Research Question

Your research question (RQ) is the anchor of your thesis: it determines what evidence you need, what methods are appropriate, and what a defensible conclusion looks like.

If you are unsure, that’s normal. Write 2–3 candidate RQs, then pick the one you can actually answer with your time and access.

What makes a strong research question?

A strong RQ is:

  • Answerable with the data, access, and time you have
  • Specific about the context, intervention/approach, and outcome
  • Researchable (it produces a structured set of possible answers, not a yes/no opinion)
  • Aligned with your methodology and evaluation design

Search vs. research (a useful distinction)

“Search” questions are often yes/no or purely descriptive. “Research” questions ask about mechanisms, conditions, trade-offs, or comparisons.

  • Search: Can real-time tracking reduce delivery delays?

  • Research: Through which mechanisms does real-time tracking reduce delivery delays (e.g., better ETA predictions, earlier exception handling, improved dispatch decisions), and under what conditions?

  • Search: How much did urban freight volume increase during the pandemic?

  • Research: Which factors explain the increase (e-commerce growth, delivery frequency, consumer behavior), and how large is each factor’s contribution?

Narrow vs. broad (finding the “right size”)

  • Too narrow: How did AGVs reduce picking time in Zone B during peak week 42 of 2023?
  • Too broad: How does automation improve logistics?
  • About right: How do AGVs affect order-picking productivity and error rates in a multi-zone warehouse, and what are the main operational constraints?

Common traps

  1. Unanswerable: “How can logistics achieve perfection under all conditions?”
  2. Opinion-based: “Which warehouse layout is best?”
  3. Vague “how can”: “How can companies reduce costs?” (reduce which costs, in what setting, using what levers?)

Checklist (use before you commit)

  • Clarity: Could a reader restate your RQ in one sentence?
  • Feasibility: Can you answer it within the project duration and data access?
  • Operationalization: Do you know what you will measure (metrics, variables, outcomes)?
  • Evaluation: Do you know what comparison/baseline you will use?
  • Scope: Is the scope defensible (and not silently hiding major limitations)?

A helpful framework: PICO (quantitative/empirical work)

  • Population: what system/context are you studying (e.g., multi-zone warehouse, urban delivery area)?
  • Intervention: what changes (e.g., algorithm, policy, new technology)?
  • Comparison: what is the baseline (current practice, alternative method, before/after)?
  • Outcome: what outcomes matter (cost, throughput, service level, emissions, robustness)?

Examples (and how to improve them)

Example 1

Weak: “What is the optimal storage location assignment and picking policy to minimize costs in the XYZ warehouse?”

Improve by:

  • Avoiding “optimal” unless you can prove optimality (often unrealistic in theses with heuristics).
  • Defining which costs and which constraints.
  • Making the comparison explicit (against what baseline?).

Stronger: “Which adaptive storage location assignment and picking policy combinations reduce picker travel distance and labor cost, under space and labor constraints in the XYZ warehouse, compared to current practice?”

Example 2

Weak: “How can warehouses reduce costs with automation?”

Stronger: “Which automation technologies reduce labor cost in multi-zone order picking, and what trade-offs do they introduce for service level and flexibility?”

Example 3

Weak: “How does a hybrid slotting model perform compared to existing strategies in minimizing picker travel and associated costs?”

Stronger: “How does a hybrid slotting model (re-slotting + healing) compare to baseline slotting strategies in reducing picker travel distance and labor cost, and how sensitive is performance to demand volatility?”

Using AI tools responsibly (optional)

AI tools can help you refine wording, surface ambiguity, and generate alternatives. Use them as critique tools, not as decision-makers. Good prompt templates:

  • “Critique this RQ for clarity, feasibility, and measurability: …”
  • “Rewrite this RQ into three variants: (1) empirical, (2) optimization/design, (3) qualitative case study.”
  • “List the implied variables/metrics and the minimal data I would need to answer this RQ: …”

Reference