How can you create research questions that are clear and testable?

Taking a pragmatic approach is a method I have found useful as it weeds out any unnecessary wording or questions that may seem relevant in the moment. Always ask yourself your own questions as if you were the subject to see if your questions best fit the purpose of your research.

A good understanding of the problem is key. You need to consider running a deep dive into the problem to make sure you fully understand all its complexities. This would help you understanding the scale of the issue at hand and its related components. Once this is well-defined, you may move on to selecting a particular aspect of the problem you wish to focus on.

The terms “scope” and “purpose” do not have the same significance. Before you formulate your research questions, you should define the subject matter of your research, along with its purpose, scope, objectives, and hypothesis. These elements serve as a foundation upon which you will elaborate your questions. However, your questions should retrieve measurable results that are used to test your hypothesis. Hypothesis are essentials, and so is the correlation between hypotheses and questions.

I think the most vital aspect of creating a research question is ensuring that it's testable. It's easy to ask big, broad questions, but they often can't be effectively answered through research. To make sure my question is testable, I always ask myself: "Can I gather data to answer this question? Can I design a study or experiment that would provide evidence for or against this question?" If the answer is 'no', it means I need to refine my question.

Audience is important because it may be multi specialities involvement. The stakeholders sometimes we may find very difficult to make them understand.

In my experience, it is imperative to synthesize your own knowledge based on broad information. This enables you to capture specific themes across a topic and then prompt more granular and specific questions to further your research.

If you “have defined your topic”, it indicates that you have a specific aspect of interest. Reviewing literature comes into play when you have ideas about potential topics and wish to assess them. Once you have defined your topic, you conduct a literature review that focuses on your topic. This review aims to summarise various knowledge and opinions from peers within the field.

When generating questions, a common mistake is to stop once you have a question you like. Try to force yourself past this to generate more questions - at least 10 or 15. If you get stuck, go with some issue development techniques, like turning the questions you already have around or inverting them, or changing their scale and scope. Also, when refining your questions, closely examine them to identify any assumptions you make inside them (e.g. 'how often do Russian actors target us' assumes Russian actors are targeting us, and is now just asking how often). Assumptions inside questions aren't necessarily bad, but they must be identified and examined to help ensure they're valid.

Unless the topic of your research specifically involves brainstorming, it is not advisable to rely on brainstorming to generate research questions. The research questions should be formulate d based on appropriate research techniques and methodology.

Getting feedback from others is vital here. If/when the analyst selects the question, they will pick one that appeals to them or that they know the most about. When a non-expert colleague selects what they think is most relevant, it is often closer to what your end consumer will also find relevant.

A broad literature survey may be needed to grasp the wider implications of a specific research question. Often, basic science questions take second priority with funding bodies, but it is important to stress that key processes that become altered in disease can be manipulated to achieve therapeutic goals. The most notorious example in my field is the development of a drug with remarkable effects in treating male erectile dysfunction that was initially formulated as an anti-hypertensive. Thus, targetting a receptor may have unknown effects in the whole organism, and this can not be overlooked when designing experiments.

I strongly recommend deconstructing your question into component parts. Often, we get bogged down trying to find an answer to the main question as a single issue, and get lost in the endless research loop thinking 'if I know everything there is to know I'll be able to answer this.' Instead, break the question into several smaller, more answerable portions. Start by asking yourself 'what would I need to know to be able to answer this question' to give yourself smaller, more easily accomplished research tasks that can be compiled into answering your main question.

Acknowledge the weaknesses in your question and be open about what the limits of your research and it’s implications are because. A savvy audience, especially one you are trying to influence, will already be doing this so your failure to be candid will appear either deceitful or ignorant. Acknowledge the limitations while emphasizing the strengths.

Be sure to start with a clear null hypothesis. To do this - assume a treatment has no effect (as in no relationship between the variables exists) and write the hypothesis to reflects this.

RQS should be driven by gaps in existing literature. If you know your participants, applicable statistical approach, and what you are seeking to resolve, there should be an easy movement to research questions. If you are having difficulty writing them, then I would suggest you do not yet have the necessary depth of knowledge on the research problem or study population. If a self examination indicates confidence in these two areas, then you may be faltering in understanding the expectations/limits of your statistical analysis tools. Knowing your independent/dependent variables is key to determining if what you propose is a match for your analysis approach. Finally, if you are still having trouble making the RQs work, repeat the process.