Discussion: View Thread

Expand all | Collapse all

A study (and tool) on what makes a research question worth asking - would value your reactions

  • 1.  A study (and tool) on what makes a research question worth asking - would value your reactions

    Posted 5 hours ago

    Dear colleagues,

    A study is only as strong as its research question. Get it wrong, and no amount of careful downstream work can rescue the project. Yet almost all of our methodological tools focus on the data (design, measurement, analysis); very few help with the question itself. It's also the most common feedback I give on papers. So I built a coaching tool to help close that gap, and I'm now studying whether it works.

    To take part: https://rq-coach.vercel.app/?c=aom-ob-2026-06

    It's AI-supported but not a chatbot you ask to fix your question: the coaching logic is grounded in published work on research-question quality (e.g., Davis, 1971; Sandberg & Alvesson, 2011; Hollenbeck, 2008; Dorobantu et al., 2024). The language model is just the engine. In ~20–30 minutes, it works through your puzzle, tension, and contribution, checks the literature against the field's leading journals (with verifiable citations), and pushes back rather than writing the question for you. I have carefully pilot-tested it before, and the LLM model is funded through my research budget. You can use the tool for free, whether for 1 or 3 hours.

    I'd value two things from this community:

    1. Participation. The tool asks you to upload a proposal you're actively developing and work through one session, plus a short voluntary survey. Sessions are anonymous, stored on EU servers, and used only for this research; consent is on the first screen, and your ideas are never shared outside the research team or published. 
    2. Your reactions. For those of you who advise students on research questions: does the way the tool diagnoses a weak question match how you'd do it yourself,  and where does it fall short? That's exactly the feedback that will shape the next version, and I'd genuinely welcome it, here in the thread or directly.

    Once it reaches a solid proof of concept, I'll release the architecture and the code, fully open access.



    ------------------------------
    Christian Troester
    Professor of Leadership and Organizational Behavior
    KLU - Kühne Logistics University
    Hamburg, Germany
    ------------------------------