Who Cares, and What Has Already Been Done (Judy's Questions 2–3)

Chapter 2 Episode 3

2. Who cares? Why should people care about this problem?

3. What have other people done about it? Why is that not sufficient?

Q2 and Q3 together assemble the foundation of your work. Q2 asks you to develop a comprehensive understanding of who is involved in the context you are studying, since an incomplete stakeholder picture will bias every decision downstream. Q3 asks what prior work has already been conducted to address the problem and where specifically it fell short. Together, they ensure you understand both the full landscape of voices to consider and the specific gaps that remain for your work to address.

Q2: Who Cares?

This question looks much simpler than it actually is. It is easy to find a few obvious stakeholders, but identifying only partial stakeholders will guarantee that you overlook critical factors and considerations that affect the validity of everything downstream, from the question you ask to the study you design to the findings you report. My interpretation of Q2 is to answer the context fully: who is involved, and whose voices must be considered to avoid building on an incomplete picture.

As an HCI researcher, what I mean by understanding the context goes beyond simply identifying the direct user of a technology. It requires understanding the socio-technical systems that depict the organizational model of the interdependent relationship between humans and technology. The people who use a tool, the people who receive information from it, the people who make decisions based on that information, and the people who are affected by those decisions may all be distinct stakeholders with different needs, constraints, and behaviors. If your stakeholder analysis captures only some of them, your research question will inherit the blind spots, your study design will miss critical variables, and your findings will carry biases that reviewers and practitioners will notice.

Let me give you a concrete example. Imagine you are designing a remote patient monitoring system using an AI-based conversational assistant, and you argue that the system should facilitate clinicians’ needs by designing the conversation flow to collect information that supports their decision-making. Fair enough, right? But if you stop at the clinician as the primary stakeholder and conduct design studies only with them, you carry a latent assumption that patients will use the system the way clinicians expect and are willing to do so. What if patients resist using the device? What if they lack the literacy to operate it properly, or cannot articulate their symptoms clearly? And between clinicians and patients, nurses are often the moderators who receive patient data firsthand and are accountable for triaging — how nurses make sense of that data is also critical to whether clinicians will be able to make an informed decision. Any of these factors being overlooked will undermine the validity and effectiveness of the system, and overlooking the perspectives of diverse stakeholders in both the design and the subsequent evaluation is a critical flaw that could have been avoided by asking Q2 more thoroughly at the start.

Gaining a comprehensive and in-depth understanding of the target context is what Q2 ultimately asks you to do, so that the question you ask, the study you design, and the findings you report do not carry biases from an incomplete picture of the landscape. Effective methodologies include a thorough literature review as well as talking with real-world stakeholders when possible. Using an AI tool to pressure-test your stakeholder list can also help surface blind spots.

Q3: What Have Other People Done About It?

Your literature survey pays off here as well, since the answer builds on Q1 and Q2 by reviewing prior work that attempted to solve problems within the same context, sometimes for a broader or narrower scope of stakeholders. You should use published papers as your evidence, and personally I find those published at top-tier venues in your target domain more credible for grounding the argument. For each piece of prior work you cite, articulate what they did, what approach they took, and what specific limitation or gap remains.

The Gap: Why Is That Not Sufficient?

Summarizing what others have done is, to borrow from NLP terminology, an extractive task — a relatively straightforward one that, as long as you have done the literature review properly, anyone can execute well, because it does not require you to generate new knowledge. The second half of Q3, identifying what others have missed, what they left unaddressed, or what assumptions they made that your work will challenge, is a much harder abstractive task. By abstractive, I mean that the task requires generating new knowledge by reading and understanding the original content. I have seen two failure patterns repeatedly, and both are worth naming.

First, many papers make the claim that “no one has studied X.” This is a common writing issue with using absolute language in academic work, and I discuss it in more detail in the earlier post on the craft of prose. When you use such absolute claims, you can never be certain that no one has attempted something similar in a venue or context beyond your search, and a reviewer who finds even one counterexample can dismiss the entire argument. A claim of novelty based on absence is often considered weak because, as the principle goes, the absence of evidence is not evidence of absence.

The second failure is listing prior work and then stating that “there is still room for improvement” or “further research is needed” without justifying what exactly is missing and why it matters. These claims are your own hypotheses without supporting evidence, which is neither persuasive nor credible. What is missing is a clear argument, grounded in what you know about the domain, for what directions are worth investigating, based on which hypotheses, and what evidence supports those hypotheses.

A strong gap statement could look like this: “existing approaches handle X but fail under condition Y because they assume or overlook Z.” The strength of this formulation is that it identifies a specific assumption or limitation in prior work and connects it to a concrete condition under which that limitation matters, which gives you a logically coherent foundation that is far harder for a reviewer to dismiss. Your approach in Q4 should then directly address the assumption or condition you just identified, so the gap and the approach form an unbroken chain.

You may be wondering: there are a thousand ways to form such a statement with logically coherent rationales, so which one should you choose, and why? That is exactly the right question, and the next episode gives you the answer.

Scoping Through Q2 and Q3

Together, Q2 and Q3 define the scope of your project from two directions. Q2 tells you who you are serving and what is at stake for them. Q3 tells you what has already been tried and where specifically it fell short. The intersection of these two answers is where your contribution lives: a specific gap in the prior work, affecting specific stakeholders, with specific consequences that justify the effort of a new investigation. If your Q2 is vague about the stakeholders or your Q3 is vague about the gap, the scope of your project is undefined, and everything downstream — your approach in Q4, your methods in Q5, and your contribution claim when you write the paper — will inherit that vagueness.

The next post covers Q4 and Q5, where you turn the gap you identified here into a concrete approach and a specific methodology, using the six-question chain introduced in the overview post as the detailed construction method for building the motivation.