CoArgue: Fostering Lurkers’ Contribution to Collective Arguments in Community-based QA Platforms

Title of the Paper

CoArgue: Fostering Lurkers’ Contribution to Collective Arguments in Community-based QA Platforms

Paper Information

• Domain: Human-Computer Interaction (HCI), Community Question Answering (CQA), Natural Language Processing (NLP)
• Keywords: Collective arguments, community-based QA platforms, Q&A systems, NLP, user engagement, interaction design, information visualization, lurker, argumentation writing, community participation

Research Background and Problem Statement

• Identified Problems or Challenges:

  • In community-based QA (CQA) platforms, the development of collective arguments relies on a sufficient base of contributors. However, most frequent users are "lurkers" who rarely post content.
  • Lurkers face challenges such as lack of confidence, low willingness to participate (e.g., perceiving their contributions as unimportant or insignificant), and difficulties in processing complex information, such as effectively digesting and producing meaningful responses.

• Significance:

  • The sustainable development of collective arguments requires increasing lurkers’ participation; otherwise, the growth and sustainability of CQA communities are at risk.
  • Researching how to lower the participation barriers for lurkers not only enhances community contributions but also provides design insights for other online communities.

• Research Motivation and Related Work:

  • Existing research supporting lurkers often focuses on specific communities (e.g., health or collaborative communities) but lacks designs tailored for loosely connected CQA communities.
  • Current tools aimed at improving argumentation skills primarily focus on generating standalone texts rather than collaboratively constructing collective arguments. Therefore, there is an urgent need to design tools that effectively support users in contributing to collective arguments.

Proposed Solution

• Proposed Solution:

  • CoArgue System: A visualization system based on natural language processing and interaction design, aimed at helping lurkers participate more effectively in collective arguments on CQA platforms through motivation and capability support.

• Innovations:

  1. Proposing an NLP pipeline to extract and organize argumentative elements in Q&A discussions, such as stance, claims, and premises.
  2. Combining various interaction designs (e.g., navigation views, chatbots, and contribution views) with information visualization to provide personalized support, addressing users’ barriers in confidence, willingness, and capability.
  3. Integrating encouragement mechanisms, content structuring, and writing guidance tailored to the needs of lurkers.

• Implementation Steps and Key Technologies:

  1. NLP Pipeline: Extracting key argumentative elements, including stance analysis (categorized as positive, neutral, or negative), claim clustering (using BERT embeddings and clustering algorithms), and extracting associated premise keywords.
  2. User Interface Design:
     • Answer View: Highlights claims and premises in content to help users quickly grasp key information.
     • Navigation View: Provides an overview of collective arguments, allowing users to view categorized claim clusters and their popularity.
     • Chatbot View: Assists users in organizing their thoughts and provides motivational support through conversational interactions.
     • Writing Interface: Pre-fills selected claims and keywords chosen by the user to reduce the difficulty of writing.
     • Contribution View: Offers real-time feedback on users’ contribution metrics to the collective arguments.

Research Outcomes

• Specific Outcomes:

  1. Compared to baseline systems like Quora, CoArgue significantly improved lurkers’ motivation, writing skills, and engagement during interactions.
  2. The content generated by users increased in both quantity and richness (especially the number of premises), and users rated their output quality and participation confidence higher.

• Advantages:

  • CoArgue demonstrated superior performance in supporting lurkers’ contributions to collective arguments compared to baseline systems, particularly by effectively enhancing users’ willingness to participate and sustaining their contributions.
  • Visualization and interaction design further reduced users’ cognitive load and provided more pathways for lurkers to engage.

• Experimental or Evaluation Results:

  • User experiments showed that CoArgue’s motivational support (users feeling more recognized) and capability support (improved understanding and generation of content) were significantly better than the baseline.
  • Users generated significantly more claims and premises compared to those using the baseline system.
  • While the overall complexity of CoArgue was higher, leading some users to perceive a learning curve, most users expressed willingness to use the system again.

• Limitations and Future Directions:

  1. The experiment participants were predominantly young users, necessitating further research on other age groups and demographics.
  2. The current study focused mainly on technology-related discussion topics; future research should expand to include more controversial or high-knowledge-threshold topics.
  3. The experiment only evaluated short-term effects; long-term studies are needed to verify whether CoArgue can sustainably reduce lurking behavior and cultivate more active users.
  4. Some users found the chatbot functionality insufficiently intelligent or repetitive; future improvements could enhance the conversational experience by upgrading the language model’s capabilities.