulative, and marginal effects) and news convergence across discursive, thematic, and categorical dimensions. Results show overall improvement in communication outcomes, but shifting efficiency across actor levels. Notably, average and marginal effects reveal a misalignment between content volume and impact. News convergence patterns vary by media tier and do not consistently enhance performance. These findings clarify the evolving roles of multi-level actors in China’s international communication and offer new insights into the operation of news convergence in non-Western contexts.

# International Communication

# Text-Mining

# News Convergence

# Filed Theory

# Communication Effects

* Journal of Journalism and Communication Studies (《新闻与传播研究》) is the top-tier Chinese journal in journalism and communication studies.
** ​ First student author.

How to Relieve Worries? A Study of Public Narratives on Sleep Health Based on Social Media Big Data. Education Media Research.(06), 97-104. doi:10.19400/j.cnki.cn10-1407/g2.2023.06.015.

Jie Li, Junheng Qian* & Rui Qing. (2023). 

This study investigates how sleep health is discussed on social media and how the public perceives it, aiming to inform more effective health communication strategies. Using Python, we scraped 43,486 sleep-related posts from Weibo and conducted sentiment analysis and semantic network analysis to examine emotional tone and thematic patterns. Results show that negative sentiment dominates (62.99%), far exceeding neutral (21.98%) and positive (15.02%) expressions. Negative posts focus on poor sleep experiences and related struggles, while neutral and positive posts explore coping methods and improvements, especially highlighting the role of exercise. Our findings suggest that future sleep health communication should leverage social media more actively, promote accurate public understanding, and respond to users’ needs with targeted, science-based information.

# Sleep Health

# Narrative

# Social media

# Network Analysis

# Sentiment Analysis

* ​ First student author.


B. Peer-reviewed Conference Papers

Junheng Qian & Zhijin Zhong. (2025).Diachronic text mining analysis of news convergence and its impact on audience news engagement across domestic and global contexts. Paper presented at the Association for Education in Journalism and Mass Communication (AEJMC) 2025 Conference, San Francisco, United States.

Dongyang Zhang, Junheng Qian & Bo Wang. (2025). How does economic domination act on media domination? The Relationships between AI Industrial Competitive Advantages, Media Attention, and Media Tone between China and the United States. Paper presented at the International Communication Association (ICA) 2025 Conference, Denver, USA.

Junheng Qian & Jie Li. (2024). How Information Overload Impacts Preventive Behavioural Intentions in Public Health Emergencies: Examining the Mediating Roles of Emotion and Health Beliefs.Paper presented at the International Communication Association (ICA) 2024 Conference, Gold Coast, Australia.

Junheng Qian. (2024). The Urban-Rural Gap in COVID-19 Vaccination Intention in China: An Analysis Based on Data from the Chinese General Social Survey 2021. Paper presented at the Association for Education in Journalism and Mass Communication (AEJMC) 2024 Conference, Philadelphia, USA.


C. Research Experience

Research Assistant | Hussman School of Journalism and Media, the University of North Carolina at Chapel Hill

05/2025-Present

Supervisor: Prof. Xinyan (Eva) Zhao (UNC) & Prof. Wenlin Liu (University of Florida)

Proposed and refined research design in collaboration with advisors.
Conducted data cleaning, mining, and statistical analysis using Python.
Participated in weekly group meetings and presented research progress.
Authored a first-author manuscript under review at a peer-reviewed journal.

Research Assistant | Centre for Computational Communication Research, Sun Yat-Sen University

09/2023-Present

Supervisor: Prof. Zhi-jin Zhong

Led and co-led multiple research projects from design to publication.
Authored one peer-reviewed publication, one revise-and-resubmit manuscript, and several manuscripts under review and conference papers.
Drafted policy reports for government and contributed to grant proposals.
Managed and curated lab-wide datasets.

Research Assistant | Faculty of Humanities, Hong Kong Polytechnic University

05/2024-11/2024

Supervisor: Prof. Cindy Ngai

Conducted comprehensive data analysis for two research projects, using structural equation modelling (SEM) and other statistical methods.
Generated clear and detailed data analysis for projects.
Collaborated with advisors to interpret results and contribute to manuscript.

Undergraduate Research Assistant | School of Public Administration, University of Electronic Science and Technology of China

09/2021-06/2023

Supervisor: Prof. Jie Li and Prof. Xiao Han

Collected and organized data for a government research project.
Assisted in faculty-led research projects, contributing to study design and analysis.
Co-authored a publication in a peer-reviewed journal.