Influence of the Research Frame on Qualitatively Derived Health Science Knowledge

Abstract

Although qualitative researchers generally acknowledge that their theoretical location and methodological orientation will influence the findings of their inquiries, it has been less well understood how the body of knowledge within a substantive field might be shaped by these factors. In this article, the authors draw on insights obtained from their experiences attempting to synthesize qualitative research findings. From that reflective process, they raise questions about the manner in which qualitatively derived knowledge from various orientations can be interpreted and understood.

While it is well recognized that the lens with which a qualitative researcher views a research problem will play a significant role in the outcome of the ensuing research, it is more difficult to work out precisely how those research results should influence our understanding of the state of the art knowledge in the particular field to which that study attempts to make a contribution. In this article, we examine many of the facets that comprise a researcher's angles of vision, or frame of reference, within a qualitative study, in an attempt to further elucidate the effects such lenses have on qualitatively derived knowledge in the health sciences. We draw on two meta-study projects in which we have been involved, one on chronic illness experience and the other on fatigue, in which the available body of qualitative research was rigorously and systematically examined for patterns and themes in the process and products of knowledge. On the basis of insights derived from these projects, we intend to challenge qualitative researchers to more accurately recognize and effectively account for the influence of their research frame on their findings, thereby strengthening the contributions that qualitatively derived knowledge can ultimately make to the health sciences and health care.

Background to our Analytic Process

Although results from our qualitative meta-study projects are reported elsewhere (Paterson, 2001; Paterson, Thorne & Dewis, 1998; Thorne & Paterson, 1998; Thorne, Paterson, Acorn, Canam, Joachim & Jillings, in press), some background will be useful to contextualize our particular insights. Using an adaptation of approaches evolved in sociology for meta-analysis across theoretical orientations (Ritzer, 1992; Zhao, 1991), we extended meta-theory, meta-method, and meta-data analysis strategies into a meta-synthesis approach called meta-study (Paterson, Thorne, Canam & Jillings, 2001). With a team of colleagues, we engaged in an extensive meta-study of all available research published between 1980 and 1996 on some aspect of living with a chronic disease. Systematically analyzing the implications of theoretical orientation, methodological strategy, and data interpretation for the 292 published reports that met our study criteria (see Appendix A), we were able to challenge many of the current and rather static conceptualizations of chronic illness experience and propose more comprehensive and inclusive interpretations that would better represent the spectrum of experience that persons with chronic illness report (Paterson, 2001). In a second meta-study of research conducted before 1999, we examined qualitative investigations into the phenomenon of fatigue within chronic illness. Using similar methods to categorize and interpret the 33 available studies (see Appendix B), we concluded that a number of distinct conceptualizations of fatigue existed within the researchers' orientation to the phenomenon and, at least in part, explained discrepancies and contradictions within the body of findings (Paterson, Canam, Joachim & Thorne, in review).

In both meta-study projects, we had the opportunity to deconstruct current knowledge claims made on the basis of qualitative research by examining and interpreting the theoretical and methodological foundations on which they were made. While many qualitative researchers seemed to surround their designs with epistemological claims consistent with subjectivism and social construction, they also typically expressed their findings in the language of ‘truth value’ and advocated application of their results as if they were fundamental to a shared reality. Because of this, we found ourselves increasingly disquieted by what qualitative health researchers are claiming their findings to represent, and we were inspired to reconsider the knowledge claims that are derived from the body of work as it evolves and develops over time. In this discussion, we draw from our experience with both of these studies to reflect on the implications of four distinct aspects of the research frame: its historical location, the manner in which the disciplinary orientation influences its logic, the theoretical and philosophical positions in which it is grounded, and the methodological strategy by which it has generated its findings.

Historical Location

While research related to health issues has been conducted in the social sciences for generations, the rapid proliferation of qualitative approaches in the health sciences began in the early 1980s. During the ensuing decades, nurses, physicians, social workers, and rehabilitation therapists have found that qualitative approaches provide an opportunity to address a wide range of health questions that are not amenable to inquiry using quantitative approaches. Throughout this relatively recent period of history, there have been tremendous shifts in how qualitative research is understood in the health sciences, the number and nature of methodological approaches that are available to investigators, and the expectations of a reading audience with regard to how methodology is located and explained.

These historical trends are important to understanding the style and form of research reports, and also for evaluating the quality of findings on the basis of reported aspects of the inquiry process. Because qualitative studies in the social sciences were traditionally understood to use methods that conformed to the standards of each discipline, many of the early health research reports included almost no discussion of the method that was used and how it was applied. In contrast, in the health sciences, qualitative researchers of the same era were attempting to publish their research reports in an academic climate in which deviations from traditional quantitative methods were not well understood and therefore required extensive explanation and documentation, not only of the methods themselves but also of the epistemological underpinnings from which they were derived.

During this same two decade period, there have been recognizable trends within the health fields related to which topics are commonly understood to warrant inquiry and how such topics are conceptualized. For example, in the early 1980s, an understanding of chronic illness as a discrete theoretical entity unto itself was still in its infancy, and the majority of studies concerned with chronic illness bounded their inquiries into discrete disease categories, occasionally but not necessarily linking their findings with those that might be relevant to other chronic diseases. As the body of research evolved, many researchers built their own inquiries on the designs and methods of others, and certain methods became more closely aligned with studies relating to certain diseases. For example, those studying diabetes were drawn to grounded theory methodology, while those studying multiple sclerosis seemed more likely to employ narrative inquiry methods. These patterns of how certain questions were considered in turn influenced what researchers thought might be interesting about various diseases and what factors or features they might be alerted to as they generated their subsequent investigations.

Over the course of this timeframe, some studies came to be held up as exemplars in the field, and became a standard foundation on which subsequent studies were grounded. For example, Charmaz's early work on loss and suffering in chronic illness (Charmaz, 1983) dominated the field for a decade or more until it was countered with equally compelling images of health within illness (Thorne & Paterson, 1998). Strauss' (1975) classic work on chronic illness trajectory depicted a pattern of experiential progression that was widely adopted as universally applicable until more complex conceptualizations began to emerge from the synthesis of the body of work (Paterson, 2001). This pattern of theoretical linkages, on which subsequent studies were expected to build, also occurred in relation to the conceptualizations used to study specific diseases. For instance, the early work linking epilepsy with stigma (Scambler & Hopkins, 1986; Schneider & Conrad, 1983) became almost ‘industry standard’ so that even today there are few alternative interpretations available in the body of qualitative research on epilepsy. Thus, although stigma is widely recognized as the appropriate conceptual representation associated with epilepsy, we know little about whether that linkage remains true within the current social context in which other diseases that evoke different judgements may have shifted the social context in which epilepsy is experienced.

In the case of other diseases, shifting trends in the popular conceptualizations associated with particular chronic conditions rose and fell in popularity over time in a discernable pattern. For example, although the traditional conceptualization linked with breast cancer was body image disturbance (Cohen, Kahn & Steeves, 1998; Thorne & Murray, 2000), the past two decades of inquiry have witnessed a shift to framing breast cancer first as a disease characterized by uncertainty (Hilton, 1988; Nelson, 1996) and later as one characterized by spiritual transformation (Coward, 1990; Peluzsi, 1997). Thus, in some diseases, certain prototypical conceptualizations seem to remain unchallenged over time while, in others, researchers interpret similar data into findings characterized by quite different conceptual structures.

As various diseases come to be identified with particular experiential concepts over time, these concepts often orient the next generation of inquiries and become the theoretical lens through which the disease is understood. So, for example, while it is difficult to discuss epilepsy without the reader assuming you are referring to stigma, the body of literature on stigma tends to be relatively silent on conditions that one might logically presume to have stigmatizing potential, such as psoriasis. Similarly, while sexuality and body image are typically presumed by health care providers to be a critical component of the breast cancer experience, it is much less likely that persons with inflammatory bowel disease or end-stage renal failure will be recognized as requiring support for sexuality and body image challenges. And, although fatigue has finally achieved some recognition as a prominent experiential element of chronic illness, it tends to be studied in relation to primary fatigue syndromes (such as CFS) and in the context of therapeutic interventions (such as breast cancer chemotherapy), rather than in diseases such as multiple sclerosis or rheumatoid arthritis, in which clinicians well recognize that fatigue is a pervasive characteristic of the chronic illness experience.

Another complication of the historical context of qualitative science in this field is the various degrees of affinity for interdisciplinarity that have evolved within the scholarship as compared to a disciplinary ethnocentricity. While some researchers explicitly read and report on the research of scientists from a range of disciplines, others focus almost exclusively on the work that derives from their own discipline. And while some researchers manage to explore research from a global perspective, others apparently limit their exploration to studies from a single nation, language group, or subset of journals. While linguistic boundaries account for this problem to some extent, it also seems likely that the Atlantic Ocean presents a much more conceptual divide than necessary, especially in this era of electronic communication.

From our analysis of these bodies of qualitative research, therefore, it seems apparent that the historical context into which each individual qualitative study emerges plays a significant role in determining what is considered relevant and interesting, which bodies of knowledge are recognised as foundational, how one ought to approach a research problem, which methodological strategies one should draw upon, and how research findings will be reported. Thus, when we try to understand the way in which the findings of any particular qualitative study contribute to our general understanding of any phenomenon, there are a number of critically important historical and contextual factors that must be thoughtfully considered.

Disciplinary Orientation

One prominent form of theoretical variation that becomes apparent when one reviews research reports across the health and social sciences is the powerful impact of the disciplinary lens that is applied to studies, even when investigators have used similar methods to study similar topics. In some instances, the disciplinary orientation of the author(s) is made explicit in the author credits, or can be implied from the intended audience of the journal in which the research is published. However, as researchers are increasingly working and writing outside their discipline, the disciplinary orientation of a specific published report may be difficult to discern. A critical read of the reference list for which authors have been cited and which others have not may be illuminating; in many instances, it will suggest that the logic of a particular discipline has been recognized as foundational background knowledge. In addition to cited references, cues in the way in which language is used and/or implicit assumptions within the research report can also reveal information about the disciplinary orientation that is being expressed in any particular study. At times, this level of scrutiny may reveal a departure from the discipline that one might have assumed from the author's academic preparation or home department; for example, health care professionals with social science training might be writing in the style of one discipline or the other or a unique application of both. The obscurity of the disciplinary perspective in many of the published qualitative research reports adds a layer of complexity to the process of analyzing their relative biases and to making sense of how those biases may have influenced the findings that are claimed as a result of the research.

The disciplinary orientation of any qualitative research study can have a profound influence on the way the research problem is framed, the way the research questions are posed, the data that are gathered and analyzed, and the findings that are produced (Thorne, 2001). As might be expected, psychologists tend to be oriented by the assumption that documenting processes of cognition and emotion are the ultimate aim of this kind of research. In contrast, sociologists tend to be unconcerned with what the individual cognitive processes might look like and focus instead on the interactional structures of behavioral patterns, social expectations, and cultural rules within health care (Gerhardt, 1990). Anthropologists tend to document patterns within beliefs about health and illness, and to locate them within the larger context of social, cultural and ethnic organizational frameworks. While the language, methods and questions deriving from these disciplines might appear to represent similar kinds of inquiry, our experiences of trying to interpret across the disciplinary divides have convinced us that research reports are strongly shaped by what the researcher understands to be the point of gaining new knowledge from a disciplinary perspective. Given the same interview or observation, each social scientist would likely ‘see’ it differently and use different reasoning processes in analyzing not only what is meaningful but also what meaning can be made of it.

In contrast to the social science disciplines, researchers working from a disciplinary perspective in the health sciences tend to generate and conduct their investigations from the perspective of gaining knowledge for application to clinical practice. Where they link their questions and interpretations to social science theoretical structures, they generally do so with a rather different intent than that of the social scientists (Wellard, 1998). As many of their research reports reveal, health science researchers tend to assume that the alpha and omega of each research project is indeed the clinical context (Miller & Crabtree, 1999). Thus, they use, and rely on, general theoretical and social science ideas, but in a manner that is considerably less theoretical and more applied than the usual standards of social science research might dictate. They might acknowledge and build on theoretical principles and common understandings, but express them with an identifiable attention to the unique and particular variations that constitute the cases on which these theoretical understandings would be applied in practice. Thus, many qualitative researchers working from a health science orientation are considerably more hesitant than their colleagues in the social sciences to align their inquiries with grand theoretical projects. As a result, their research results may be regarded with scepticism by their more theoretically inclined colleagues.

Interestingly, with the trend toward increasingly large and interdisciplinary research teams, many studies explicitly capitalize on the perspectives of different social and health science disciplines, and in some cases, the research reports reflect that hybrid species of varying orientations. Our investigations into disciplinary perspectives has therefore also led us to consider the effect that disciplinary publication standards have had on the form and style of research reports as well as on the way that authors frame their work in order to make them acceptable for publication and dissemination. For example, some social science journals have long traditions of discouraging detailed discussion of the application of method, especially when the methodologies inherent to the discipline can be assumed. The written reports of research in such journals can create a serious challenge to those with different disciplinary backgrounds attempting to audit the reasoning process by which the authors have reached their conclusions. Some health science journals, in contrast, typically demand a critical reflection on all aspects of the research process, such that reports seemed to provide considerably more detail about the researcher as instrument than about the findings of the research. Thus, in trying to critically interpret what sense can be made of the findings of any particular study, the astute reader must be able to appreciate not only the nature of the disciplinary grounding of any study, but also the traditions of that discipline's scholarly conventions.

Theoretical and Philosophical Grounding

Although it has become commonplace to articulate a primary distinction between qualitative and quantitative methods in health research, this classification sometimes blurs the wide range of epistemological and ontological variations within and between the methods that are considered qualitative (Frankel, 1999). Such variations arise in relation to the manner in which a study is grounded within existing literature, the assumptions that are made with regard to what constitutes knowledge, and the claims that can be justified on the basis of research findings. While disciplinary orientation represents an important theoretical and philosophical foundation that helps us understand the meaning of findings from specific qualitative studies, it is typically embedded within a much larger variety of theoretical and philosophical positionings that influence the research process and product. Thus, having established that scholars from various disciplines decidedly do not think alike, we can now turn to similarly powerful variations in thinking that may be independent of discipline. Each of these theoretical and philosophical frames can strongly influence the research. By understanding their influence, we are in a position to consider more critically the meaning of what has been studied and the claims about what is and is not known.

Proponents of qualitative methods differ with regard to the degree to which they expect a study to be grounded in prior theoretical and research literature (Sandelowski, 1993). In some instances, it is assumed that a study will build on all knowledge within a substantive field of study (such as illness-related fatigue). In others, it is recognize that an appreciation for existing empirical knowledge is essential to locating any new research question. However, various authors may interpret this requirement as including only research-based papers (such as those using a particular fatigue measure), or only qualitative investigations (such as those that have explored the experience of fatigue as articulated by those who know it first-hand) as relevant to the foundation of any new inquiry. Finally, there are some methods for which it is recommended that the researcher avoid serious attention to the pre-existing knowledge for fear of biasing the inductive reasoning processes that might occur during data collection and analysis. For example, some species of existential phenomenology and heuristics suggest bracketing all preconceptions to the extent possible and exploring the topic with a fresh perspective rather than running the risk of recreating the conceptualizations that have been argued by others (Schwandt, 1997). Because of these significant theoretical variations, it can be difficult for proponents of one method to recognize the credibility and value of findings of another, and to interpret the results of a particular study in the context of those deriving from different approaches (Smith, 1990).

Within and between disciplines, there is also considerable variation with regard to the way in which various authors understand the proper use of theory. Depending on method, discipline, and personal preference (or perhaps that of the dissertation supervisor!), some qualitative researchers explicitly locate their inquiries within a particular strand of theoretical development, and sustain that orientation throughout the report. Others seem reluctant to contaminate the originality of their findings by providing an overly detailed theoretical location from which their study was launched. Still others provide clues as to theoretical leanings, but insufficient information for the reader to fully understand the manner in which theory may have been integrated. In some instances, authors justify their studies on the basis that nothing is known about the phenomenon under question, and therefore qualitative methods are warranted. Unfortunately, such claims often ignore the presence of scholarly writing on certain aspects of the phenomenon that may take a different form, derive from a different discipline, or use different inquiry methods than those chosen by the author for the current review.

An even more curious feature of the theoretical orientation of qualitative research is that some authors explicitly locate their inquiries within a particular position in theory development and then apparently abandon all efforts to link, apply, or interpret their findings in that context. Thus, in some instances, theory may be useful for defending that a particular study is justifiable within a discipline's knowledge development regardless of whether that theory represents a grounding to which the researcher actually ascribes. Thus, a critical observer is well advised to thoughtfully consider not only which theory has been used to ground the research but also how it has or has not shaped the research process (Sandelowski, 1993).

A more subtle, but equally potent, theoretical grounding arises from the fundamental epistemological and ontological positions that the researcher takes in relation to the nature of the phenomenon under study and knowledge about it. In particular, the relative meaning and importance of objective and subjective knowledge has a critical impact on the questions, designs, and interpretations of qualitative research. If a researcher is engaged in fieldwork, interviewing persons afflicted by fatigue due to a particular health condition for example, the subjective impressions offered by such persons might variously be interpreted as evidence of objective reality, as an intersubjective construction between the researcher and the researched, or as their unique experience of a dominant socially constructed discourse. Different researchers take very different philosophical positions on such questions as where truth resides and whether there is an objective truth to be had in relation to the phenomenon (Altheide & Johnson, 1994; Kvale, 1995; Lincoln, 1995). For example, while one might assume that pain or fatigue are real simply because someone claims them to be, another might depend on clinical measures or behavioral signs to corroborate subjective reports. Such distinctions can sometimes be found in the overt claims that researchers make about their philosophical positionings, and at other times they can be inferred from their methodological choices, such as claiming uninterpreted raw data as findings, or using member-checks to validate their theoretical interpretations.

Thus, in understanding the potential contribution of a particular qualitative study, the explicit and implicit theoretical and philosophical positionings of the author can explain a great deal about why the findings are articulated and explained in the manner that they are and what sense the author believes one might make of them. Whether the author is presenting the reader with a truth claim or an alternative way of understanding a phenomenon makes an important difference in the way the research results can be interpreted and applied to ongoing inquiry or to practice.

Methodological Strategy

While the methodological approach adopted by any qualitative researcher is inevitably highly dependent on his or her historical location, disciplinary orientation, and philosophical perspective, there are aspects of the research method that warrant further consideration in trying to make sense of the contributions that several studies might make to a particular substantive field. In our chronic illness meta-study, we scrutinized almost 300 qualititative studies on some aspect of adult chronic illness experience. Within the social sciences, we found that most researchers adhered to the methods most common in their discipline, so that psychologists used phenomenological methods and anthropologists used ethnographic approaches and so on.

Within health sciences research, we found interesting patterns and trends in the use of various methods to address various questions, and noted that those patterns had indeed shifted over time as various claims gained theoretical dominance and popularity. Of those studies that explicitly claimed a methodological direction (64 percent of the total sample), 29 percent were grounded theory studies, 25 percent were phenomenologies, and 13 percent were ethnographies. Another 33 percent used a wide range of less prominent methods, such as narrative analysis, qualitative case analysis, discourse analysis, participatory action research, naturalistic inquiry, feminist methodology, and biographical analysis. While it was quite common in the 1980s in both the social and health sciences to publish ‘generic’ qualititative research, or studies in which the method was not specified, there is a decided preference for ‘named’ methods as the fields mature.

As one explores the relationships between method and study focus, design and findings, other insights emerge. For example, we have observed that narrative analysis methods are quite often used to study a particular constellation of diseases (such as HIV/AIDS, MS and cancer) but almost never in relation to other diseases such as diabetes, despite the fact that a disproportionate share of the chronic illness research derives from this population. Such observations caused us to question assumptions about who ‘has a story to tell’ about their illness as opposed to whose disease is understood to be more mundane or straightforward. One can also notice fascinating slippages between methods, so that some researchers explicitly apply grounded theory methods to studies of ‘lived experience’ and some phenomenologies are explicitly grounded within social theory. Thus, ethnographies do not necessarily contribute to sociocultural theory, grounded theories do not necessarily contribute knowledge of basic social processes, and phenomenologies do not necessarily tap subjective or phenomenal understanding. Further, we have noticed that studies purportedly using quite different methodological direction can be remarkably similar in their design. In particular, studies in the health sciences often apply general interpretive description approaches (Thorne, Reimer Kirkham & MacDonald-Emes, 1997) regardless of their methodological claims, searching for common patterns and themes within shared elements of a chronic illness experience, and theorizing the application of those common themes to an infinite set of unique cases.

Although all aspects of qualitative research are influential on the product, sampling perhaps deserves special attention with regard to how it creates the frame on which research findings are reported. In many instances, the unstated disciplinary, theoretical, and philosophical orientations that a researcher brings to an inquiry can be extrapolated by his or her sampling logic. For example, we were intrigued by how various researchers defined their samples when their intent was to study chronic illness in a general sense. In some instances, they drew from a mixed population, perhaps including individuals with such conditions as rheumatic, cardiovascular, and endocrine disorders. In other instances, they studied individuals with one disease and theorized a relationship to other diseases. Some studies included individuals with physical disabilities while others explicitly exclude these. Rarely were persons with chronic mental disorders unrelated to the physical illness included in these kinds of studies. Until fairly recently, most studies of chronic illness drew upon samples of persons whose diseases were not understood as ‘terminal.‘ As technological and other scientific advances have changed the trajectory of many such diseases through the 1990s, cancer and HIV/AIDS have started to appear quite regularly in the constellation of diseases that researchers consider chronic (Paterson et al., 2001).

To illustrate this issue further, we have noticed that some chronic diseases attract a disproportionate share of qualitative research interest, and therefore contribute more strongly than do other diseases to general chronic illness theorizing. As Bland (1999) noted, there has been almost no research enthusiasm for discovering what it is like to live with a chronic venous leg ulcer, despite its prevalence, the inherent chronicity of the condition, and the powerful degree to which it influences all aspects of a person's life. Similarly, Chronic Obstructive Pulmonary Disease and hypertension are rarely a focus of qualitative inquiry, while asthma and cardiovascular disease are quite popular. Among the most commonly studied chronic diseases are breast cancer and diabetes (both Type I and Type II), and while they clearly comprise significant subgroups within the chronic illness population, they cannot fully represent all of the major dimensions of chronic illness that are critically important in shaping what life is like. For example, both represent chronic illnesses in which lengthy periods of relative stability may make people more accessible to researchers. As Conrad (1990) pointed out, much of the available research into chronic illness experience seems to have been done on populations accessed through ambulatory outpatient clinics. This may suggest that convenience rather than theoretical or demographic interests play a role in determining the samples that are most commonly represented in the research literature.

Another observation that can be made on the basis of this body of work is that diseases that influence verbal communication, such as stroke, amyotrophic lateral sclerosis, or Parkinson's disease, make disproportionately unpopular research topics. This suggests to us that, besides the convenience factor, qualitative researchers may reveal a preference for conducting studies with adults who are easily articulate (Paterson et al, 2001). With the exception of studies that explicitly study a particular ethnic or social grouping's distinct experience with chronic illness, the sample populations of many studies are heavily weighted in the direction of white, educated, middle-class women, regardless of the extent to which that subgroup is representative of the total demographic picture of any particular chronic condition. Thus it might be argued that the subjective experiential knowledge that we have compiled may be rather strongly influenced by dominant values, beliefs, and assumptions common to that particular group and, because of this, much of the diversity inherent in chronic illness experience may have been rendered invisible within our research literature.

In addition to the general issues associated with what has been studied and what has not, a review of the body of work reveals some additional sampling issues that may influence our ability to extract general knowledge from the research-based literature. Among these is the frequency with which small samples are used in qualitative inquiries. While the depth and richness that a qualitative researcher typically seeks is more accessible when the sample size is constrained, most researchers would agree that there is a considerable trade-off with regard to the degree to which the findings are understood as generalizable, or relevant beyond the immediate study population. However, many researchers appear to lose sight of the size and nature of their samples in their enthusiasm to convey their confidence that their conceptualizations of the phenomena have far-reaching implications (Thorne, 2001). Thus, far too many research reports articulate implications applicable to populations that grossly exceed those for which they have theoretically sampled and can justifiably draw conclusions. They may ignore the possibility that there are relevant variations between urban and rural populations, between face-to-face interview data and data obtained by other methods, or between samples tapped on a single occasion as opposed to those followed over time. They may not recognize that there are numerous gaps in our knowledge base with regard to the impact of such discrete variables as age, gender, economic advantage, or ethnicity, as well as less explicit variables such as outlook on life, geographic or social location, and health system context.

Finally, another complicating factor within this body of research is the unfortunate tendency among some researchers to publish multiple reports on what appear to be the same data set without explicitly articulating common samples (such as would be the case in secondary analysis) (Paterson et al., 2001) By examining the demographics reported in what otherwise appear to be presented as discrete studies (that is, they claim distinct research questions, theoretical frameworks, analytic strategies and findings), one often suspects that researchers have simply split different kinds of conclusions from a single study into multiple research reports. Because publishers may be less enthusiastic about accepting manuscripts that claim to be partial findings, the culture of the research industry may have inadvertently contributed to this problem. Further, because the acceptable length of journal article research reports has been reduced in recent years due to economic pressures on publishers, investigators may not feel that a single journal article is sufficient to capture the depth and detail required for an effective and scholarly qualitative research account. Thus, while multiple publication is understandable, the failure of many researchers to acknowledge openly the interrelations between their data sets and findings creates a difficulty for those who are trying to make cumulative sense of entire bodies of literature. If the multiple research report represents constructions by the same researchers (instruments) on the same sample, the weight of conclusions from multiple publications should not rely more heavily in our interpretation than would those from a single written report, although sheer volume of writing sometimes affords an unwarranted impression of credibility. From our consideration of this large body of research, therefore, we believe that matters of sampling may be critically important to our shared understanding of any particular phenomenon.

Conclusions

Thus, our adventures in meta-study have afforded us an opportunity to reflect on the various ways in which historical context, discipline, theoretical orientation, and method shape individual qualitative research studies. We have become convinced that the research frame, or the angle of vision from which the researcher approaches a particular study, is critically important to understanding why a piece of research becomes what it does and what sense we might make of its contribution to the field of knowledge. It seems clear that, in closely examining each study in relation to each of these dimensions, we can learn a great deal about the researcher and, hopefully, something meaningful about the phenomenon about which the researcher is trying to generate new knowledge.

Clearly the theoretical and disciplinary direction of qualitative research in general is dynamic and evolving. The knowledge that we think we can extract from the products of such research is fluid and slippery, changing in both form and nature even as we re-examine it from yet another angle. As we continue to feel the pressure to account for the foundational knowledge that qualitative research offers to health care and health science, we would be well advised to steer clear of tight meta-syntheses and grand generalizations. Qualitative research does offer a different window on the world from that of its quantitative counterparts. But, despite an explosion of enthusiasm for qualitative inquiry into matters of health and illness, what it is that we actually ‘see’ when we look through this window will not be uncomplicated or consensual.