The Use of Theory in Qualitative Research

What Is Theory, and Why Might Qualitative Researchers Sometimes Treat Existing Theory With Caution?

Although definitions of theory vary, there is general agreement that theory involves the expression of relationships between abstractions to arrive at understanding of a phenomenon, though theories may vary in scope, complexity, and level of abstraction (Anfara & Mertz, 2015; Maxwell, 2013). Theory is valued as the way in which academic understanding is differentiated from other forms of understanding, moving beyond the collation of facts and description or moralizing judgments (Wilson & Chaddha, 2010), and can also be a guide to action (Maxwell & Mittapalli, 2008). In some approaches to qualitative analysis (e.g., grounded theory; Glaser & Strauss, 1967), building theory that is grounded in data is prized as the goal of research. However, as Maxwell and Mittapalli (2008) noted, some kinds of theory sit uneasily with many approaches to qualitative research and would not commonly be part of the theorizing of qualitative researchers. This includes theory that decomposes social and psychological phenomena into discrete variables and aims for explanation of cause-and-effect relationships between these variables, in order to make generalizable predictions.

Most research methods textbooks point out that when we refer to qualitative research, we are usually indicating not just an interest in non-numerical data but a broad research paradigm, though one that includes some variation (e.g., Braun & Clarke, 2013). This paradigm differs from quantitative research in the underlying philosophical assumptions made about the nature of the world we are investigating (ontology), the kinds of knowledge that are both useful and possible and how we might gain this knowledge (epistemology), and hence the best ways to conduct research. In many disciplines, qualitative research emerged as a point of resistance to the use of methods from the natural sciences (positivism) in quantitative research for understanding psychological and social phenomena (Howitt, 2016). Following the early distinction of the sociologist Max Weber, qualitative researchers often take understanding rather than explanation as the goal of the human sciences (King & Brooks, 2017). Instead of assuming that the causes and effects of human actions and experiences can and should be studied objectively using the methods of the natural sciences (measurement, control of variables, standardized procedures), qualitative researchers often (though not always) take an interpretivist approach, being interested in understanding the meanings that people attach to their experiences and practices (Gray, 2018). The assumption is that these meanings will be varied and will develop differently in different contexts (Cresswell & Cresswell, 2018). Therefore, the psychological and social worlds being investigated are not assumed to be straightforwardly predictable and governed by general laws of cause and effect (Howitt, 2016). This position is often accompanied by a degree of relativism, so that analysis of qualitative data is viewed as an interpretation. Just as research participants may construct different meanings about a situation, so might the researcher construct one of many possible interpretations of the participants’ meaning making (Braun & Clarke, 2013). Qualitative methods have also largely arisen within naturalistic forms of inquiry, which often (though not always) aim for a holistic approach and are skeptical of theories that suggest that (a) individuals can be studied separately from the contexts in which they live and (b) phenomena can best be understood by isolating and measuring component parts, such as specific thoughts and beliefs (Howitt, 2016). Therefore, many researchers working within a qualitative paradigm would be unlikely to assume that the goal of research is to develop or test theories that can arrive at universal truths, capable of predicting the operation of discrete variables across a range of contexts. They would also be skeptical of the determinism implied in this type of endeavor, as it seems to ignore the notion of people as agents who create outcomes based on idiosyncratic meanings (Maxwell & Mittapalli, 2008).

To illustrate with an example, the theory of planned behavior (Ajzen, 2002) is a popular conceptual framework in quantitative research investigating predictors of breastfeeding. The theory decomposes these predictors into positive or negative attitudes toward a behavior (e.g., breastfeeding), beliefs about norms related to the behavior, and expectations of behavioral control and suggests that together they predict intentions, which predict the behavior. Therefore, the theory of planned behavior would conceptualize breastfeeding as an individual behavior that is a consequence of individual decision making, based on individual beliefs, though some of these beliefs might be about social norms. It is proposing “broad brush” causal relationships applicable to a wide range of behaviors across different contexts. Therefore, the theory does not lend itself to the more contextualized and relational understandings of human action that are often the aim of qualitative research (Horrocks & Johnson, 2014), for example, to understand the personal and cultural meanings that inform infant feeding and how these are shaped and negotiated in specific contexts. Therefore, it is no surprise that this theory is explored more often in quantitative than in qualitative breastfeeding research. In a study testing the relevance of the theory of planned behavior to breastfeeding, qualitative methods of data collection might be used initially to gather information for constructing a measure of attitudes to breastfeeding (e.g., Giles et al., 2007). However, direct verification of the theory would be obtained by measuring attitudes, expectancies, intentions, and behaviors and establishing statistical relationships between these.

The above discussion glosses over some important distinctions between qualitative approaches. It is not the case that qualitative research is never concerned with questions of causality. Some researchers using an approach called realist evaluation (Pawson & Tilley, 1997) aim to develop theory about causal mechanisms in healthcare interventions through qualitative data, but with a focus on complexity rather than linear cause and effect. The assumption is that causal mechanisms are likely to be multifaceted and difficult to observe and define and are experienced and understood differently by different stakeholders. In contrast to quantitative research, theory is developed to capture variability across contexts. For example, as part of a realist evaluation of breastfeeding support groups, Hoddinott, Britten, and Pill (2010) used qualitative data to develop an explanatory model that accounted for differences in successful implementation of groups and in breastfeeding outcomes. Similarly, researchers using grounded theory may build theories to examine complex causal processes. For example, Hunt and Thomson (2017) used this approach to develop understanding of why breastfeeding women often do not access peer support.

Despite some variation in qualitative researchers’ approaches to theory, there are clearly some kinds of theory that are unsuited to the aims of most qualitative research, even if the theory appears relevant to the topic being investigated. Qualitative research tends to ask questions about the nature of phenomena or how complex processes work and are perceived or how meanings are constructed (Braun & Clarke, 2013), and theory is likely to be useful to the extent that it sheds light on these questions. However, qualitative researchers sometimes exercise a more general caution about bringing any theoretical concepts to the research process, regardless of the nature of these, because of a concern that these can color understanding of the phenomena being investigated in unhelpful ways. As Bendassolli (2014) noted, this assumption probably has its roots in phenomenology (discussed below) and grounded theory, but it is more widespread. Many, though not all, qualitative researchers would assume that if their aim is to understand the multiple meanings that others construct about the world, then it makes sense to start with others’ accounts rather than theory (Cresswell & Cresswell, 2018). Moreover, the impetus for the development of qualitative methods has often come from a desire to “give voice” to marginalized and disempowered groups. Any theoretical interpretation risks impeding this by imposing the researcher’s meanings on those of the participant (Willig, 2017), particularly as there is a tendency for theory to incorporate the understandings of dominant social groups (Burr, 2015). As Dodgson (2018) noted in relation to breastfeeding, structural inequalities can result in such different life experiences that it is already difficult for members of more privileged groups to understand the experiences of those who are more marginalized. Uncritical use of theories that focus on individual behavior and how to change this (e.g., theory of planned behavior) can further exacerbate this blindness, by obscuring the operation of social inequalities and power dynamics, making the different and unequal contexts for mothering less visible (Horrocks & Johnson, 2014).

Therefore, qualitative research has often had an uneasy relationship with theory. At the same time, though, some have argued for greater use of theory in analyzing qualitative data (e.g., Anfara & Mertz, 2015), whilst other researchers have pointed to the inevitability of researchers bringing prior conceptual frameworks to the research process and the usefulness of articulating rather than ignoring these (Bendassolli, 2014; Maxwell, 2013).

How Theoretical Frameworks Can Inform Qualitative Breastfeeding Research

So far, we have focused on differences in philosophical assumptions between quantitative and qualitative research. However, there are also considerable variations in these assumptions within qualitative research (Denzin & Lincoln, 2018). Much more has been written about underlying conceptual issues in relation to qualitative methods as opposed to quantitative methods, and various qualitative paradigms have been developed that articulate different ontological and epistemological assumptions. As Maxwell and Mittapalli (2008) noted, these are often referred to as “theoretical frameworks” or “theoretical perspectives” and therefore will be considered here, though these broad perspectives might be more appropriately understood as normative or orientating frameworks that indicate particular goals for research. The intention of these frameworks is not so much to provide theoretical concepts for making sense of particular empirical observations but, rather, to indicate what kinds of data and concepts might be useful.

When these frameworks are expressed at highly abstract levels, qualitative researchers may refer to a position of, for example, neo-positivism, critical realism, limited realism, contextualism, interpretivism, constructivism, social constructionism, radical constructionism, or pragmatism, though textbooks on qualitative research (e.g., Cresswell & Cresswell, 2018; Denzin & Lincoln, 2018; Gray, 2018; King & Brooks, 2017) divide and label these positions differently so that terms can overlap. Although these philosophical frameworks might seem rather esoteric, engagement with them can help researchers to clarify whether, for example, they are treating participants’ accounts as straightforward reports of breastfeeding-related events and practices that have a “reality” outside these accounts (neo-positivism), or whether the focus is on how different participants view a complex world from different perspectives (critical realism). Alternatively, a researcher may view what participants say about their breastfeeding experiences as a partial window to their subjective experience and lived personal meanings (contextualism), as the manifestation of culturally shared understandings (social constructionism), or as a series of discursive moves that construct a particular account of their infant feeding choices (radical constructionism). These types of considerations can give an analysis a much sharper focus, whereas avoiding these ontological and epistemological decisions can result in a muddled account of findings.

Other broad frameworks incorporate one or more of the above philosophical positions but are less abstract and more clearly tied to particular research methodologies. This is perhaps more often what qualitative researchers mean by theoretical framework (Anfara & Mertz, 2015), and examples include phenomenology, narrative psychology, discourse analysis, grounded theory, and ethnography. For example, researchers have made good use of phenomenological approaches to try to understand the lived experience of breastfeeding. The aim of phenomenology is to understand phenomena as they appear in consciousness (Brooks, 2015), and (somewhat paradoxically, given that phenomenology is often referred to as a theoretical framework) it espouses an approach to analysis that is inductive rather than theory led (Howitt, 2016). The more descriptive forms of phenomenology (e.g., Giorgi, 2009) aim for rich description of participants’ experiences, following Husserlian ideas about placing the foundations of knowledge on an understanding of concrete phenomena, rather than on theoretical abstractions. Therefore, this approach advocates “bracketing” prior assumptions, including those that are derived from theory. However, other forms of phenomenology (e.g., Smith, Flowers, & Larkin, 2009) view interpretation as an inevitable part of lived experience and aim for a balance between understanding participants’ interpretations and also, unlike descriptive phenomenology, offering researcher interpretations that are sometimes drawn from prior theory. Use of the different types of phenomenology has encouraged a deeper probing of experiences of breastfeeding, aiming to move beyond everyday assumptions and medicalized frameworks, for example, drawing attention to the complexity of emotional responses to breastfeeding and breastfeeding support (e.g., Guyer, Millward, & Berger, 2012) and the relevance of existential security to women’s experiences of breastfeeding (e.g., Palmér, Carlsson, Brunt, & Nyström, 2015).

There is not the scope here to discuss the full range of theoretical-methodological frameworks available. However, examining one more in a little detail illustrates the extent to which use of different qualitative frameworks can guide analysis in radically different directions. Discourse analysis aims to explore language as a form of social action. Unlike phenomenological research, analysts adopting this perspective treat spoken and written accounts as a text that performs a function rather than as a “window” to subjective experience (Burr, 2015). Their aim is to grasp something of the ways in which different forms of talk about breastfeeding construct different versions of reality that can make different forms of action seem reasonable. Discourse analysis has drawn attention to the difficult moral work that women engage in to maintain or repair their identity when talking about infant feeding and the limited, often dichotomous, ways of talking about breastfeeding that are available (Ryan, Bissell, & Alexander, 2010). Other examples include Burns, Schmied, Fenwick, and Sheehan’s (2012) exploration of midwives’ talk about human milk as “liquid gold.” Their clear analytic focus on language highlights how emphasizing the nutritional aspects of human milk in midwives’ conversations with mothers resulted in the practices of breastfeeding and their relational aspects becoming less visible.

Since the 1980s, there also has been an increased interest in using broad theoretical frameworks that draw on feminism and critical theory to inform research with a social justice agenda (Cresswell & Cresswell, 2018). Although we considered above how conceptual analysis can risk imposing the meanings of dominant groups on participants from more marginalized groups, “thinking with theory” can alternatively challenge researchers to move beyond their taken-for-granted assumptions and see the world in different ways, avoiding the unwarranted reproduction of common-sense notions through qualitative analysis (Jackson & Mazzei, 2018; Wacquant, 2002). For example, use of critical theory can enable researchers to consider the role of less obvious power dynamics in producing social inequalities related to breastfeeding (Dodgson, 2018). In a recent issue of JHL, Thomas (2018) demonstrated how critical race theory could draw attention to structural racism and microaggressions, enabling analysis to move beyond an individualized understanding of the barriers to accreditation faced by lactation support providers of color. Similarly, Johnson, Williamson, Lyttle, and Leeming’s (2009) use of a poststructuralist feminist perspective was able to show how milk expression was not simply an individual “choice” but a way in which women were struggling to bring their subjectivity in line with cultural ideologies of motherhood in order to create the “good maternal body.” Although not necessarily having an explicit social justice agenda, social ecological theory has also been used to guide qualitative data collection and analyses of breastfeeding to ensure that individual experiences are understood in the context of broader systems (e.g., Dunn, Kalich, Henning, & Fedrizzi, 2015). However, despite many additional examples of the use of theoretical frameworks in qualitative research to extend thinking about the contexts and experiences of breastfeeding, these frameworks are not often used when researching breastfeeding interventions. This is perhaps because they have been developed in relation to naturalistic inquiry and do not lend themselves to the question “Does it work?” Using these frameworks could, though, encourage more penetrating questions when researching interventions (e.g., Phenomenology: “What was the experience of taking part, and what did this mean to participants?”; Ethnography: “How does this intervention fit with local cultural practices?”; and Discourse analysis: “What does this intervention communicate about infant feeding, and how does it do this?”; Leeming, Marshall, & Locke, 2017).

Using More Specific Theoretical Concepts to Illuminate Qualitative Findings

Sometimes, as well as using broad orientating frameworks, qualitative researchers draw on more focused theories or concepts. These can be used as a “lens” to enable specific aspects of the data to be viewed from a new perspective, or parallels to be drawn with other literature. For example, in a recent ethnographic study, Dowling and Pontin (2017) used the notion of liminality to examine their observational and interview date on long-term breastfeeding, showing how this constituted an uncertain transitional state “betwixt and between” previous and future ways of life. Drawing on anthropological literature, they noted how being out of step with cultural expectations around weaning and transitions in early motherhood could mean an uncertain, ambiguous, and even marginalized place in the social world—“a temporary identity for which our culture has no name” (p. 70). Boyer (2018) also demonstrated the value of drawing on a range of theoretical concepts for understanding the social and cultural context shaping qualitative data. She used concepts from cultural geography (e.g., “affective atmospheres” and “public comfort”) to take her analysis beyond common-sense notions of embarrassment to demonstrate how the discomfort around public feeding relates to intersubjective emotional experiences, which can leave women feeling out of place or as if they are not looking after the comfort of others. As Wilson and Chaddha (2010) noted, drawing on a range of theories at the point of analysis can provide a set of illuminating conceptual lenses without the risks of trying to “shoehorn” the data into one theoretical framework that has been chosen in advance.

Conclusion

Qualitative researchers exploring breastfeeding, early motherhood, and maternal and infant care could sometimes have a closer engagement with theory than they do. The above examples illustrate how exploratory research is not necessarily a phase of investigation prior to theoretical and conceptual development. Actively drawing on theory during qualitative analysis, and reflecting carefully on philosophical assumptions, can provide the researcher with a new and clearer perspective on the data and can enhance transferability of findings and the contribution of the research to wider academic understanding and to practice. However, it is worth thinking carefully about what a theory or conceptual framework will contribute to the research. Articulating a sophisticated theoretical framework is of little use unless this has a direct bearing on the aims and outcomes of research. Where the aim of a study is simply to summarize the views that stakeholders have about an intervention or to provide rich description of participants’ subjectivity, then theorization may not be what is required. As Anfara (2008) noted, theory should be used reflectively, for although it can reveal, it can also conceal by privileging some understandings over others.