Solving the Chaos of Qualitative Research Varieties

Correspondence: papers@team.qeios.com — Qeios will forward to the authors

Introduction

I will probably not surprise the reader by saying that thoroughly understanding the plethora of qualitative research varieties is a complicated methodological endeavor. To ease the task, many authors have attempted to classify them. Numerous articles and textbooks that describe the main varieties have already been published and are helping students as well as beginning researchers to navigate this treacherous landscape. Or are they, really?

For some reason, the presented ‘main’ varieties are rarely the same ones, and their classifications are very much inconsistent. For example, Creswell & Poth^[1] describe five qualitative research ‘approaches’, which are narrative research, phenomenology, grounded theory, ethnography, and case study. Merriam & Tisdell^[2] talk about six common qualitative research ‘designs’ or ‘approaches’, namely basic qualitative research, phenomenology, grounded theory, ethnography, narrative analysis, and qualitative case study; while Žydžiūnaitė & Sabaliauskas^[3] identify seven qualitative research ‘methodologies’: qualitative content analysis, phenomenology, phenomenography, thematic analysis, grounded theory, narrative analysis, and ethnography. But Patton^[4] outdoes them all by describing as many as 16 different ‘theoretical approaches’.

You may have noticed that authors use terms like methodology, approach, design, strategy, or even theoretical tradition to describe the same concepts. However, failing to draw clear distinctions between these terms leads to numerous misunderstandings and misapplications in empirical research.

The epiphany to write this article came when I was confronted with material that compared thematic analysis with phenomenology, case studies, action research, and other varieties. I simply could not accept that all these could ever fall into the same category. So, I set out to distinguish the different categories that would allow logically sorting the above-mentioned varieties. The outcome was a comprehensive framework.

Definitions of central concepts

Qualitative research

The endeavor of sorting the whole mess requires reconsidering the very starting point of it. We are so used to hearing the term ‘qualitative research’ that we do not question what makes a study qualitative anymore. What is the necessary element that must be qualitative so that we would call the research qualitative? Let’s say, it’s the data. Then, every research that involves qualitative data (interviews, transcriptions, documents) would have to be regarded as qualitative. But we must be mindful that any qualitative data can also be analyzed quantitatively. We can count the occurrences of a certain word or a phrase in a document, we can search for combinations, and lots of other things. These are the basic principles of quantitative content analysis.

The other way of thinking is that for us to call a study qualitative, it is the analysis that needs to be qualitative. Yes, I am aware that quantitative research has strict requirements regarding the usage of statistical methods. I am also aware that this path of thinking will not appeal to researchers specializing in quantitative research. From their perspective, analyzing quantitative data in a qualitative way is simply poor practice. Nevertheless, this is commonly practiced in several research traditions, including but not limited to action research.

In a broad sense, data is a captured ‘slice of the world’. Certain data makes certain knowledge accessible, but it is through the analysis that the knowledge is obtained. If we had the best data possible but chose a poor method of analysis, the claims about knowledge would be very limited. Since it is the knowledge obtained that is important and not the knowledge accessible, qualitative research must be any type of research where data is analyzed using qualitative methods.

Research design

In the following sections, I will describe terms that are often used interchangeably. Different authors assign different meanings to them and, therefore, may not agree with my definitions. The provided definitions are grounded in semantics, the established usage, and pragmatics. My aim is to establish clear distinctions and boundaries between the otherwise overlapping terms.

I define research design as the complete plan of the study, including structural, methodical, and practical considerations. Please note that ontological and epistemological considerations are not included in the designing process. These considerations are considered to be part of the theoretical framework of a study. Ontological and epistemological considerations may influence research design but are not part of it. Ethical considerations and the approval of an IRB committee are included in the practical considerations of a study.

Research strategy

By defining research design as an overarching term, other similar terms are seen as its parts. Yet, it may not be easy for the reader to accept this decision, since design is often associated with the practicalities of research. The reason for exalting research design and not research strategy is that structural and methodical considerations must come in the designing process and not prior to that. Looking at the research object through the pre-chosen structural or methodical lenses narrows the field of accessible knowledge. There are many researchers who are experts in a single methodology and its philosophy and therefore employ it repeatedly. This should not be viewed as a weakness; however, I would suggest researchers describing the considerations for choosing the research strategy and explicitly stating that the decision was ultimately made due to the methodological background of the researchers.

Therefore, research strategy is defined as structural and methodical considerations of a study.

Research approach

In my framework, research approach is the structural considerations of a study, captured in five dimensions: focus, orientation, relation to the object, researcher intrusiveness, and scale (for easier memorizing, I refer to them as FORIS). These dimensions will be explained in the next section. The term ‘structural considerations’ refers to the idea that a research approach of a certain study is constructed from chosen elements.

Research methods

From all the similar terms, research methods have the clearest semantic background; thus, they are defined as the applied replicable procedures in a study, mainly regarding data collection and data analysis. A more detailed explanation will follow later on.

Research techniques

Research techniques denote applied solutions in a study that do not exhibit the characteristics of procedural clarity and replicability. Techniques represent partially transferable decisions in gaining access to data sources, collecting data, analyzing data, and presenting the findings. We should distinguish techniques from the practical considerations of a study. Despite lacking procedural clarity and being irreplicable, techniques are transferable and intercontextual. They are helpful advice, the practice that has worked well in the past.

Research methodology

In contrast to research methods that correspond to specific procedures of a study, research methodology should be understood as the methodical, epistemological, ontological, ethical, or even axiological tradition regarding a specific research approach or method. It is not specifically related to the research that is performed. Methodology can be applied to a study, but it manifests through the approach, methods, techniques, and practical considerations.

Research framework

The research framework is the cultural, theoretical, methodological, or other context in which the study is positioned, and whose knowledge base it is aiming to broaden. We must be careful when applying this term in terms of methodology, ethics, or analysis. According to this definition, the key element is positioning in a certain context. Therefore, a methodological framework would imply positioning the study in the tradition of the chosen research approach, but it is not the term for the description of the strategic aspects of a study.

Constructing the approach

Just to remind the reader, the research strategy is divided into two parts in the proposed framework: research approach and research methods and techniques. I have presented in the introduction section that different authors identify different varieties of qualitative research. Most of these varieties fall into the category of research approaches. While different research approaches have certain similarities, some have many distinct features. Therefore, an attempt was made to distinguish all the necessary characteristics of every main qualitative research approach and classify them into five dimensions. The dimensions and corresponding elements (or choices) of research approaches can be seen in Table 1.

In this framework, the research approach of a particular study is a decided combination of research approach elements. Every qualitative study must have at least one research focus, at least one research orientation, at least one type of relation to the object, exactly one identified type of researcher intrusiveness, and exactly one type of scale. It is possible to explore the object employing several elements of the first three dimensions. However, I would suggest that this decision be reconsidered, as scope in research rarely signifies the quality of it. I am aware that most qualitative research does not address all the dimensions mentioned above. In some articles, the elements of the dimensions are clear from the context, and sometimes they are not. Personally, I have noticed that focus and orientation are the two dimensions that are unidentifiable most often.

Research focus does not imply that other aspects in the research field are not present when conducting a study; rather, it denotes where the attention of the researcher is. It is possible to focus on two or all three elements, but not at the same time. Rather, the attention must then be divided either by jumping between one and the other, or by exploring different elements consecutively. If a study is focused on meaning, it is answering questions like ‘what does a phenomenon mean for a subject?’. An interaction-focused study would explore relationships (including power relationships) or communication. However, in a process-focused study, the change of a phenomenon in relation to time is important.

Concerning research orientation, five elements have been identified relating to temporal considerations (Figure 2). Four out of the five have a single data collection timepoint. Cumulative orientation is the exception, where data is collected throughout the study. Exploratory orientation is based on variation in experience of study participants, while essence-oriented research attempts to capture the necessary and sufficient qualities of what makes a certain phenomenon. Evaluative orientation, despite collecting data at a single timepoint, is concerned about how the present moment came to be. On the other hand, predictive research aims at describing possible scenarios of future development. Research with cumulative orientation should be distinguished from process-focused research. Process in the dimension of focus does not necessarily require cumulative data collection (e.g., grounded theory approach).

Moreover, relation to the study object denotes the physical positioning of the researcher relative to the study object. It is self-explanatory that in self-centered research, the study object is directly related to the experiences of the researcher. Similarly, immersive research describes approaches where the researcher is closely related to the study object, but the object is nevertheless outside of the first-hand experience of the researcher (e.g., ethnography). The third element, however, requires a clear distinction between the concepts of subject, object, and setting. The dimension ‘relation to the study object’ should not be confused with positioning in relation to study subjects or the study setting. This way, distanced research represents the type of studies where the researcher has no relation to the study object, but not necessarily to the study subject. Therefore, collecting data through interviews, despite making an obvious relation between the researcher and the study subject, should be considered distanced research.

The dimension of researcher intrusiveness represents the active and deliberate manipulation of the study object. Low intrusiveness indicates that research is aiming to suppress the researcher impact (or observation impact) on the data. On the other hand, highly intrusive research follows the principle of understanding through change (e.g., action research or therapy). This dimension is not trying to address the researcher impact as enacted, but only how the object is approached; therefore, intrusiveness is not conceptualized in a continuum.

Lastly, scale is also an important consideration in constructing a research approach. Single-case studies involve only one research participant, situation, or setting. When multiple cases are merged together either by combining or juxtaposing them, I suggest calling them merged-cases studies. The extensive study, though rarely performed, is different from the merged study because it attempts to capture broad domains of information, sometimes employing quantitative research sampling techniques and drawing generalizing conclusions.

Choosing methods and techniques

It is not within the scope of this article to present the extensive list of methods and techniques of qualitative research. Partly because there can never be a definite end to such a list. However, it is important for the purposes of this article to present the different types of methods and techniques and provide several examples for easier understanding. I have identified four types of methods and techniques: for accessing data sources, for collecting data, for analyzing data, and for presenting the findings.

Since qualitative research is empirical in its nature, it always includes some kind of data. This data must be accessed by the researcher. It might be as simple as buying a certain book or scheduling an interview with a close friend. But most of the time, it is more difficult than that. In some cases, the required documents or research participants are not accessible at all due to ethical, legislative, or political reasons. As you may have noticed, gaining access to data sources is very study-specific and therefore cannot be described methodically. Therefore, we can only talk about techniques for gaining access to data sources, but not methods.

Having access to data sources, the researcher then collects the data. It really depends on what kind of data it is. If the data is pre-recorded (e.g., documents, videos), then collecting it is simple and requires neither methods nor techniques. Otherwise, the researcher must acknowledge that capturing and collecting data always prioritizes some aspects of the world while ignoring others. Since data majorly influences the study findings, it is preferable to collect them using pre-identified methods rather than using techniques. It does not mean that the methods for data collection must be pre-validated or pre-approved by the scientific community. Rather, they must be described with procedural clarity. Moreover, collecting the data includes choosing, acquiring, or creating the study instruments. For example, if the data is collected through observation, it must be pre-decided what aspects of the research setting the researcher will be focusing on. The same applies when the data is collected using self-reflection. Failing to pre-identify the aspects of the researcher’s attention is considered poor research practice.

An important question to consider is how unstructured interviews should be regarded. From what has been said before, the paramount idea is that the researcher should be prepared. Therefore, conducting an unstructured interview should always be justified. Additionally, in studies of cumulative orientation (e.g., action research), data collection methods and instruments may be created in the course of the study. But even cumulative research should mind identifying data collection methods during the reflective stages and not the ones of action.

Also worth considering is the often-misinterpreted idea of inductiveness necessitating unpreparedness. If the aim of the researcher is to avoid his or her pre-understandings distorting the data during the collection process (e.g., in phenomenological methodology), an even greater preparation of the researcher is required. The researcher must not only identify his research object and create or adopt a procedure for studying that object but also recognize his pre-understandings and decide how their bracketing will be attempted. Therefore, in this case, data collection methods encompass both the exact procedure for data collection and the bracketing of the researcher’s pre-understandings.

What concerns data analysis, there are many well-established methods and methodical traditions. To name a few: thematic analysis, qualitative content analysis, narrative analysis, discourse analysis, conversational analysis. Most of these methods are currently considered as families of related methods and can therefore be subdivided into several varieties. For example, in the family of thematic analysis, we have at least three family members: reflective thematic analysis^[5], codebook thematic analysis^[6], and coding reliability thematic analysis^[7]. The important thing to underline is that, as with data collection, we should prioritize methods over techniques. A researcher is always welcome to adapt a chosen method of data analysis to an extent he or she feels suitable, as long as procedural clarity is retained and any change to the established method is justified.

Finally, I will explore the techniques for presenting the findings. Personally, I find writing up the most interesting part of qualitative research. The reason is that the presentation of findings in a qualitative scholarly article can adopt many different forms and even allows artfully interpretative rhetoric^[8]. It is here that the creativity of the researcher can be expressed. Besides, the presentation of the study results may not even be in the form of a scientific text. We should consider oral presentations, poetry, fiction, theatre productions, or even presentations in practice as ways of presenting research results. But the framework that this article is offering is not concerned with the forms of presentation, but rather the techniques.

To a certain degree, techniques for presenting the findings are usually established in the tradition of the chosen data analysis method. It makes sense because data analysis methods provide certain types of outcomes. For example, thematic analysis results in themes, qualitative content analysis results in a categorization matrix^[9] or a coding frame^[10]. But even then, very rarely do we find the exact requirements of how such a presentation should look. This is why we could never call the practices of findings presentation methods.

Among the techniques, we often hear graphical-systemic description being praised. This technique not only allows the reader to quickly get the idea of what the article is about but also increases the probability that the article will be interpreted correctly. From the authors’ side, graphical-systemic description allows visually seeing the flaws in the data analysis and improving it before sending the article to the reviewers. The other techniques worth considering are tabular summaries, contrasting cases, embedding quotes in the description of the findings, employing metaphors, and many others.

Explaining the main varieties through the framework

Central to this framework is the idea that it is not the naming of a research design or approach that is important but the underlying essence of it. In other words, it is the structural characteristics that the study possesses that make it what it is. In this section, I will provide definitions of most of the qualitative research varieties identified by Creswell & Poth^[1], Merriam & Tisdell^[2], Patton^[4], and Žydžiūnaitė & Sabaliauskas^[3] through the relatively simple framework of this article.

Action research is a qualitative research approach distinguished by its cumulative research orientation, self-centered relation to the object, and high researcher intrusiveness. Even though most of the studies that employ this approach have a scale of one case, action research methodology does not abandon the idea that action research can be performed on a merged-case study scale. This is usually called participatory action research^[11]. Therefore, one-case and merged-case action research should be considered as different types of action research.

Phenomenology is a qualitative research approach distinguished by its focus on meaning and orientation towards essence. I am aware that by explaining phenomenology through this framework, I am essentially defaming it. It loses its mysteriousness and refined outlook. Yes, defining it in two fundamental dimensions does not do justice to the philosophical foundations of phenomenology. However, ignoring or not being aware of those philosophical traditions has become common practice in qualitative research. Some authors even explain phenomenology as simply being a research approach that focuses on experiences or meaning (e.g., Crocker et al.^[12]; McKay & Kirk^[13]). That being said, this definition of phenomenological research makes phenomenology easily comprehensible and accessible. It also acknowledges studies that do not provide rich epistemological and ontological considerations. By the way, this definition does not leave room for descriptive phenomenology as developed by Husserl. It is because in social sciences, description without interpretation is logically impossible. Through language, context, and researcher influence, social sciences are inherently interpretative.

Phenomenography is a qualitative research approach distinguished by its focus on meaning and exploratory orientation. Again, ‘exploratory’ here denotes an interest in variation.

Narrative research is a qualitative research approach distinguished by the focus on process, evaluative orientation, and distanced relation to the study object. Also specific to this approach is that the data is always analyzed using narrative analysis. Narrative research often assumes a constructivist epistemological and ontological stance, but not all researchers explore these questions. Therefore, the level of researcher intrusiveness varies depending on the study design.

Ethnography is a qualitative research approach distinguished by cumulative orientation, immersive relation to the object, and low researcher intrusiveness. Low researcher intrusiveness does not mean that the researcher does not affect the study setting, but rather that the aim of the research is to minimize this effect. Ethnography usually has a scale of one case, but not necessarily. Ethnographic research performed on a merged-case scale is usually called multi-site ethnography^[14].

Autoethnography is a qualitative research approach distinguished by cumulative orientation, self-centered relation to the object, and low researcher intrusiveness. Since there is a difference in necessary elements, autoethnography cannot be considered a sub-approach of ethnography. Autoethnography is the middle ground between ethnography and action research.

Grounded theory is a qualitative research approach distinguished by its focus on process. Every other element can differ depending on the purposes of the study. In terms of research orientation, grounded theory can be used to study variation, the essence of a particular process, or in a cumulative, evaluative, or predictive way. In addition, while grounded theory often delves into meaning and interaction, the central aspect is how the meaning and interaction came to be, that is, it is the process of meaning-making and interacting that is important. We must note that in practice, grounded theory is often used as a data analysis method, rather than an approach.

Thematic analysis, qualitative content analysis, discourse analysis, and narrative analysis are considered data analysis methods. Yes, they may have established traditions of related research approaches. And yes, some approaches may be more compatible with a certain data analysis method than other ones. But that does not put these methods in the category of an approach.

A case study is neither an approach nor a methodology, as is often presented by various authors. Rather, it is simply a characteristic of qualitative studies that describes their scale.

Hermeneutics is not considered a separate approach, rather ontological and epistemological considerations. Several methodologies (hermeneutic phenomenology, hermeneutic narrative research) place importance on these considerations, but they rarely deviate from the overarching ontological and epistemological considerations regarding qualitative research in general. While a study may take into account the ontological and epistemological position established in the tradition of a particular research variety, in many studies, this is not the case.

Examples of practical usage

The reader of this article may have been thinking that this article merely offers a novel viewpoint to a well-explored field; that it is yet another attempt to explain what many have already done. Each from their own perspective. If this is the case, then the reader may have sized up this framework too quickly.

Instead, this article is a call to the academic community. A call to reconsider its practice. Having criticized the current practice and offered a new perspective, there is yet another step to be taken. The step is to demonstrate examples of this framework in practice. This I will do by examining and translating the ‘methods’ sections of three methodologically different qualitative articles. The reader should not think that I am criticizing the articles I have chosen. Quite on the contrary, these are articles that I have identified as having well-described and good-quality designs. An attempt will be made to show that the framework of this article allows simplifying and clarifying these descriptions.

The first article employed a phenomenological approach and was published in Teaching and Teacher Education^[15]. It has a section named ‘Method’ with subsections ‘Design’, ‘Participants’, ‘Ethical considerations’, ‘Interviews’, and ‘Data analysis’. Applying the new framework, we would structure this section differently. To begin with, the section would be named ‘Design’. It would have four subsections: ‘Approach’, ‘Practical considerations’, ‘Data collection’, and ‘Data analysis’. In the article by Schmidt et al., it is stated that ‘This study is grounded in descriptive phenomenology’. This would translate to ‘The approach of this study is characterized by a focus on meaning and an orientation towards essence’ or put more simply: ‘The study aims to describe the essential components of meaning that the participants ascribed to X’. But what about the study being ‘descriptive’? I argue that it is not part of the approach, but rather of the chosen data analysis method. In this case – Giorgi’s descriptive phenomenological method^[16].

The next article I have chosen as an example has the title ‘The school restaurant: ethnographic reflections in researching children’s food space’ and was published in the International Journal of Qualitative Studies in Education^[17]. The author states that the ‘<…> study on Peartree Academy [a school restaurant] involved adopting an ethnographic case study <…>’. In other words, the study’s orientation was cumulative, the relation to the object was immersive, researcher intrusiveness was low, and the scale was a single case. As usual with ethnographic studies, data was collected using observations and interviews. In fact, observations have become so closely related to ethnography that many authors identify them as part of the research approach. However, according to the proposed framework, observation is a data collection method and cannot be considered a part of the research approach.

The third article applied grounded theory to study clinical nurses’ process of coping during COVID-19 and was published in the Journal of Clinical Nursing^[18]. The authors describe the study approach as follows:

Grounded theory is an inductive method of creating a theory from data^[19], where researchers foster critical reflection throughout the process by asking questions, seeking clarification, and actively listening to participants’ stories. Researchers also draw on personal experiences to facilitate theoretical development, as active agents in the research process^[19].

Please note that the authors described grounded theory here as a method, rather than as an approach. That does not do justice to the ideas set out in the referenced book by Corbin and Strauss, where grounded theory is regarded both as a method and an approach. However, this is a description that fits well with my framework. I would like to remind the reader that as an approach, grounded theory has only one necessary element – it is focused on process. In the article by Nowell et al., however, grounded theory is considered a data analysis method. One that results in a theory. Unfortunately, by describing grounded theory as a method, this article avoids the question of research approach entirely. I would suggest reporting this study as having a focus on process, a distanced relation to the object, low researcher intrusiveness, and a merged-cases scale.

Conclusions

This article is an attempt to identify the necessary and sufficient criteria for explaining the main qualitative research varieties and to sort them using one framework. It is an attempt to bring conceptual clarity to fundamental concepts, which often overlap. At the heart of this framework is the idea that the quality of qualitative research lies in the clear understanding of the study design, including all structural and methodical considerations, and not in the naming of the chosen methodology. I argue that research strategies are not chosen at all, but rather constructed according to the aims of research.

This framework separates ontological and epistemological considerations from the designing process altogether. It simply is not part of the design. It may be part of the methodological tradition, or it may be an integral part of a certain research approach, but the researcher needs not contemplate ontological and epistemological questions him- or herself. Such attempts are often unsuccessful and might even diminish the value of an empirical study.

Most of the qualitative research varieties fall into the category of research approaches and can be defined relatively simply using the five dimensions delineated in this article. In most cases, qualitative research varieties can be explained in two or three necessary dimensions, and varying the elements from additional dimensions results in sub-approaches.

There are, however, many questions left unanswered. Firstly, claims about knowledge that different research varieties report must be reviewed in the light of this framework. Moreover, since this framework allows constructing 270 one-element research approaches and as many as 9384 approaches, if we allow selecting several elements in the dimensions of focus, orientation, and relation to the study object, the question arises as to how many of them are functional. And how many novel approaches can be identified simply by combining different elements? Lastly, the historical context of how certain varieties gained recognition and whether certain forces had influence on their establishment also requires exploration.