Nature Human Behaviour launched a series of 'How to' Comments in 2025, offering expert guidance to researchers and readers of the journal. This is the first of a series of blogs summarizing the key findings of these Comments for the Communities’ Research Publishing Toolkit.  

What makes a ‘good’ research question? 

You may be wondering what characteristics determine a good research question. Is it precise? Does it challenge previous assumptions?  

Good research questions: 

• Drive and capture enthusiasm  
• Reflect innovative thinking 
• Reframe unanswered questions 
• Challenge previous assumptions 
• Demonstrate how past methods may have limited understanding 

The process of developing a research question 

Megan A. K. Peters outlines a multiphase approach to developing a good research question, including 4 main phases: 

Phase 1 

In the beginning, incite your curiosity. Let your mind take you to weird and wonderful places that inspire you – perhaps starting with something you’ve read or heard. 

Brainstorm ideas. You can write down your thoughts, allowing your ideas to form and mature. The more the merrier! Walking by yourself and vocalising these thoughts (out loud) may help with the creative process. Spoken words can be powerful when forming ideas and connecting the dots. 

When you’re brainstormed out, critique your ideas. Look at your arguments and explore their flaws. How might others critique you? Talk with people who understand your field, but are not specialists, to help you see things from another perspective. 

Phase 2 

What do you know about the existing body of knowledge? Time to dive in.  

• What questions have been explored that are like yours?  
• What did they find and how did they do it?  
• Read the literature and take notes, linking papers and ideas. 
• Develop your own shorthand for notetaking. 
• Use automated tools sparingly to avoid losing connections in your mind-map. 
• Consider using free tools like Obsidian to link ideas and visualise connections. 
• Tag papers using ‘primary’ or ‘secondary’ to your likely question. 
• Read widely – do not narrow your reading list (yet!). 

Phase 3 

Next, refine. Aim to create a specific set of circumstances through which your question can be addressed. 

Return to the papers that you marked as ‘primary’ to your likely question and go over them in detail. Identify the ones that fall close to your vision. Read them again and ask what they might have done differently, instead of what they actually did. 

Generate a list of alternative explanations or experimental designs that might challenge existing findings. This could be subtle, or more obvious. It could involve a different analysis, data collection method, target population or task.  

Phase 4 

The final product is almost here!  

Before you finalise your question, consider potential confounds or alternative explanations. 

Specify the paradigms, tools, and approaches you will use to answer your question, keeping in mind other plausible explanations or what others might find. By considering the potential confounds or other interpretations, you can develop a methodology that rules them out. 

You should now be left with a clear, focused and tractable (i.e. a good) research question.  

To learn more and discover things to avoid, you can read the full ‘How to’ Comment, ‘How to develop good research questions’, by Megan A. K. Peters, here. You will find the Research Publishing toolkit here, alongside other frequently asked questions about the publishing process.