Lesson Specification: Module 2.1

Metadata

Field	Value
Module	Module 2.1
Title	Designing Your Research Question
Part	Part 1: Main Curriculum
Estimated Duration	30 minutes
Nexus Tool(s)	None (pure methodology)
Prerequisites	Module 0.2 (Research Methodologies)

By the end of this lesson, students will be able to:

Construct a well-scoped research question using established frameworks (PICO for health sciences, SPIDER for qualitative research, or the Heilmeier Catechism for engineering/CS).
Distinguish between a topic, a research gap, and a research question — and explain why only the latter drives a thesis.
Perform a preliminary gap analysis by combining keyword searches with citation network analysis to identify where the field has under-explored territory.
Evaluate their own research question against the "FINER" criteria (Feasible, Interesting, Novel, Ethical, Relevant).

A Topic is a broad area: "Machine learning in healthcare."
A Gap is a specific observation about what the field has NOT done: "No study has compared transformer architectures to RNNs for predicting ICU readmission in pediatric patients."
A Research Question is a precise, answerable query that fills that gap: "How does transformer-based prediction of pediatric ICU readmission compare to RNN-based models in terms of AUROC and calibration?"
The Mistake: Most first-year PhD students confuse a topic for a question, leading to years of unfocused work.

PICO (Population, Intervention, Comparison, Outcome): Gold standard for clinical and health sciences research. Forces you to specify exactly who, what, vs. what, and measured how.
SPIDER (Sample, Phenomenon of Interest, Design, Evaluation, Research type): Adapted for qualitative and mixed-methods research where "intervention" doesn't apply.
Heilmeier Catechism (DARPA): For engineering and computer science. Asks: What are you trying to do? How is it done today? What is new in your approach? Who cares? What are the risks?

Step 1: Conduct a broad search using scholar-search-kit.
Step 2: Build a citation network using scholar-graph-kit. Look for clusters with few connections (under-explored intersections).
Step 3: Read the "Future Work" sections of the most-cited papers (identified by PageRank). These sections are literal roadmaps to publishable research questions.

Feasible: Can I realistically answer this with my resources, time, and access to data?
Interesting: Will my committee, my field, and potential employers care about the answer?
Novel: Has this exact question been answered already? (Check your dedup results!)
Ethical: Can I pursue this without harming participants or violating IRB/ethics board policies?
Relevant: Does the answer advance knowledge or practice in a meaningful way?

Pain Point: "Six months into your PhD, your advisor asks: 'So, what exactly is your research question?' You pause. You thought you had one. You say, 'I'm working on machine learning in healthcare.' Your advisor frowns. That's a topic, not a question. You just realized you've been working without a compass."
Wow Moment: "What if you could use the same tools we learned in Module 3 and Module 4 — search and graph analysis — to systematically identify gaps in the literature and construct a research question that is almost guaranteed to be publishable?"
Real-World Scenario: "Your thesis proposal is due in 3 months. By the end of this lesson, you will have a draft research question that passes the FINER criteria, built using a repeatable, evidence-based process."

Builds on: Module 0.2 (Research Methodologies — quantitative, qualitative, mixed).
Uses tools from: Module 2.2 (scholar-search-kit) and Module 3.2 (scholar-graph-kit — citation networks).
Leads to: Module 3.4 (Conducting a full SLR).

Hulley, S. B., et al. (2013). Designing Clinical Research. (FINER criteria)
Cooke, A., Smith, D., & Booth, A. (2012). Beyond PICO: the SPIDER tool for qualitative evidence synthesis. Qualitative Health Research.
Heilmeier, G. The Heilmeier Catechism. DARPA.