DeltaMV

Designing a Good Research Question (Projects Often Fail Here) 

By DeltaMV Insights TeamReviewed by DeltaMV Academy EditorialUpdated 7 min read

Module 2 · 3 of 7Market Research Methodology & Behavioural Science FrameworksAll articles
  1. What Is Behavioural Science, and Why Does It Matter in Healthcare? 
  2. Describe, Understand, Shape: DeltaMV’s Framework for Turning Data Into Decisions 
  3. Designing a Good Research Question (Projects Often Fail Here) 
  4. Sampling 101: Who You Talk to Determines What You Learn 
  5. Make It Emotional, Make It Social, Make It Easy: DeltaMV’s Approach to Behaviour Change 
  6. Health Belief Model & Theory of Planned Behaviour, Explained Simply 
  7. Nudge Theory in Healthcare: Small Changes, Big Impact 

A good research question turns a business decision into something research can actually answer. Most disappointing studies fail here, not in fieldwork. This article walks through six steps to design a clear, answerable question, the mistakes that spoil it, and a checklist you can use before any project starts. 

Key Takeaways

  • A research question is not a topic. “GP attitudes to our new inhaler” is a topic. A question has a clear thing to find out, about a clear group of people. 
  • Start from the decision the research will inform. If no answer could change what you do, the question is not worth asking. 
  • One main question with two or three sub-questions is usually enough. A long list of questions is a sign the thinking isn’t finished. 
  • The questions you ask respondents come last. They are the output of this work, not the starting point. 
  • This article covers designing the question only. Choosing who to speak to is covered in the next article.

The Process, End to End 

Most research projects start with a topic or a worry, such as “Why aren’t doctors prescribing more?” That is a fair place to begin, but it can’t be researched yet. It has to be sharpened into a question that points to a method, a group of people, and a possible action. 

The project brief states the business decision. The research question sits one step below it. It says exactly what the study needs to find out so that the decision can be made. 

It helps to see the full ladder. The example below is illustrative. 

Level Example Its job 
Topic GP attitudes to a new inhaler Names the area. Nothing more. 
Business decision Do we fund GP education or a patient support service next year? States what will be decided. 
Research question Among GPs who tried the inhaler, why do most stop prescribing it, and how common is each reason? States what we need to know. 
Sub-questions How many tried and stopped? What reasons do they give? Which of those could we realistically address? Splits the question into answerable parts. 
Respondent questions “Tell me about the last patient you moved off it.” Comes last. Written once the levels above are clear. 

Why This Step Gets Skipped 

Writing the question takes only an hour or two, yet it is often rushed. There is usually a launch date, a budget to spend, and a feeling that everyone already agrees on what they want. 

They often don’t. Put three people from the same team in a room and ask each to write the question on a sticky note. You will usually get three different questions. Finding that out on day one is cheap. Finding it out at the final presentation is not. 

Step by Step 

  1. Start with the decision, not the topic. Ask: “What will we do differently depending on what we learn?” If the team can’t name a decision, stop here. Interesting is not the same as useful. 
  1. Name who and what. Write down exactly whom the question is about (for example, GPs who have tried the product) and what you want to know about them (for example, why they stopped). Vague groups lead to vague answers. 
  1. Write one main question. Keep it to a single sentence with one idea in it. If you need the word “and” to join two different ideas, you probably have two questions. 
  1. Add two or three sub-questions. These break the main question into pieces. Some will be about “how many” and some about “why”. That difference matters later, because it points towards the right method. See Qualitative vs Quantitative Research in Module 1. 
  1. Run the “so what” test. List the likely answers and what you would do with each one. If every answer leads to the same action, the question won’t help, so rewrite it. See the table below. 
  1. Check it can be answered. Ask whether the people you need can be reached, and whether the time and budget are realistic. A brilliant question that can’t be answered in six weeks needs to be narrowed. 

Here is the “so what” test for the inhaler example. 

If the answer is… Then we would… 
Most GPs stop because their patients struggle with the device Fund a patient training and support service 
Most GPs stop because of worries about side effects Fund GP education and evidence sharing 
The reasons are mixed, with no clear pattern Look at GP segments first, before spending 

Each answer leads somewhere different. That is what makes it a good question. 

A Quick Word on Behaviour 

Where possible, ask about what people do, and then why. “Why did you stop prescribing?” is easier to answer honestly than “How do you feel about the product?” This follows from the first article in this module: what people say and what people do can differ, and what they do is usually easier to act on. 

Common Mistakes 

  • Mistaking a topic for a question. “Understand the diabetes market” can go on forever. Without a clear end point, the study collects everything and answers nothing. 
  • Asking two questions in one. “Why do GPs stop and what do they want from us?” has two answers that may point in different directions. 
  • Building in the answer. “How can we show GPs that our device is easy?” assumes the device is easy. A good question leaves room for surprises. 
  • Choosing the method first. If someone says “let’s run a survey” before the question is written, the survey will shape the question instead of the other way round. 
  • Asking something no answer can change. If the budget is already committed, research won’t move it. Be honest about which decisions are still open. 
  • Making it too big. A question that tries to cover every market, every audience, and every barrier will produce a thin answer to all of them. 

A Checklist to Use Right Away 

  • Can you write the business decision in one sentence? 
  • Is the main research question a single sentence, with one idea in it? 
  • Does it name who the question is about? 
  • Are there no more than three sub-questions? 
  • Does each likely answer lead to a different action? 
  • Can it be answered with the time, budget, and people available? 
  • Have you written down what the study will not answer? 

If you want a template, use this: “To decide [decision], we need to know [what] among [whom], so that we can [action].” 

The Cheat Sheet

A research question is a clear, single sentence that says what you need to find out, about whom, so that a specific decision can be made. Start from the decision, keep one main question, and check that each possible answer would lead to a different action. If it doesn’t, rewrite it.

A DeltaMV Perspective

A pattern worth illustrating: a team asks for research to “find out why doctors don’t like our product.” On the surface, it sounds like a clear request. 

But when the team is asked what they would do with the answer, it turns out they are choosing between two investments: more education for doctors, or a better service for patients. “Why don’t doctors like it?” can’t separate those two options. 

The question is rewritten around the decision: among doctors who have tried the product, what is the main reason they stop, and how common is each reason? Now each answer points to an action. 

The study ends up smaller, because it speaks only to doctors who have tried the product. It is also more useful, because the team knows in advance what each result would mean. 

Composite/illustrative example; not a specific client or project.

FAQs 

What is the difference between a research objective and a research question? 

An objective says what the study will do, such as “to understand why GPs stop prescribing.” A question says what it needs to find out, such as “why do most GPs who try the product stop?” Many teams use both, but the question is the one that keeps the study focused. 

How many research questions should a project have? 

One main question and two or three sub-questions is a good rule. More than that usually means the project is trying to answer several different decisions at once. Consider splitting it into separate studies. 

Can the question change once the project has started? 

A little, yes. Early interviews often show that a question needs sharpening. Large changes, though, usually mean the original decision wasn’t clear. Agree changes openly with everyone who signed off the first version. 

Who should write the research question? 

Both sides, together. The business team knows the decision. The research partner knows what can be answered and how. A question written by only one side usually has a gap. 

The Bottom Line

A good research question is short, specific, and tied to a decision. Most projects that disappoint were not badly run. They were pointed at the wrong question. A little extra time spent here saves far more later.

What are you trying to change?

Let's explore the behavioural insight behind it.

Connect with us