Write the research design appropriate to your research problem

As we explained last week, Motivated Academics are curious to brainstorm and test new research ideas. But once you come up with your research idea and develop your hypothesis, how do you know what type of research design to use so that you collect reliable data? That’s where research design comes into the academic picture! 

I want to emphasise one thing – the generation of new research ideas is one of the most crucial aspects that you will need to learn as an academic. Each research project and scholarly paper depends on your ideas. But great ideas without efficient research design process and research execution are just that – great ideas without meaningful outputs. 

That’s why Motivated Academics are eager to TEST their innovative concepts and ideas to make sure these bring tangible benefits to the economy, society and environment we all live in.  

How do you test your ideas then? Well, that’s what research design is for! 

What is research design? 

Let me first define what’s research design. According to Oxford Reference, a research design is “a detailed proposal relating to a defined piece of [research] endeavour, which includes a definition of a problem, subject, or hypothesis for investigation; the background and context to the investigation; the proposed means and methods of the investigation; the work plan and timetable; details of the proposed investigators, management arrangements, and quality control procedures; and a table of costs.”

This definition was originally presented for archaeological endeavours and activities, but I believe it can be extended to any research area. At the end of the day, it is all about defining the design of a scientifically sound method that will allow you to produce a sufficient amount of data to test your hypothesis. 

How to design your research in 6 simple steps? 

Now that you know what research design is, let me share a simple 7-step framework that will help you develop a well-planned research design. I’ll try to support this research design with example where possible. 

Step 1: Reflect on your hypothesis

The hypothesis you’ve developed for your research includes a piece of crucial information about your research question and variables that you want to evaluate.

Therefore, your research design should start by reflecting on your research hypothesis. You should be able to answer the following questions: 

• what variables or parameters you should investigate?
• what information do you need to collect in your research to test your hypothesis?
• how much data do you need to test your hypothesis? 

This information is crucial in the next step. 

Step 2: Decide on research approach

Now there are two main approaches to your research approach: qualitative approach or quantitative approach. 

The qualitative approach aims to characterise the relationships between the dependent and independent variables using mathematical tools to correlate the observations from your data. This usually generates quantitative data that are numerical and help you evaluate different case studies.

The quantitative approach aims to characterise the relationship between the dependent and independent variables using observations and non-mathematical data (pictures, recordings, text, transcripts, interviews, focus groups) to draw insights and conclusions.

You can also use mixed methods that combine both qualitative and quantitative approaches. With this approach to research, you can get more comprehensive answers to your research problem and you can deduct more informed conclusions. 

Step 3: Select the type of research design

Wondering what are 4 types of research design? Here is the answer. 

For quantitative research design, you can use experimental design, quasi-experimental design, correlational design, and descriptive design for your quantitative research. Here’s a brief overview of these research design types: 

An experimental design seeks to determine a cause and effect correlation of variables that you study. Your experimental research design usually involves independent and dependent variables, and you are trying to understand how the change in the independent variable influences the dependent variables. The samples you use are often randomised. Your experiments are usually performed in a controlled environment.

A quasi-experimental design is similar to the experimental design, but it relies on pre-existing groups (non-randomised samples). Your experiments are usually performed in a natural environment. 

A correlational design focuses on determining whether a correlation exists between the variables under question and how strongly these are correlated. It heavily relies on a statistical analysis of your data. 

A descriptive design relies on understanding the characteristics of a single sample, such as averages and standard deviations. This type of design may not need a hypothesis as usually, you are focusing on understanding the existing data. 

For qualitative research design, the most commonly used types of research design include case study, ethnography, grounded theory, and phenomenology. 

Case study design aims to understand and explain the experience of a defined sample via direct observation and interaction with that sample. The focus is placed on the description of the experience. 

Ethnography design aims to describe the characteristics of a selected culture. This is achieved via a selection of variables and characteristics to consider in the study and data collection via literature review, culture immersion, direct observation and interaction with subjects. The outcome is the description of culture. 

Grounded theory design aims to discover the problems and challenges in society and how members of society deal with these. It involves an iterative process of “formulation, testing and re-development of propositions until a theory is developed”. The data is collected via interviews or observations. The outcome is a new theory that is supported with data. 

Phenomenology design aims to describe the experiences from the individuals perspective. It aims to understand the feelings and experiences with regard to a specific phenomenon. Usually, there are no defined steps in terms of the data collections to allow space for creativity. The outcome is the understanding of the subject’s point of view and structural explanation of the specific phenomenon. 

Step 4: Define your population and sampling method

Once you decided on the research design process that will provide the best set of data to test your hypothesis, it’s crucial to define the amount of data, aka. population, and sampling method. 

I often find that researchers tend to generate way too much data, which makes it difficult to decide what data to include in your chapter or research article.

Therefore, it’s crucial to decide the characteristics of the population you want to study at the outset of your research to maintain focus. For example, your population can be characterised by a specific demographic or comprise a given set of technologies. 

Once you understand what data you need to collect to test your hypothesis, you need to decide how this data will be sampled. This is in particular important when you’re dealing with large populations, e.g., people in a specific country, and it is impossible to get data on all of them. Instead, you will collect data based on a representative sample of that population. But let’s leave a discussion like this for our further articles. 

To sample data from your population, you can use two main sampling methods. 

Probabilistic sampling relies on using random methods to generate data, usually involving some sort of mathematical representation of the population you evaluate. It allows you to truly understand the correlations and characteristics of the population via thorough statistical analysis. 

Deterministic sampling, aka, non-probabilistic sampling, relies on non-random selection of the sample and is usually easier to perform. However, it is subject to bias depending on how your sample is selected. 

Step 5: Select data collection method 

Now that you know what data you need to collect, it is important to decide how exactly you are going to collect your data. This is one of the most crucial aspects of the research design process, as you will actually decide what kind of tools you will use in your research. 

At this stage, it is crucial to distinguish between the primary data and secondary data.

The former focuses on collecting new data via surveys, questionnaires, interviews, observations, mathematical modelling and so on.

The latter uses the data that has already been published in the literature. Therefore, the collection of the secondary data largely relies on literature review, database search, raw data mining and so on. 

Step 6: Design your data collection process

We’re almost there! So you know what methods you’re going to use, what parameters to measure, and how to sample them. Now it’s time to determine the minimum amount of data necessary for your research. 

Remember, you only need to generate a sufficient amount of data that is necessary to prove or disprove your hypothesis. Oversampling will neither give you more additional information nor improve the accuracy of your results – it will just consume more of your time during the data collection period. 

To determine the minimum amount of data necessary for your research, you may use rules of thumb in your area (i.e. 3 repetitions of the same experiment), Taguchi design or fractional factorial design. 

I feel that it is important to mention that you also need to develop a data management plan as early as possible in your research. This is especially true during questionnaire design or qualitative design, as you need to make sure you properly deal with the sensitive data and adhere to your funder’s data management requirements. 

Importantly, when you decide on the amount of data to collect, your research design must have the following characteristics:

• Objectivity: the research methods you use in your work should allow you to objectively test your hypothesis. This means that your results should be free from bias and should not include subjective opinions. This is difficult to achieve, as we tend to promote our interests. We also make assumptions in our work. Therefore, for others to understand what you did and why you did it, it is essential to explicitly state the assumptions and approach of your work. It is important to ensure that your work could be replicated by other researchers so that they could verify your results and conclusions, and built on your research

• Repeatability: the research method you use in your work should be reliable and accurate. This means that whenever you conduct similar research under similar conditions using the same method, your results should be similar and lead to the same conclusions. This is especially true for statistical and experimental work, so your research design must include a sufficient number of observations to allow for uncertainty analysis. This is to ensure that your results are representative and not influenced by the research design. This is also important for any sort of computational or analytical research, where you develop models or correlations to represent the system or concept that you analyse. Why? Well, you need to make sure that the methods you used are valid, which means these represent the actual system accurately.

• Validity: the research method that you select needs to support you in testing and verifying your hypothesis, and subsequently answering your research questions. Your methods should help you to produce and analyse your data objectively to draw conclusions in the context of your hypothesis and research question. When selecting methods ask yourself what kind of data you’ll need to test your hypothesis.

• Generalisation: finally, it is important that you select the appropriate size of your test sample so that the analysis you do is representative and can be generalised to the entire population, not just the sample that you’re considering. For example, if you’re analysing whether sharing research via social media increases the visibility of your profile, you should include at least 100, or in ideal case 1000, researchers in your study so that you could generalise the findings.

Step 7: Develop approach to data analysis

Now that you’ve designed the data collection process, it’s time to decide how you are going to analyse your data. You can do it by using the qualitative and quantitative data analysis tools, following the research design approach you selected in Step 2. 

For the qualitative data analysis, you’ll use your statistical analysis skills including descriptive statistics, inferential statistics, and hypothesis testing. 

Descriptive statistics will help you to understand the characteristics of your data, especially in terms of statistical distribution, the central tendency of data and variability of data. The inferential statistics will help you understand the relationships between your data, and generalise the characteristics of the population based on your sample. Finally, using hypothesis testing you can understand whether the correlations you derived are statistically significant. 

For qualitative data analysis, you may consider using thematic analysis or discourse analysis. The former focuses on the understanding the data content and its wider implications to determine key themes. The latter focuses on contextualising the data and is commonly used to understand the communication patterns. 

Of course, other qualitative and quantitative tools do exist. You would usually find these via a comprehensive literature review of your research field. 

Conclusions

Now you understand why research design is a crucial aspect of each research project, especially at the early stage of your PhD degree.

Having a clear path to the delivery of your research will reduce the anxiety and uncertainty associated with your work. Follow the 7-step process outlined in this article to develop your research design! 

But there is one thing that I want to share with you here that helped me to successfully complete my PhD and publish my work in top journals. When you follow your initial research design process, you’ll generate new ideas and avenues – don’t be worried to pivot from your initial plan as you go through your project. 

Always follow the most exciting ideas, as long as they are consistent with the main objective of your project! 

What kind of research do you pursue? Qualitative or quantitative? 

PrevPrevious5 ideas to deal with common PhD challenges

Next10 tips for delivering an academic presentation with impactNext