How to come up with a research idea

7 min readOct 4, 2017

Here are the steps:

Identify a topic on which you want to work. Set it up as a question. For example, imagine you want to work about migration, acculturation and mental health. A possible question might be, “How does migration influence the association between acculturation and mental health?” Or you may ask the same question in a different way, “How does the relationship between migration and the mental health of the migrants are influenced by acculturation?” This is the first step where you will need to ask yourself a lot of questions; generate as many questions as you can.

In the second step you will need to start drafting a paper to give these questions the life of their own. There are three parts to it — start with a set of facts, develop a theory from the facts, and from the theory generate hypotheses that you’d like to test. This is usually what we do in Health Sciences. You may also start working with the facts to begin with and then develop a theory or let a theory emerge that you may want to test in other ways - for example using qualitative inquiry.

In the third step, you start drafting the facts. When you write these “facts”, that is, information bits that will inform your theory building, you should work from nouns and verbs. This is the craft of writing your position. Avoid adjectives as much as possible unless necessary. Do not use adverbs. This is a self-imposed constraint that we have seen work with us.

How would you write it? There is no fixed style and over time, you will see that you will develop your own style of work. You can either work on a chronological basis, that is, start with historic origins of a concept and work forwards in time. For example, let’s say you are investigating the role of migration and health. Your thesis or argument is to test healthy migrant effect as a hypothesis. Start with the history of the healthy migrant effect, and progressively examine the role of healthy migrant effect till present. This sets up a timeline where you can start reviewing the flow of concepts. Another idea might be to use concept maps.

Start with Facts

In the beginning are facts.

In the middle are facts.

In the end are facts.

Everything else is like sandwich fillings.

Only ideas and questions matter.

Vacuous opinions do not matter. If you are inclined to express opinioins, keep them to yourself, jot down in a journal, and methodically cull them. Challenge yourself at every step.

At all stages, stick only to the facts; then theories that you derive from facts taken together and linking all the facts; hypotheses (kind of kernels of facts) that you will test at some point. Do this repeatedly. When you read a paper, discard everything but facts.

When you note facts, note facts. Not your own mind’s reactions’ products. Learn to recognise your own mind.

To do this, read slowly.

To read and think slowly, learn to breathe slowly.

Deliberately slow down. Walk slowly. Think slowly, pause often, take time to drink water, and think.

Satyajit Ray is reputed to have said that when he met Rabindranath Tagore, he found Tagore spoke every word as if they were scripted, every word measured, every word with its nuance¹

Cultivate Mindfulness.

But how?

Niklas Goeke wrote an essay on mental biases (here is the link). Read it.
Check out Maria Popova’s book on Sherlock Holmesian thinking, how to think like Sherlock Holmes (link here)
Daniel Kahnemann and Amos Tversky wrote “Thinking Fast and Slow” discussing the mechanics of the different ways of slowing down your thinking and engaging the faster, intuitive, reflex to a more carefully considered slower controlled mode of thinking. Read it.
Maria Popova proposed a five-step process to think new ideas and creativity (link here)

Bare bones facts. Statistics, data. For health related topics (our emphasis here), write epidemiology:

What is the health topic?
Who are affected?
Where are these people?
Why should we be worried about it?
What causes or leads to the health issue you are going to work with?

Ask lots and lots of questions

Ask questions. Link topics with questions.

Simon Sinek in his “Start with Why” wants us to ask three fundamental questions on any project we start or work with:

Start with “Why”. Why are you interested in the topic? Keep it in the centre
Then move out and answer the question “How”, i.e., how will you do it
Finally, wrap around let the answer as to “What”, or “what is it that you are going to work on” emerge.

I have seen about eight out of 10 of my students start the other way round. When they start writing their proposals, they start with a “what”. That is, they already have an idea and they are sold to it. Rather than asking yourself why would you carry out this research, or why are you interested in pursuing the PhD, or that Master degree, if you are sold to an idea, you may find yourself confused as you start writing your ideas on the proposal.

Concept maps?

Yes concept maps (Figure 1)

Concept maps are about concepts. You have two different types of concepts: abstract concepts and concrete specific types of concepts. Start with the most abstract concept you want to work with (it does not matter, you can start with highly specific concepts as well). By abstract, I mean un-measurable, entities that you can “conceive” but kind of all encompassing, organising concepts. Say something like “Climate Change” as a concept. You then start making more concrete concepts like “Temperature Rise”; get even more specific “4 degrees increase in global temperature” and so on. You can use pen and paper to write or create these concepts; or you can use post-it notes. Use cubes of post it notes and tear pages off post-it notes to create concepts and then lay them on a floor or a wall. Then once you have gathered concepts then move them around.

The other thing you can do is to use digital tools. I use both cmap software and scapple. I actually love scapple as it is intuitive and works great. Up to you. Anything that will help you to draw concepts and move them around will work. Your first concept map will be wrong. So, move around the concepts till you get the right kind of alignment. Remember the three rules:

How to generate concept maps

Start with an idea parking zone. Put lists of all your ideas, concepts, thoughts in the form of single words or phrases and stack them somewhere. I usually use a white board, or floor, or wall as a holding space and post-it notes.
Start thinking. Arrange the concepts from most abstract to most concrete from top to bottom. Your concepts are going to be nouns.
Connect pairs of concepts using verbs. When two concepts are connected by a verb, then the sentence is referred to as a proposition. A proposition can be a starting point for writing a paragraph or even the theme of a paragraph. Say, you want to work on climate change and health. Your top level concept is climate change and another level is global warming. You can write something like

“Climate change” causes “Global Warming”.

This sentence is a proposition. OK, so now you have concepts and propositions. The real issue here is that you can connect concepts using arrows going in all directions and any directions you like. It is this connectivity that matters. You can see from here that if you have a list of 10 concepts in say two to three levels, you will have 45 connectors and propositions.

3. Aim to have about 15–25 concepts, not more than 25 concepts and not less than 15 concepts per idea. That way you will have between 100 and 300 propositions that you can use for writing your notes.

There is a flow or order that you must follow when you start crafting your proposal. Start with what we know or the facts as stated. All facts must be accompanied by the source information. Next, tie a theory that aims to identify if there are conflicting information that must be resolved, or if you construct a theory tying together all the available facts, what can you say about the problem that you decide to address. From here, start generating one or more hypotheses and work from there. So in section one, write about everything we know about topic T, in section 2, write about the questions that arise based on what we have known and accumulated in section 1, and then in section three, offer your own solution. This solution can be of two different types (in health sciences):

X causes Y ( for example, smoking causes lung cancer). If X causes Y, you will need to consider three fulfilling criteria at least where X and Y will have a valid association:
That the occurrence of X and Y could not be because of chance. This is rarely an issue for criticism as most published studies are well powered; however, when you start out unless you propose to investigate such an association, you will need to be mindful.
That, there may be some form of bias that has led to an apparent association between X and Y, in reality, X and Y may not be associated at all. This is a valid argument and you will need to investigate this angle of criticism for all studies that you evaluate in health sciences. If you deal with observational studies, learn to identify and list possible selection and information bias or response biases and try to uncover if these might explain the association or apparent association we get to see.
That, a third variable, Z, might actually explain the association between X and Y, such that, Z is associated both with X and Y and does not come in the path between X and Y. Z is referred to as a confounding variable and test if the authors have covered confounding variables sufficiently enough to justify whether another study may shed insight into it.

So, there are different ways to go about looking into the apparent associations or even lack of associations and construct theories that might fit into a position.

See Krishna Dutta and Andrew Robinson’s Tagore: Myriad Minded Man.