Establishing the Object and the Logic of Inquiry in Qualitative Sociological Research

Introduction

What are you studying? Or more precisely, what are you aiming to understand through this study? These are questions often posed when evaluating or commenting on sociological research. Answering these questions require that we have established a clear object of our research. Social phenomena are inherently complex and multifaceted, and frequently, the object of study – the phenomenon under investigation – is not clearly defined. We discuss in this article how a lack of clarity about the object under investigation can be problematic, both for the research process and for the evaluation of a study and its contribution.

Establishing the object of study is a topic that is rarely addressed in the methodological literature. Leifusrud & Sohlberg (2021) contend that “how the ‘initial’ research object is identified and constructed (…) is a theme seldom problematised since the choice of research objects is often taken for granted and habitually made” (pp. 2-3). Our aim in this article is to illustrate the importance of establishing the object of study and how doing so enhances our ability to assess the sociological value of empirical research. We discuss methodological strategies for examining and establishing the object of a study, and the importance of the object for the entire research process. To illustrate the importance of a clearly defined object of study and its implications for the entire research process, we focus on the phenomenon of voluntarism – specifically, the role of non-profit organizations and volunteers in delivering social welfare.

The choice of illustrative example is partly motivated by what Milligan and Conradson (2006) term the voluntary turn in European welfare policy. Although the prevalence of volunteering is far lower in Europe (29%) than, for example, in Africa (61%) (Manda et al., 2025), expectations are rising. Amid growing concerns about the sustainability of welfare states, many countries are increasingly turning to voluntarism as an alternative to both public and market-based welfare provision (Askheim & Askheim, 2024; Dowling, 2024; Milligan & Conradson, 2006; Wolch, 2006). Since the 1980s and 1990s, New Public Management reforms have led to the marketization of public services across European welfare states, a trend further intensified by the 2008 financial crisis (Crespy, 2016). However, austerity measures and the rise of powerful market actors extracting significant profits from tax-funded welfare services (Kuusela & Tarkiainen, 2025; Pelling, 2025) have prompted a renewed interest in community-based and voluntary sector solutions (Gibbon & Hilber, 2022; van Dyk, 2021). This shift has also spurred increased scholarly attention on voluntarism.

Voluntarism is, however, a multifaceted phenomenon. As with any multifaceted sociological phenomenon, this raises the question of how we can study it. What exactly constitutes voluntarism, and is it a sufficiently coherent phenomenon to serve as the object of empirical inquiry? We argue that voluntarism is perhaps better understood as a topic. By topic we mean an issue or a problem that is too extensive or complex to serve as a research object of a single study. How, then, can voluntarism be rendered accessible to sociological research, and how can voluntarism be represented as a researchable object in a way that is both politically relevant and epistemologically sound?

We propose a line of inquiry and a set of methodological prerequisites for establishing a researchable object and identifying appropriate data to represent it. Our discussion evolves around five critical phases of the research process, which also structure the article: (1) establishing the object, (2) bounding the object, (3) identifying instruments of observation, (4) establishing empirical facts, and (5) extending the interpretation. Our aim with this line of inquiry is to bring more attention to the methodological importance of the research object and encourage researchers which study sociological phenomena utilizing qualitative data to think more critically about their object of study. The article concludes with a general discussion of methodological criteria for conducting relevant and coherent sociological research.

Voluntarism: a Complex Sociological Phenomenon

The selection of voluntarism as an example in this article is not only due to its significant political – and consequently research – interest. Voluntarism serves as a useful illustrative case of a sociological phenomenon because it is often taken for granted, despite being conceptually complex and difficult to define. As David Horton Smith noted over fifty years ago (1972, p. 3), “Voluntarism represents a category of human activity that is so varied it defies adequate description.”

Voluntarism is known by many names. It is often described as the third sector, alongside the public and private sectors (the market). It is labelled non-governmental and non-profit. Yet it is unclear what actually defines or constitutes voluntarism. Sometimes, voluntarism is used interchangeably with volunteerism – the act of engaging in activities voluntarily, without coercion or financial compensation (Ellis, 2026). However, volunteerism also occurs within public and private sectors, and organizations in the voluntary sector, such as NGOs, often rely heavily on paid staff. Moreover, many equate the voluntary sector with civil society (van Dyk, 2021), which is sometimes understood to include also the private sector, while some equate voluntarism with community (Fenger et al., 2023), which points more to informal organization and a distinctive part of civil society.

Voluntarism not only evades a universally agreed definition, but its boundaries with the public and private sectors are also becoming increasingly blurred. Fenger et al. (2023) even argue for a new conceptualization of voluntarism. The rise of new public governance has opened space for greater participation by both non-profit and commercial actors in public governance. Private companies now support or develop volunteer initiatives as part of their corporate social responsibility, and various forms of social entrepreneurship are emerging that pursue public goals using business methods. These changes contribute to the blurring of boundaries between sectors and are evident also in Norway (Eimhjellen et al., 2018; Vannebo & Grande, 2018). This blurring of boundaries is also linked to changes in forms of volunteering. While organized and membership-based volunteering is in decline, we now see the emergence of more flexible, and ad hoc-based forms of volunteering, partly driven by digital and social media and platform organization (Eimhjellen et al., 2018; Fenger et al., 2023; Hansen & Slagsvold, 2020). We also see a reemergence of food banks, soup kitchens, charity shops, and seasonal fundraising – what Kessl & Schoneville (2021) refer to as the “new charity economy.” Finally, there is a tendency to view voluntarism and voluntary engagement as a fairly homogeneous sphere, ideologically speaking. However, in a time of increasing polarization, we also find voluntary initiatives and civil movements with highly conflicting goals and values, as Ayres et al. (2024) warn from Latin America and the Caribbean.

These inherent and evolving complexities provide the rationale for using voluntarism as an example in our broader discussion on how to define and establish the object of a study, and the methodological implications of failing to establish a clear object of a study. We depart from the premise that our knowledge interest is in concerns about the sustainability of the welfare state and the political hope invested in voluntarism as a preferred solution to this challenge. In this context, some politicians may seek knowledge about voluntarism’s capacity and readiness to meet political expectations, while others may be more interested in the potential consequences for the welfare state of shifting greater responsibility onto voluntary efforts.

We argue, however, that voluntarism, in itself, does not constitute a researchable object, and that its multifaceted nature makes it impossible to observe in its entirety. Voluntarism is better understood as a topic that entails several possible research objects. This insight points our attention to a more general discussion on the importance of establishing a clearly defined object for empirical study. We argue that to study multifaceted topics such as voluntarism we first need to ask ourselves; What aspects of voluntarism can we observe, and study empirically, in a way that will provide knowledge relevant to what we seek to understand?

Establishing the Object

So, what requirements must be placed on an object for a sociological study? We will here take the reader through a line of argument, to make clear what we mean by defining a researchable object that can lay the foundation for a study that will generate knowledge relevant to the knowledge interest stated above.

Robert Merton, 1987, (p. 1) claims that establishing the phenomenon first “involves the universally accepted doctrine that phenomena should be shown to exist or occur before explaining why they exist or how they come to be”. In other words, a phenomenon needs to exist or occur for us to research it empirically. It is, however, not difficult to find historical examples of research on objects we today consider non-existent. Parts of research on moral deficiency and race in the late 19th and early 20th centuries are good examples of such inquiries. Even today we are confronted with phenomena whose existence can be quite contentious, phenomena that are so politicized or normatively defined that it is uncertain to what extent they can be said to exist and lend themselves to empirical investigation. Examples of such phenomena include co-creation and sustainability, which are concepts that refer to political ambitions rather than concrete empirical phenomena, but often are subject of investigation in research on governance and transformation of the public sphere.

Our main point here, following Merton (1987) is that to be made an object of study, a phenomenon must be shown to exist or occur in a way that can be observed, that is, through data that can represent the phenomenon. To return to our example, while voluntarism can be said to exist, in order for us to research it we still need to define how it can be observed. As discussed, voluntarism is a rather unclear phenomenon, which exists in different forms and is often described by different conceptions and definitions, not all of which are easily made accessible for empirical investigation. Establishing the research object therefore requires thorough work, which according to Bourdieu (1992), is by “no doubt the most crucial research operation”, yet the most ignored (p. 224).

To illustrate how a research object is established in sociology, Swedberg (2021) makes a distinction between observans (the process of observation) and observandum (that which is to be observed). This distinction highlights a crucial aspect of defining the research object in a study. While it might seem that identifying what is to be observed (observandum) precedes the process of observation (observans), the reality is more complex. No social phenomenon that captures the researcher’s attention is naturally given. Social phenomena are all laden with layers of preconstructions, be it common-sense, political, or theoretical, as seen in the case of voluntarism. Defining what is to be observed therefore requires a thorough process of observation, where one examines and deconstructs the initial topic or problem at hand, something which Bourdieu (1992) describes as “a break with common sense.” Bourdieu asserts that the preconstructed ” [i]s everywhere. The sociologist is literally beleaguered by it, as everybody else is” (p. 235). Therefore, constructing an object of study requires breaking free from this beleaguering to create an object that the researcher can take responsibility for. It involves active observation and reflexive scrutiny of the phenomenon in question to determine not only what is to be observed (the object) but also how it can be observed (the empirical design of a study and the data which represent the object)—to bring forth the knowledge we are seeking.

Merton’s doctrine that the phenomenon must be shown to exist or occur should also be understood in relation to a social world, where we relate to phenomena based on the meanings we attach to them (Blumer, 1986). Reed (2011) makes an interesting distinction between minimal and maximal interpretations, pointing our attention to that establishing the phenomenon to be observed is not merely a process of selection, but also a process which involves interpretation. For instance, when observing children at play, many observers can easily agree that this is what is being observed. We will thus accept it as a fact, even if it is an interpretation, acceptably a ‘minimal’ interpretation. Human behaviour must be interpreted to make sense, and it must make sense to become the object of a sociological study. Interpretation is thus integral to the process of establishing an object. To be able to say what children play, why they play the way they do, and the deeper meaning of the form of play we observe requires a more in-depth interpretation, moving towards what Reed refers to as a maximal interpretation, which we arrive at through dialogue with theory.

A phenomenon appropriate for sociological study must therefore exist or occur in a way that can be observed, and it must make sense, yet be constructed independently of our common-sense conceptions. But there is more to this. Not all phenomena lend themselves to sociological study. For example, while we can observe a single entity – such as a voluntary organization or a specific project – we must ask: what constitutes the sociological phenomenon we seek to understand? A single organization does not, in itself, qualify as such. Rather, the phenomenon may be the organization’s similarity to other entities of the same type, that which makes it an organization or a particular kind of organization. Alternatively, the phenomenon may be a recurring feature found in certain types of organizations, making these organizations the sites for investigating it. In essence, the sociological phenomenon is something that repeats; a pattern of behaviour that can be detected across multiple entities.

Swedberg (2021) argues that, for an analysis to be sociological, “the research object needs to be selected out and constituted from a sociological perspective” (p. 15). He admits that there is no universal agreement on what procedures this entails, but in “modern sociology “, he says, “it is often argued that what should be analysed and explained are patterns of social behaviour” (p. 13, our cursive). Patterns of people’s behaviour, such as the habitual forming of organisations, reflect that social behaviour is shaped through social interaction within social and cultural contexts. Analysing and explaining these patterns therefore requires a contextual analysis of the interactions and processes that produce them. As we have shown, the landscape of voluntarism oriented toward welfare provision is in transition, which reflect changes in patterns of social behaviour. To analyse and explain these patterns, we need to focus on how and why they emerge. This, in turn, places additional demands on the research object, namely that it can be contextually and historically situated, and that we remain attentive to the relations and processes which are integral to understanding the emergence and manifestations of the object.

This line of argument is in line with Desmond (2014), who advocates for a relational and processual approach to constructing objects for social research. He criticizes what he refers to as substantialist perspectives for imposing “static and atomistic categories onto a world made up of bunches of intertwining interconnections” (p. 551), and quotes Emirbayer (1997), stating that “ [actors] are inseparable from the transactional contexts within which they are embedded” (p. 551).

Returning to our example of voluntarism we see the importance of these insights. The landscape of voluntarism is historically produced and shaped within specific social and cultural contexts. Rather than choosing an ahistorical or substantialist perspective and examining patterns as static categories, we should instead focus on establishing an object which encompasses the relationships and processes through which these patterns are produced within a wider social context. To study process is to study relationships and vice versa. So, how can we construct an object of study that meet these criteria and has the potential to extend our knowledge about voluntarism in a way that is both epistemologically valid and politically relevant? The first step is to determine which aspects of the topic of voluntarism are most important to investigate, and which empirical data could represent these. As suggested earlier, politicians could be interested in examining either (a) the capacity and readiness of voluntarism to meet political expectations, or (b) the potential consequences for the welfare state of shifting greater responsibility to volunteering. Both questions carry significant political relevance.

The first option, however, requires extrapolation from the current situation. Statisticians might estimate future capacity based on historical trends, but such estimates would be highly uncertain and would not constitute an epistemologically sound sociological study. In other words, it is not possible to establish a sociological observandum that provides valid knowledge about the future. While it may be feasible to construct an object of study that sheds light on the realism of political expectations, the second approach seems more promising. Rather than assessing the future capacity of voluntarism to relieve the public sector of welfare responsibilities, it is more sociologically interesting and empirically feasible to examine the consequences of shifting greater responsibility to voluntarism. Voluntarism is already undergoing change, and these changes can be observed. Understanding what drives these changes and how they affect the welfare state can also provide crucial and relevant knowledge for politicians to evaluate the transformations that have taken place and whether these are politically desirable.

While the transformations we are interested in exploring occur in other countries, we will here focus on Norway. We therefore propose the following research question which allows us to define a concrete object of study:

What is driving the transformation of voluntarism in Norway and how do these changes relate to changes in the welfare state?

Our object of study, then, is the processes of change which are involved in the transformation of the landscape of voluntarism in Norway. To answer the research question requires that we observe the relationships between relevant actors in these processes, and how these relations evolve over time. The research question is formulated with the possibility that the welfare state itself is a driving force in these transformations, and that the influence of actors is bidirectional and that changes on either side potentially could reinforce one another. However, to understand what drives changes, we must also observe processes and relationships that contribute to stability in the landscape.

Bounding

After having constructed an object for our study and clearly defined a research question we pose to it, we need to establish the context for our study, which includes defining the scope of observation. We can refer to this as bounding the object under observation (Hansen, 2022). Bounding includes establishing the time period as well as the space (landscape or field) we contextualize this object within. If we are studying processes of change, we also need to specify the kinds of actors and relations that are embedded in these. Bounding the object in time and space also defines the boundaries between the object and the context of a study. The time outside the delimited period constitutes the historical context, while the kinds of actors and processes other than those specified constitute the social context.

If we choose the processes transforming the landscape of voluntarism as our object of study, how should we go about bounding our object? First, in terms of time period, the study requires a starting point, which allows us to determine the temporal and contextual foundations for an empirical investigation of the phenomenon under study. Foucault (1972) argues that when establishing (or in formation of) the object of study, we first need to establish a chronological break in history, a temporal break, which marks a change in practices or ideas which can be observed and represented through data. Foucault refers to such breaks as chronological in that they mark a discontinuity. ¹ Identifying discontinuities helps us to see the limits of a process and significant transformations which mark changes in relations or processes.

Establishing a chronological break does not only allow us to justify a starting point for our analysis. It also enables comparison in that we can establish a domain and a period from which transformations of ideas or practices can be observed historically. In our example, we should try to identify the time we first begin to observe changes in forms of voluntarism in Norway, when these changes gain momentum and explicitly create openings for, encourage or favour new forms of organization or operation of voluntarism.

Second, bounding also includes defining the kinds of actors and relations we seek to investigate. We have already argued that studying social processes is studying relationships, and vice versa. Social change can only be understood through the relationships that produce them. If we are to study the relations embedded in voluntarism it makes sense to look for transactions or points of contact between actors. Which actors are involved in volunteer work at various points in time, when do they appear on the scene, and what is their role in relation to other actors? We would need to pay particular attention to the production of coordinated action (Desmond, 2014), seeking to map the interfaces between public, commercial and voluntary actors to understand how networks between actors are formed and which factors contribute to instigate, maintain, change or break relations between actors. We would also be interested in mapping the various forms such networks take, and instances of hybridization of networks or activities, that is, where features of different networks or activities combine to form new initiatives. While these core processes define the object of our study, it is equally important to recognize other relationships and dynamics that may influence it, such as political developments and factors shaping the scope of voluntary engagement. These considerations form part of the broader context necessary for understanding what drives the transformation of voluntarism.

In summary, bounding the object consists of establishing a chronological break to define a temporal starting point for a study, and specifying the types of actors and relationships we need to investigate. Bounding helps us define the boundaries between object and context, and the scope of data needed to represent the object. Yet, even if the object is carefully bounded from the outset, in the process of analysis we should be open to considering alternative points of departure which can extend and improve our explanations of the object under study (Hansen, 2022).

Identifying Our Instruments of Observation

Once we have established the context for our study and defined the scope of observation for our object (the observandum), we need to prepare for the process of observation (observans) by identifying the instruments by which we observe. What data can speak about the object, and how can they best be generated?

In a qualitative study, various forms of data can be utilized to explore the processes and relationships involved in transforming the landscape of voluntarism. Accounting for the historical dimension requires retrospective data covering the defined time period, and ideally prospective data as well, by following certain processes as they unfold over time. As for relationships, ethnographic observations allow us to witness encounters between key actors. For example, observations might include meetings and negotiations where relationships and collaborations between actors are discussed and formed. Such processes can be tracked over time through ethnographic field observations. Observations of volunteer events can also provide rich data on the organization and different representations of voluntarism.

Documents are another valuable source of qualitative data and are often essential for understanding change over time. Documentation is a fundamental aspect of organized human activity, and materials such as archives, meeting minutes, newsletters, correspondence, and media reports can offer a rich and detailed account of organizational developments and relationships between actors. These sources provide valuable insights into the background, processes, and relational dynamics associated with changes and new initiatives. Documents are also the only source of timely data on changes over time, unless a longitudinal study is conducted.

Documents are also an important source of concurrent development. Not all relationships can be observed directly, but may be inferred through written communication, or identified through how different actors and their relationships are represented in various contemporary texts. Moreover, documents not only report on what happens but can also instigate relationships between actors by proclaiming that networks or new constellations should form.

Interviews with representatives of the involved actors can also serve as a crucial source of data. They offer insights into how actors present themselves, articulate their goals and values, assess opportunities and constraints, and refer to other actors. Additionally, interviews can reveal relationships or connections to other stakeholders that might not be captured through other sources. At their best, interviewees can act as their own ethnographers (Mol, 2002).

When combined with other data sources, interviews provide a comprehensive view of how various actors are positioned within the landscape of voluntarism, highlighting conflicts and struggles, and illustrating how relationships have evolved over time. While retrospective interviews can be uncertain, their integration with textual sources can enhance our understanding of historical developments. Other qualitative data, such as pictures or videos, can also be valuable in identifying relationships between actors and significant events.

Quantitative data can also inform a qualitative study and speak to shifting political priorities and changes in relationships between actors. Quantitative data, such as budgets, indicate how much resources different actors have available, how resources are divided between actors, and how the distribution of resources has changed over time. Quantitative data indicating membership in organizations and the number and types of people engaged in different volunteer initiatives can also be valuable. Such data tell us about transformations in the field and whether and when initiatives develop to form more flexible forms of volunteer engagement.

Using multiple data sources can provide a more reliable picture of relationships over time by allowing for triangulation. All these types of data can point in different ways to how and why patterns change over time. We have to carefully select the data and instruments needed to speak about our object of study, and argue for our selection based on how they represent the object. In our example, we need to focus particularly on the processes occurring within and between existing and emerging voluntary actors – including social entrepreneurs and hybrid organizational forms – and on data which allow us to depict their relationships with public and commercial actors. What new actors are entering the field, and what changes in organizational forms are emerging? How are these changes perceived and articulated, both publicly (in writing) and orally (in interviews)?

Establishing the Facts

Having defined the duration and scope of the ‘landscape’ under investigation and established the instruments for observing the object – specifying the types of data that can inform about the object and how they can be generated – we are prepared to establish the facts about the object. Identifying the instruments of observation and establishing the facts are somewhat intertwined, and these two phases may be hard to distinguish from one another. It is, nevertheless, important to recognise how a clearly defined object, and careful evaluation of how it can best be observed, guide the choice of observational instrument - ultimately determining the facts we produce about an object.

Some may feel uneasy discussing facts, fearing association with a realist or positivist stance. However, in this context (and in the social sciences in general), facts refer to observations as they make sense to the observer. The observer interprets what is observed. Just as the object of study is a “minimal interpretation” (Reed, 2011), so too are the facts about it. Researchers engage in minimal interpretations as they describe observed phenomena as they are understood, but without attributing to them a deeper or theoretical meaning.

What would establishing the facts entail in our example? It would involve recording and describing, as thoroughly and accurately as possible, everything that can be interpreted as relationships, and ruptures and changes in such relationships between the selected actors over the defined time period. It will also involve concrete information about circumstances that may have influenced such relationships and changes therein – including contextual information that enriches the factual picture that is drawn of the changes in the landscape of voluntarism over time.

Relationships should be understood broadly, and may include anything from direct collaboration between actors, alliances and agreements, to actors being part of the same network or association. Connections can also include ways in which actors refer to each other, verbally or in writing. Anything that could constitute a connection between two or more relevant actors that could potentially influence their position or dispositions is relevant. The connections should be described in sufficient detail to be later analysed in relation to other connections. Even simple references to, or comments about, other actors in someone’s interview or written sources can hold significance. A comprehensive overview of the frequency and various nature of relationships, how relationships have evolved over the defined time period, and how they contribute to procedural change would represent the facts about what has transpired between the actors during the period.

When dealing with large and complex data, accounting for the facts can be challenging and may require organization and collation to create a clear picture. Some degree of categorization or thematic organization – an “arrangement of evidential signs” (Reed, 2011) – may be necessary to determine factual patterns. Quantifiable data will typically be presented in tables or graphs. One should present the facts about the object in a manner that is as universally acceptable as possible, before moving on to more abstract or theoretical interpretation. Clarity about the facts – what is to be analysed – is what allows others to assess the validity of our analysis.

We recognize that even establishing the facts requires a degree of interpretation to make sense. For example, determining the nature of a relationship requires both interpretation and critical reflection to avoid bias while maintaining a focus on identifying the factual. One should explain the criteria for how the nature of relationships between actors are categorized and show examples of how these criteria are met to be transparent and to gain the greatest possible agreement on the facts.

Some might think of establishing facts as a somewhat rigid approach. Our point here has been to illustrate that establishing the facts involve a set of critical choices which have implications for the study at large, and that the critical inquiries involved in the various phases of the research process are interconnected and interdependent. In our example, the factual account of the scope and nature of the relationships between the actors over time should, taken together, and with sufficient context, provide a basis for analysing the dynamics of the processes transforming the landscape of voluntarism. However, a full-fledged and comprehensive analysis requires a more extended interpretation.

Extending the Interpretation

Interpretation is about making sense of a situation, a text, or a pattern, etc. We have already established that a phenomenon must make sense to constitute an object for sociological study. This means that interpretation occurs at the stage of establishing the object and continues when establishing the facts about the object, as discussed above. Analysing these facts involves expanding and deepening this interpretation. We ask ourselves what our observations mean, why relationships are the way they are, what processes we observe, and why they exhibit certain patterns. Extending the interpretation is a reflexive and creative process involving abstraction and theorizing, but it may also require further organization and processing of the data to facilitate a deeper and theoretically informed analysis of the facts.

Depending on the type of interpretation envisioned, more thorough or sophisticated sorting and organization of data may be required to enable the retrieval of sequences and patterns, facilitating an active dialogue with the data. Imagine sitting with your lengthy transcribed interviews, what Kvale (1996) refers to as the “1,000-Pages Question.” Attempting to engage with a stack of paper is futile; it must be ‘opened up’ and made accessible to allow for dialogue with the data. Making data accessible could for example involve coding, segmenting, and summarizing the data, or creating sequences of data appropriate for analysis. There is no right way to make the data accessible for analysis. Rather, the character of the data, and the analytical approach we choose for the study determines the best approach. Although this practical work is essential and may be described as part of the analysis, it is important to recognize that this work remains at the factual level. It refers to what the data has shown to exist or occur and remains true to the depictions of the object as they appear in our data. Organizing and resorting these facts is about facilitating the best possible dialogue with the data, which in turn enables a deeper interpretation of one’s observations of the object.

We often reach for theories and concepts to build extended interpretations. Moving from minimal interpretations to what Reed (2011) refers to as maximal interpretations begins with asking questions about (or to) the evidential. It involves entering into a dialogue with the material, which requires asking good questions and organizing the data in a way that enables answers. Deepening or extending the interpretation, or moving towards a maximal interpretation, also involves activating a more abstract vocabulary and attempting to make connections between evidence and theory. This requires that we engage in a creative and highly reflexive dialogue with both data and theoretical ideas. For example, how social change is perceived, depends on what theories of change we use (Tawodzera et al., 2022). While theories and concepts are integral to analysis, they also are invoked in the construction of the object, as they often direct our attention to what to look for and where to look.

Returning to our example, we can begin by asking questions about the possible dynamics at play within the space of relations bounded for the study. How do relations of power play out in the landscape of voluntarism? What are patterns of alliances and dependencies? How has the distribution of responsibilities and tasks between actors changed over time? How do the dynamics affect the interfaces between the state and non-governmental actors involved in the production and delivery of welfare?

These are all examples of questions that might be relevant to ask to analyse the material and to move towards a more comprehensive answer to the research question. How can we then extend our interpretation of the processes and relationships we have observed to explain what drives them and how they relate to changes in the welfare state? First, this requires rich contextual information, particularly about political measures and developments, not only within the defined period but also further back in time to capture the historical context. Such information can reveal whether changes in the voluntary landscape are responses to political changes or manoeuvres, or vice versa. It also requires dialogue with other researchers’ interpretations of similar processes and engagement with relevant theory. In our example, choosing one single, overarching theory would probably be unfortunate, as it risks limiting rather than supporting the analysis. Instead, we should seek for theoretical contributions or concepts that emerge through a critical and creative dialogue with the data, a dialogue that challenges dominant conceptions and categorizations in the field, including distinctions between voluntary, public, and commercial actors, as well as traditional understandings of the welfare state. We need to open the possibility of uncovering deeper processes, alternative dynamics, or cultural shifts, while remaining firmly grounded in the empirical facts about the object. Our ultimate aim should be a holistic understanding of why the landscape of voluntarism develops or transforms the way it does and how these transformations relate to and impact changes in the welfare state. An additional aim could be to contribute to the development of theory, for example, by identifying weaknesses and/or possible improvements in the theories we use.

Concluding Discussion

In this article, we have argued for the importance of a clearly constructed object in qualitative social research. The research process could have been described somewhat differently. The phases could have been more or different, thereby revealing other important aspects of the research process. The illustrative example used could also have been different, thus highlighting other challenges and opportunities. The organization of the article must be understood in light of the purpose of conveying the importance of establishing the research object and its significance for the various stages of the research process. The selected phases highlight some of the most important choices in the research process and how they are interconnected, and that what connects them all is the object being investigated.

We argue that the importance of a clear research object is integral to the logic of an empirical study. When designing a study, we need to know which object is to be observed, and that it is this object that is actually observed through the data we plan to generate. This means ensuring that the data we generate can represent the object in a way that is suitable for answering the questions posed about it. The validity of a theoretical analysis of empirical data rests on clarity in the actual patterns or features (facts) to be analysed. The audience needs to know what facts form the basis for our analysis. These facts rest on thorough and relevant observations, which in turn presuppose clarity about what is to be observed, that is, a clearly defined and bounded object of study. Conversely, if the object is unclear, it cannot be bound. If the object cannot be bound, it becomes unclear what one should observe and difficult to identify which data can represent the object. If what is to be observed is unclear, we cannot establish credible facts about the object, which jeopardizes the possibility of conducting a sound theoretical analysis.

While others have discussed the importance of establishing researchable objects in sociology, our primary contribution here is to map out its implications for the entire research process, as illustrated in the figure below (Figure 1). We have argued that distinguishing between clearly defined research objects and topics, or research problems, enables us to more critically assess how empirical findings contribute to sociological knowledge. By discussing methodological strategies for examining and establishing the object of a study, our aim has been to emphasize the importance of the research object for conducting relevant and coherent sociological research.OPEN IN VIEWER

Some may think these steps seem a bit overstructured and rigid, since in practice the research process is often quite messy. We admit that a research process is a lot about trial and error, and that you often have to make changes along the way. However, our experience is that part of the mess that for example many doctoral students experience is due to insufficient work being put into the construction of the object of study. This again may be related to the fact that the significance of this work has not been thematized to a great extent in methodological literature. It is well established that the research process is an ongoing reflexive process in which the researcher must make a series of choices. We argue that a clear and well-defined research object makes this process more intuitive and the choices easier to make and justify.

Establishing the object is in itself a process, requiring both critical (self)reflection and theoretical work. It usually starts with a topic that catches the researcher’s interest, either because it poses a problem or because the researcher finds it theoretically interesting. The researcher may also have been commissioned by someone to study a given topic or problem. But as we have argued, topics and problems are usually too broad or complex to serve as an object of study. They cannot be properly observed. Instead, what we encourage researchers to do is to engage in the critical work of constructing an object that can be observed and studied and that constitutes an important aspect of the problem or topic that motivated the study in the first place.

Some may argue that the object can be clarified during the research process. We have shown that constructing the object of study requires a process of observation to gain clarity about what to observe. In ethnographic studies this may require observations in the field. The logic, however, is the same; data must be able to represent the object that one ends up studying.

Obtaining the necessary data can sometimes be challenging. Archives may be incomplete or inaccessible, and interviewees may steer the conversation away from the intended topics. In such cases, researchers must consider which objects can be meaningfully represented by the data that is available. This approach may preserve the study’s validity, but it often falls short of delivering the insights originally sought.

The same applies to the credibility of the analysis. Credibility is arguably the most fundamental criterion in both qualitative and quantitative research. Credibility is conditioned by transparency and by others being able to scrutinize the analysis and design of the study. However, qualitative studies cannot be replicated in the same way as quantitative research or controlled experiments. This, we believe, makes the transparency and rigor involved in establishing the research object even more critical in qualitative studies. The work involved with establishing a clear object, and trust in that the data subject to analysis and interpretation actually represent this object, allows peers and others to critically examine our interpretations and our analysis. We believe that placing greater emphasis on the research object, and its critical role in analysis, will ultimately enhance both the quality and credibility of qualitative research. Consequently, the role of the research object should be emphasized in methodological literature and teaching practices.

Let us conclude by highlighting two key insights from this article. On the one hand, it is essential to acknowledge that no single study can provide comprehensive answers to a broad and complex problem or topic. Treating the entire topic or problem as the research object – or taking the object for granted –can obscure the research process and analysis, ultimately undermining both the credibility and the contribution of the study. On the other hand, by accepting this limitation and committing to the rigorous work of constructing a clear and defensible research object, researchers can generate meaningful and credible knowledge about a specific aspect of the broader issue that initially motivated the inquiry. The focused approach we have outlined in this article not only facilitates informed discussion of the topic in light of new insights brought forward by a study, but also opens the door to identifying what Merton (1987) refers to as “specified ignorance”—that is, the recognition of “what is not yet known but must be known to lay the foundation for even more knowledge” (p. 1).