If you’re learning about research skills and methodologies, you may have heard the term “research onion”. Specifically, the research onion developed by Saunders et al in 2007.
If you’re learning about research skills and methodologies, you may have heard the term “research onion”. Specifically, the research onion developed by Saunders et al in 2007. But what exactly is this elusive onion? In this post, we’ll break Saunders’ research onion down into bite-sized chunks to make it a little more digestible.
At the simplest level, Saunders’ research onion describes the different decisions you’ll need to make when developing a research methodology – whether that’s for your dissertation, thesis or any other formal research project. As you work from the outside of the onion inwards, you’ll face a range of choices that progress from high-level and philosophical to tactical and practical in nature. This also mimics the general structure for the methodology chapter.
While Saunders’ research onion is certainly not perfect, it’s a useful tool for thinking holistically about methodology. At a minimum, it helps you understand what decisions you need to make in terms of your research design and methodology.
The onion is made up of 6 layers, which you’ll need to peel back one at a time as you develop your research methodology:
The very first layer of the onion is the research philosophy. But what does that mean? Well, the research philosophy is the foundation of any study as it describes the set of beliefs the research is built upon. Research philosophy can be described from either an ontological or epistemological point of view. “A what?!”, you ask?
In simple terms, ontology is the “what” and “how” of what we know – in other words, what is the nature of reality and what are we really able to know and understand. For example, does reality exist as a single objective thing, or is it different for each person? Think about the simulated reality in the film The Matrix.
Epistemology, on the other hand, is about “how” we can obtain knowledge and come to understand things – in other words, how can we figure out what reality is, and what the limits of this knowledge are. This is a gross oversimplification, but it’s a useful starting point (we’ll cover ontology and epistemology another post).
With that fluffy stuff out the way, let’s look at three of the main research philosophies that operate on different ontological and epistemological assumptions:
These certainly aren’t the only research philosophies, but they are very common and provide a good starting point for understanding the spectrum of philosophies.
Positivist research takes the view that knowledge exists outside of what’s being studied. In other words, what is being studied can only be done so objectively, and it cannot include opinions or personal viewpoints – the researcher doesn’t interpret, they only observe. Positivism states that there is only one reality and that all meaning is consistent between subjects.
In the positivist’s view, knowledge can only be acquired through empirical research, which is based on measurement and observation. In other words, all knowledge is viewed as a posteriori knowledge – knowledge that is not reliant on human reasoning but instead is gained from research.
For the positivist, knowledge can only be true, false, or meaningless. Basically, if something is not found to be true or false, it no longer holds any ground and is thus dismissed.
Let’s look at an example, based on the question of whether God exists or not. Since positivism takes the stance that knowledge has to be empirically vigorous, the knowledge of whether God exists or not is irrelevant. This topic cannot be proven to be true or false, and thus this knowledge is seen as meaningless.
Kinda harsh, right? Well, that’s the one end of the spectrum – let’s look at the other end.
On the other side of the spectrum, interpretivism emphasises the influence that social and cultural factors can have on an individual. This view focuses on people’s thoughts and ideas, in light of the socio-cultural backdrop. With the interpretivist philosophy, the researcher plays an active role in the study, as it’s necessary to draw a holistic view of the participant and their actions, thoughts and meanings.
Let’s look at an example. If you were studying psychology, you may make use of a case study in your research which investigates an individual with a proposed diagnosis of schizophrenia. The interpretivist view would come into play here as social and cultural factors may influence the outcome of this diagnosis.
Through your research, you may find that the individual originates from India, where schizophrenic symptoms like hallucinations are viewed positively, as they are thought to indicate that the person is a spirit medium. This example illustrates an interpretivist approach since you, as a researcher, would make use of the patient’s point of view, as well as your own interpretation when assessing the case study.
Pragmatism highlights the importance of using the best tools possible to investigate phenomena. The main aim of pragmatism is to approach research from a practical point of view, where knowledge is not fixed, but instead is constantly questioned and interpreted. For this reason, pragmatism consists of an element of researcher involvement and subjectivity, specifically when drawing conclusions based on participants’ responses and decisions. In other words, pragmatism is not committed to (or limited by) one specific philosophy.
Let’s look at an example in the form of the trolley problem, which is a set of ethical and psychological thought experiments. In these, participants have to decide on either killing one person to save multiple people or allowing multiple people to die to avoid killing one person.
This experiment can be altered, including details such as the one person or the group of people being family members or loved ones. The fact that the experiment can be altered to suit the researcher’s needs is an example of pragmatism – in other words, the outcome of the person doing the thought experiment is more important than the philosophical ideas behind the experiment.
To recap, research philosophy is the foundation of any research project and reflects the ontological and epistemological assumptions of the researcher. So, when you’re designing your research methodology, the first thing you need to think about is which philosophy you’ll adopt, given the nature of your research.
Let’s peel off another layer and take a look at the research approach. Your research approach is the broader method you’ll use for your research – inductive or deductive. It’s important to clearly identify your research approach as it will inform the decisions you take in terms of data collection and analysis in your study (we’ll get to that layer soon).
Inductive approaches entail generating theories from research, rather than starting a project with a theory as a foundation. Deductive approaches, on the other hand, begin with a theory and aim to build on it (or test it) through research.
Sounds a bit fluffy? Let’s look at two examples:
An inductive approach could be used in the study of an otherwise unknown isolated community. There is very little knowledge about this community, and therefore, research would have to be conducted to gain information on the community, thus leading to the formation of theories.
On the other hand, a deductive approach would be taken when investigating changes in the physical properties of animals over time, as this would likely be rooted in the theory of evolution. In other words, the starting point is a well-established pre-existing body of research.
Closely linked to research approaches are qualitative and quantitative research. Simply put, qualitative research focuses on textual, visual or audio-based data, while quantitative research focuses on numerical data. To learn more about qualitative and quantitative research, check out our dedicated post here.
What’s the relevance of qualitative and quantitative data to research approaches? Well, inductive approaches are usually used within qualitative research, while quantitative research tends to reflect a deductive approach, usually informed by positivist philosophy. The reason for using a deductive approach here is that quantitative research typically begins with theory as a foundation, where progress is made through hypothesis testing. In other words, a wider theory is applied to a particular context, event, or observation to see whether these fit in with the theory, as with our example of evolution above.
So, to recap, the two research approaches are inductive and deductive. To decide on the right approach for your study, you need to assess the type of research you aim to conduct. Ask yourself whether your research will build on something that exists, or whether you’ll be investigating something that cannot necessarily be rooted in previous research. The former suggests a deductive approach while the latter suggests an inductive approach.
So far, we’ve looked at pretty conceptual and intangible aspects of the onion. Now, it’s time to peel another layer off that onion and get a little more practical – introducing research strategy. This layer of the research onion details how, based on the aims of the study, research can be conducted. Note that outside of the onion, these strategies are referred to as research designs.
There are several strategies you can take, so let’s have a look at some of them.
Experimental research involves manipulating one variable (the independent variable) to observe a change in another variable (the dependent variable) – in other words, to assess the relationship between variables. The purpose of experimental research is to support, refute or validate a research hypothesis. This research strategy follows the principles of the scientific method and is conducted within a controlled environment or setting (for example, a laboratory).
Experimental research aims to test existing theories rather than create new ones, and as such, is deductive in nature. Experimental research aligns with the positivist research philosophy, as it assumes that knowledge can only be studied objectively and in isolation from external factors such as context or culture.
Let’s look at an example of experimental research. If you had a hypothesis that a certain brand of dog food can raise a dogs’ protein levels, you could make use of experimental research to compare the effects of the specific brand to a “regular” diet. In other words, you could test your hypothesis.
In this example, you would have two groups, where one group consists of dogs with no changes to their diet (this is called the control group) and the other group consists of dogs being fed the specific brand that you aim to investigate (this is called the experimental/treatment group). You would then test your hypothesis by comparing the protein levels in both groups.
Next, we have action research. The simplest way of describing action research is by saying that it involves learning through… wait for it… action. Action research is conducted in practical settings such as a classroom, a hospital, a workspace, etc – as opposed to controlled environments like a lab. Action research helps to inform researchers of problems or weaknesses related to interactions within the real-world. With action research, there’s a strong focus on the participants (the people involved in the issue being studied, which is why it’s sometimes referred to as “participant action research” or PAR.
An example of PAR is a community intervention (for therapy, farming, education, whatever). The researcher comes with an idea and it is implemented with the help of the community (i.e. the participants). The findings are then discussed with the community to see how to better the intervention. The process is repeated until the intervention works just right for the community. In this way, a practical solution is given to a problem and it is generated by the combination of researcher and community (participant) feedback.
This kind of research is generally applied in the social sciences, specifically in professions where individuals aim to improve on themselves and the work that they are doing. Action research is most commonly adopted in qualitative studies and is rarely seen in quantitative studies. This is because, as you can see in the above examples, action research makes use of language and interactions rather than statistics and numbers.
A case study is a detailed, in-depth study of a single subject – for example, a person, a group or an institution, or an event, phenomenon or issue. In this type of research, the subject is analysed to gain an in-depth understanding of issues in a real-life setting. The objective here is to gain an in-depth understanding within the context of the study – not (necessarily) to generalise the findings.
It is vital that, when conducting case study research, you take the social context and culture into account, which means that this type of research is (more often than not) qualitative in nature and tends to be inductive. Also, since the researcher’s assumptions and understanding play a role in case study research, it is typically informed by an interpretivist philosophy.
For example, a study on political views of a specific group of people needs to take into account the current political situation within a country and factors that could contribute towards participants taking a certain view.
Next up, grounded theory. Grounded theory is all about “letting the data speak for itself”. In other words, in grounded theory, you let the data inform the development of a new theory, model or framework. True to the name, the theory you develop is “grounded” in the data. Ground theory is therefore very useful for research into issues that are completely new or under-researched.
Grounded theory research is typically qualitative (although it can also use quantitative data) and takes an inductive approach. Typically, this form of research involves identifying commonalities between sets of data, and results are then drawn from completed research without the aim of fitting the findings in with a pre-existing theory or framework.
For example, if you were to study the mythology of an unknown culture through artefacts, you’d enter your research without any hypotheses or theories, and rather work from the knowledge you gain from your study to develop these.
Ethnography involves observing people in their natural environments and drawing meaning from their cultural interactions. The objective with ethnography is to capture the subjective experiences of participants, to see the world through their eyes. Creswell (2013) says it best: “Ethnographers study the meaning of the behaviour, the language, and the interaction among members of the culture-sharing group.”
For example, if you were interested in studying interactions on a mental health discussion board, you could use ethnography to analyse interactions and draw an understanding of the participants’ subjective experiences.
For example, if you wanted to explore the behaviour, language, and beliefs of an isolated Amazonian tribe, ethnography could allow you to develop a complex, complete description of the social behaviours of the group by immersing yourself into the community, rather than just observing from the outside.
Given the nature of ethnography, it generally reflects an interpretivist research philosophy and involves an inductive, qualitative research approach. However, there are exceptions to this – for example, quantitative ethnography as proposed by David Shafer.
Last but not least is archival research. An archival research strategy draws from materials that already exist, and meaning is then established through a review of this existing data. This method is particularly well-suited to historical research and can make use of materials such as manuscripts and records.
For example, if you were interested in people’s beliefs about so-called supernatural phenomena in the medieval period, you could consult manuscripts and records from the time, and use those as your core data set.
As you can see, there is a wide range of choices in terms of research strategy. The right choice for your project will depend largely on your research aims and objectives, as well as the choices you make in terms of research philosophy and approach.
Onion Layer 4: Choices
The next layer of the research onion is simply called “choices” – they could have been a little more specific, right? In any case, this layer is simply about deciding how many data types (qualitative or quantitative) you’ll use in your research. There are three options – mono, mixed, and multi-method.
Let’s take a look at them.
Choosing to use a mono method means that you’ll only make use of one data type – either qualitative or quantitative. For example, if you were to conduct a study investigating a community’s opinions on a specific pizza restaurant, you could make use of a qualitative approach only, so that you can analyse participants’ views and opinions of the restaurant.
If you were to make use of both quantitative and qualitative data, you’d be taking a mixed-methods approach. Keeping with the previous example, you may also want to assess how many people in a community eat specific types of pizza. For this, you could make use of a survey to collect quantitative data and then analyse the results statistically, producing quantitative results in addition to your qualitative ones.
Lastly, there’s multi-method. With a multi-method approach, you’d make use of a wider range of approaches, with more than just a one quantitative and one qualitative approach. For example, if you conduct a study looking at archives from a specific culture, you could make use of two qualitative methods (such as thematic analysis and content analysis), and then additionally make use of quantitative methods to analyse numerical data.
As with all the layers of the research onion, the right choice here depends on the nature of your research, as well as your research aims and objectives. There’s also the practical consideration of viability – in other words, what kind of data will you be able to access, given your constraints.
What’s that far in the distance? It’s the time horizon. But what exactly is it? Thankfully, this one’s pretty straightforward. The time horizon simply describes how many points in time you plan to collect your data at. Two options exist – the cross-sectional and longitudinal time horizon.
Imagine that you’re wasting time on social media and think, “Ooh! I want to study the language of memes and how this language evolves over time”. For this study, you’d need to collect data over multiple points in time – perhaps over a few weeks, months, or even years. Therefore, you’d make use of a longitudinal time horizon. This option is highly beneficial when studying changes and progressions over time.
If instead, you wanted to study the language used in memes at a certain point in time (for example, in 2020), you’d make use of a cross-sectional time horizon. This is where data is collected at one point in time, so you wouldn’t be gathering data to see how language changes, but rather what language exists at a snapshot point in time. The type of data collected could be qualitative, quantitative or a mix of both, as the focus is on the time of collection, not the data type.
As with all the other choices, the nature of your research and your research aims and objectives are the key determining factors when deciding on the time horizon. You’ll also need to consider practical constraints, such as the amount of time you have available to complete your research (especially in the case of a dissertation or thesis).
Finally, we reach the centre of the onion – this is where you get down to the real practicalities of your research to make choices regarding specific techniques and procedures.
Specifically, this is where you’ll:
What’s important to note here is that these techniques and procedures need to align with all the other layers of the research onion – i.e., research philosophy, research approaches, research strategy, choices, and time horizon.
For example, you if you’re adopting a deductive, quantitative research approach, it’s unlikely that you’ll use interviews to collect your data, as you’ll want high-volume, numerical data (which surveys are far better suited to). So, you need to ensure that the decisions at each layer of your onion align with the rest, and most importantly, that they align with your research aims and objectives.
The research onion details the many interrelated choices you’ll need to make when you’re crafting your research methodology. These include:
