DISCERN: Designing Decision Support Interfaces to Investigate the Complexities of Workplace Social Decision-Making With Line Managers

Title of the Paper

DISCERN: Designing Decision Support Interfaces to Investigate the Complexities of Workplace Social Decision-Making With Line Managers

Paper Information

• Research Area: Human-Computer Interaction (HCI), workplace social decision support system design
• Keywords: organizational decision-making, collaboration and social computing, technology probes, user engagement, decision visualization, interaction design, workplace analysis

Research Background and Issues

• Identified Issues
  • Decision-making needs of lower-level managers (e.g., line managers) are significantly underestimated, despite them comprising 60% of the management workforce and supervising approximately 80% of employees.
  • Existing decision support tools focus primarily on isolated decision scenarios with greater social distance, lacking support for social decision-making scenarios requiring close interaction.
  • Current decision support tools are predominantly based on numerical analysis and predictive models, which are less effective in intimate, highly interactive decision-making contexts.
• Importance of the Issues:
  • As workplaces become increasingly complex and data-driven, lower-level managers require decision-making capabilities that go beyond simple data analysis, balancing decisions with maintaining relationships among team members.
• Research Motivation and Related Work
  • In HCI research, digital systems for social decision-making often rely on three assumptions: stakeholder-centric data collection, numerical representation of values, and outcome prediction systems. However, these assumptions are poorly suited to highly interactive decision-making contexts.
  • This study examines the applicability of these assumptions in middle-management decision-making and develops tools to address the shortcomings of existing tools in supporting interactive decision-making.

Solution

• Proposed Approach
  • Developed a technology probe called DISCERN—an extended tool based on Microsoft Excel.
  • Supports externalization of multi-level decision processes, using a tree structure to represent decision goals and optimization priorities, combined with tabular analysis tools to store and process raw data.
  • DISCERN addresses current tool limitations by offering fuzzy characteristic support, multi-level interaction, and a "qualculation" feature that integrates qualitative and quantitative approaches.
• Innovations
  • Shifts traditional decision support from informatization to interactivity, transforming dynamic stakeholder perspectives into decision representations that are easier to collaborate on and interpret.
  • Promotes transparency and discussability in decision-making processes by using discrete weights, tree structures, and imprecise quantification tools that preserve ambiguity and flexibility.
• Implementation Steps
  1. Needs Exploration:
     • Conducted surveys and semi-structured interviews to deeply understand the decision-making behaviors and needs of lower-level managers.
     • Defined the tool development direction: supporting externalization rather than replacing face-to-face interactions.
  2. System Development:
     • Based on preliminary research, developed DISCERN using tree structures and interactive visualization techniques.
     • Integrated the GPT-3.5 model to provide attribute suggestions and feedback for decision-making.
  3. User Validation:
     • Observed managers’ usage behaviors in simulated scenarios through user enactment methods.
     • Compared DISCERN with Excel to analyze the tool’s support for high-level goal reasoning and interaction.

Research Outcomes

• Specific Results
  • DISCERN provides more flexible and realistic support for organizational decision-making, particularly in complex or highly interactive scenarios.
  • Managers can more easily define decision hierarchies, express object ambiguity, and balance quantitative and intuitive decision-making using DISCERN.
• Advantages Over Existing Solutions
  • Enhances transparency in multi-level thinking: TREE structures express high-level decision goals, while links to spreadsheets maintain the precision of raw data.
  • Provides support for ambiguity: Discrete weights and non-numerical expressions reduce the likelihood of disrupting consensus.
  • Adapts to iterative decision-making: Supports continuous logic from informal initial stages to formal analytical phases.
• Experimental or Evaluation Results
  • User enactment demonstrated that DISCERN is more effective than traditional tools (e.g., Excel) in expressing high-level goals and incorporating stakeholder perspectives.
  • However, DISCERN was found to have limitations in supporting early brainstorming and problem exploration flexibility.
• Limitations and Future Directions
  • Limitations:
    • The study primarily focused on managers in tech companies, with insufficient exploration of management needs in other sectors such as service industries.
    • Limited exploration of real-world usage scenarios through hypothetical user enactments.
  • Future Directions:
    • Develop tools to support early unstructured stages such as brainstorming.
    • Design more flexible and adaptive decision-making tools to support real-time improvements and experimental decision processes.
    • Explore stakeholder co-creation in tool design to ensure diverse perspectives are reflected in tool development.

Conclusion

This paper systematically reveals the complexities of workplace social decision-making and the limitations of existing tools. Through the DISCERN technology probe, it demonstrates the future direction and design insights for digital social decision support tools, providing an important reference pathway for future research and tool development across multiple domains.