Starting Your Research

This post is a revised version of an article I wrote in 2016.

Having supervised many postgraduate and undergraduate research students and consistently providing a set of tips and advice to each new student embarking on a research project, I decided to compile and document these guidelines. While these tips are tailored for my students, I trust they may prove beneficial to others as well. However, make sure to consult with your supervisor, as they may employ a distinct approach to initiating research.

Identifying a Research Problem

Identifying and understanding a research problem to work on is the hardest part of any research endeavour. Instead of fabricating a hypothetical problem, seek out real-life problems in your surroundings. Look for a felt need to enhance an unsatisfactory solution, improve its efficiency or scalability, or even devise an alternative solution/process to tackle the problem more effectively. Look for problems within your home, neighbourhood, workplace, university, social circles, volunteer organisations, or even during your daily commute. Conduct a brief online search on the problem and potential solutions. If people are talking about it, it means the problem is relevant. Don’t be disheartened if solutions already exist; strive to devise an even better one.

Discuss with people about the problem(s) you observe, encouraging them to share their perspectives on the issues they’ve identified. Get their feedback on the problem(s) you are contemplating, as others may offer insights that haven’t crossed your mind. While this means “fail fast” for many of your ideas, it also translates to swift learning within your area of interest.

A well-defined research problem should be relevant, worth solving, and manageable. The justification for relevance often entails articulating a research gap, signifying an area/topic within a particular field of study that has either not been thoroughly investigated or is yet to be explored. Additionally, there must be inherent value in addressing and bridging this research gap, ideally aligning with industry demands rather than necessitating a proactive imposition of the research onto the industry.

I encourage you to formulate a couple of potential research problems before the initial discussion with the supervisor(s). Although you and your supervisor(s) may or may not pursue these specific problems, engaging in this exercise prompts you to contemplate identifying a pertinent problem or felt need in your surroundings. The ability to recognise and identify problems is a crucial skill imparted through the research process. For instance, the capacity to pinpoint issues in a product developed by the company you work and propose simple solutions enhances your value as a collaborative team member, potentially translating into tangible benefits, including financial ones.

The potential problems you articulate also help your supervisor(s) in identifying your preferred areas of focus. It’s important to note that, by the end of our initial discussion, several of the initially proposed problems may be set aside for various reasons, such as practical constraints, resource limitations, excessive theoretical nature, lack of a substantial research component, or misalignment with program, personal, or mutual interests. Throughout your conversation, I may also introduce project ideas aligned with the issues you bring forward. Following subsequent rounds of in-depth discussions and a brief literature survey, you could narrow down your focus to a specific research area. Ultimately, you will refine and modify 1-2 problems that align with both your and your supervisor’s interests.

Here are some resources for identifying a research problem:

Conducting a Literature Survey

Once you’ve pinpointed a field of interest and identified 1 or 2 potential research problems, the next step is to dive into extensive reading across the domain, problems, and existing solutions. This phase is referred to as the literature survey. Conducting a thorough literature survey serves as a vital foundation, providing insights into the current state of the field, the significance of the problem at hand, various influencing factors and parameters, and potential approaches to solving it.

Moreover, a comprehensive literature survey can prevent the awkward situation of unknowingly replicating existing work. Imagine reaching your thesis defence or presenting at a conference, only to have one of the audience members or reviewers point out that similar research had already been completed! Avoiding such situations ensures that your work remains current, innovative, and contributes meaningfully to the existing body of knowledge.

Initially read widely from any potential source, including conference and journal papers, white papers, industry articles, Wikis, blogs, and more. If your field of interest is well-established, give priority to formal literature like conference and journal papers. However, if it’s an emerging area or if industry advancements outpace academic publications, grey literature such as white papers, blogs, and Wikis becomes particularly relevant.

Maintain organised records of the literature you delve into, noting details like titles, authors, publication sources, and URLs. Consider adopting techniques like highlighting and note-taking for physical or digital copies of research papers. Annotate PDFs or save web pages as PDFs for efficient note-taking. Alternatively, craft a concise summary in a separate digital journal. Post-it is not a good idea as they may not survive your long research endeavour.

As you build your foundational knowledge, narrow your literature search towards specific topics and problems that align with your research problem(s). Pay attention to emerging patterns and themes within the literature you explore, and use these insights to hone in on your research problem. The more you delve into the literature, the more refined your research problem becomes. In some instances, you might even find the need to completely reshape your initial research problem.

Throughout this phase, an undergraduate project should encompass a minimum of 8+ key literature sources, while an MSc student should aim for 15+ key papers. For those pursuing a PhD, a comprehensive literature review may involve covering 30+ key publications. Total number of papers you are likely/expected to list in your proposal is twice as this. Here are some resources to guide you in conducting a thorough literature survey:

The following tools are useful for keeping track of literature sources:

Formulating the Problem Statement

Once you have a preliminary understanding of your research problem, it’s time to start documenting it. Many students believe they know their research problem, but when pressed to explain it to someone else or put it into writing, they often encounter challenges. This can be a clear sign that a deeper understanding of the research problem is needed. According to Einstein, having a well-defined problem means you’ve completed most of your research, as working on a clearly outlined problem is far more manageable than dealing with a vague one. While it might be challenging to precisely define your research problem, it’s crucial to make an early attempt, even knowing that your problem statement may evolve or change as you progress and gain a better understanding of the problem and its limitations.

You need to articulate and specify your research problem clearly, aka the problem statement. Your problem statement should effectively address 3 fundamental questions, in any suitable order:

  • “What” is your problem?
  • “Why” is it important?
  • “How” are you planning to address the problem?

While these 3 questions come from a different domain (see Start With Why – Simon Sinek TED talk), their relevance extends to the realm of research.

Initially, try to come up with a 1/2 page write-up addressing the above 3 questions. Usually, it’s easier to start with why. It’s often helpful to begin with the “why,”. Presenting your thoughts in a paragraph format is recommended. If you encounter challenges when putting your research problem into words, consider it a sign that further clarity is needed. In such cases, don’t hesitate to discuss the difficulties with your supervisor. Another beneficial practice is attempting to articulate your problem to colleagues who may not be well-versed in your field of study. Being able to explain your problem to someone with a basic understanding of Computer Science, but not necessarily an expert, is a valuable skill. Additionally, delving into more literature can aid in refining and better defining your research problem.

Once you’ve gained more clarity on your problem statement, it’s time to craft a detailed version, typically spanning around 300 words. In a project proposal or thesis, this detailed articulation is commonly found in Chapter 1 or 3. For project proposals or literature surveys, the elaboration is usually concise and integrated into Chapter 1. In a thesis, a comprehensive problem statement is typically located in Chapter 3. In creating your detailed problem statement, make sure to include the following components:

  • Background/motivation: Offer sufficient background to the problem at hand and discuss its significance. If applicable, briefly discuss a couple of use cases to further contextualize your research problem. Use this section as a motivational preamble to your research problem.
  • Problem statement: Clearly articulate the problem you aim to tackle. While real-world problems are often intricate and expansive, a research project typically addresses a well-defined sub-problem within the larger issue. Explicitly outline the scope of your planned work and state any underlying assumptions. Supervisors and review panels often provide feedback on problem statements, commonly noting that the scope is too much. This is the section where you delve into the previously mentioned “what.”
  • Research gap: Considering the current body of knowledge related to the problem and its potential solutions, as discovered through your literature survey, identify what is missing. Highlight the aspects that make your problem-solving efforts worthwhile (i.e., “why”).
  • Proposed solution: Provide a succinct overview of your proposed solution (i.e., “how”), accompanied by specific goals and objectives. Justify your belief in the efficacy of your solution and discuss the approach to achieving each objective. Even before diving into the research, it’s crucial to have a preliminary understanding of how you plan to address the problem—this preliminary plan is known as the research methodology. Include high-level diagrams illustrating the solution and its main components. Using a diagram is an effective way to visually communicate your research methodology, which is the process of tackling your problem.
  • Research question: Craft a single-sentence research question that encapsulates the problem you aim to solve. Typically starting with phrases like “What is…” or “How to…,” the research question serves as a concise focal point for your investigative efforts.

Occasionally, the problem statement is formulated in 2 distinct steps, known as the problem statement and the purpose statement. You can compose them together or separately, depending on the requirements of your degree program, the nature of your study, and your supervisor’s preferences. Here are some helpful resources:

Writing the Literature Survey

The literature review is like a growing tree, where a couple of key papers lead to many relevant-looking ones, and those, in turn, guide you to even more. It’s an ever-expanding process. However, there comes a point when you need to put a halt to your quest for additional literature. This pause is best initiated once you’ve developed a comprehensive understanding of your area of interest, and your research problem becomes clear.

For those who strive for perfection, the temptation to continue delving into more papers may persist, potentially delaying the actual research process. Conversely, some might jump into their research after perusing only a handful of literature sources, risking overlooking crucial insights. Finding the right balance is a challenge, but paying attention to cues from your supervisor(s) can provide valuable guidance on when to slow down the literature review.

In practice, a literature review isn’t a one-and-done affair. It persists until you finalise the last version of your thesis. Throughout various phases, you may find yourself seeking additional literature for different purposes:

  • Solution design: Crafting a solution often involves exploring numerous components. Whether it’s deciding on algorithms, databases, tools, middleware, or survey techniques, you may need to explore multiple alternatives before settling on the most suitable ones.
  • Solution implementation: As you implement your proposed solution, conduct simulations, or carry out surveys, you might need to seek further literature. This could involve discovering configuration/simulation parameters or refining your survey questions for more effective results.
  • Writing the thesis: When you begin writing your thesis, the need for additional literature arises. It’s crucial for building robust arguments supporting your problem statement and proposed solution. This additional literature also helps in filling any gaps, ensuring a clear and comprehensive narrative.
  • Long-term research: For extensive research projects spanning several years, such as PhD studies or part-time research, staying abreast of new publications is essential. Subscribe to relevant websites, track new papers from conferences and journals, and leverage tools like Google News or Scholar alerts to stay informed.

Once you’ve absorbed the relevant literature, documenting your findings becomes imperative. This documentation typically becomes Chapter 2 of your thesis. While preparing this chapter, critically evaluate the literature, covering seminal papers, alternative designs, emerging patterns, and conducting comparisons. Summarise the content in your own words to reduce similarity scores and avoid potential plagiarism. Be sure to incorporate a discussion on how one piece of work is interconnected with others, highlighting both its advantages and drawbacks. If you find yourself struggling and constantly referring to the original papers, it indicates a need to revisit and thoroughly understand the ideas presented. This is where effective reading strategies, such as highlighting and annotating, come into play to ensure your write-up tells a clear, concise, and cohesive story.

Before starting the literature review, create a detailed outline. Arrange the topics based on their importance and interconnections. This not only helps in identifying any missing elements but also ensures a coherent flow in your writing. It’s a more efficient approach than rearranging topics after completing the entire write-up. This planning stage may also help identify additional literature sources to explore. Ensure your literature review encompasses the following elements:

  • Summary of related solutions: Provide a summary covering various solutions, technologies, algorithms, and tools relevant to your research. Even if your focus is on a specific technology or tool, explain other related options. Justify your decision to work with the chosen technology or tool and clarify why it was preferred over others.
  • Motivation and justification: Utilise the literature survey to both motivate and justify your research problem. Clearly articulate the reasons behind your problem and demonstrate its significance within the existing body of knowledge.
  • Visual aids: Include figures and tables where applicable. Avoid using screenshots from original sources; instead, create your visuals to enhance clarity and originality.
  • Critical evaluation: Critically assess each technology, algorithm, and tool. Discuss their respective strengths and weaknesses, assess the credibility of assumptions, identify any missing elements, delve into performance results, and explain the applicability or inapplicability of each solution to your specific problem.
  • Solution summaries and comparisons: Summarise each solution/technique and present a comparative analysis between comparable ones. Consider presenting this information in the form of a table for clarity.
  • Coherent structure: Ensure the contents of your review are related, connected, and maintain a smooth flow. Start with independently explainable topics, progressing to those with dependencies on previously discussed subjects. For example, if your project involves processing big data from Internet of Things (IoT) applications, begin by explaining what big data and IoT are before delving into similar solutions.
  • Proper citation: Accurately cite all relevant literature sources in-text to uphold academic integrity and acknowledge the contributions of others. Depending on your program, consider using referencing styles such as IEEE, ACM, or APA.
  • Language and style: Pay attention to the language and style. Opt for an academic writing style as it prioritizes clarity and conciseness in expressing ideas and findings. Be careful when using language editing and generative AI tools, as they often adopt a more collaborative style, potentially distorting the intended meaning and compromising accuracy.

Explore the following resources for guidance on academic writing and composing an effective literature survey: