The term qualitative methods refers to a variety of research techniques and procedures associated with the goal of trying to understand the complexities of the social world in which we live and how we go about thinking, acting, and making meaning in our lives. These research practices, which emphasize getting close to participants and trying to understand how they (and we) view the world, include, among others, participant observation, interviews, life histories, and focus groups; autoethnographic, phenomenological, narrative, and most ethnomethodological and feminist approaches; particular forms of documentary, content, discourse, and conversational analysis research; and some action research.
Qualitative researchers may be placed along a broad continuum ranging from an orientation akin to positivist science to one more akin to art and literature. In between is a vast middle ground where elements of both orientations are present. Moving along the qualitative continuum from science to art and literature, one finds practitioners who see social life as something out here to be discovered independently of the researcher, those who view social life as something constructed through interaction and engagement with the world, and those who focus more closely on the person describing social life and the modes and practices of description (see Crotty 1998; Denzin 1997). Across the continuum, the focus changes from studying others who are assumed to be uniquely separate from the researcher, to examining interactions between the researcher and others, to including the positionality, politics, and story of the researcher who interacts with others.
Currently qualitative work enjoys a burgeoning interest across social science disciplines including anthropology, sociology, communication, education, social work, and nursing. The result is a growing sense of a qualitative community unconstrained by disciplinary boundaries. As this community grows and forms its identity, the spectrum of possibilities broadens, creating new alternatives for qualitative research and, in the process, raising vexing and controversial questions (see Denzin and Lincoln 1995; Snow and Morrill 1995). Given the interpretive turn (Rabinow and Sullivan 1987) in social science, more and more researchers are applying art-based criteria to their work; at the same time, new computer programs, such as NVivo, allow for more systematic and rigorous coding of qualitative data. We view these differences and ensuing conversations as strengthening qualitative research and representing a coming of age for qualitative work (Bochner and Ellis in press).
We organize our discussion of qualitative methods according to three ideal types representing points on the qualitative continuum from the science/ causal pole through the middle ground to the artful/interpretive pole. These categories are intended as a useful means of dividing the territory, but they are also arbitrary and should not be taken as a literal map of the field. Rather, we encourage readers to envision a wide expanse of methodological approaches and to view the boundaries we have constructed as permeable, flexible, and fleeting, with many qualitative researchers likely to position themselves in more than one category. As the authors, both of us have engaged in middle ground and artful/interpretive qualitative work; our current allegiance lies primarily in the artful/interpretive region of the continuum.
Qualitative Research as Science
At one end of the qualitative spectrum, researchers approach qualitative research as an extension of quantitative inquiry. Their goal is to produce propositional knowledge about human behavior generalizable to specific populations. They see truth as something ‘‘out there’’ to be discovered and explained by the researcher. Positioning themselves as neutral, disinterested parties in the research process, they want to be able to control and predict subsequent behavior of samples within the population being investigated. These researchers follow the building-block, foundational model of scientific knowledge, viewing inquiry as a linear progression, with each new discovery adding on to available explanations. Qualitative researchers in this tradition express many of the same concerns as their quantitative counterparts, including an interest in random sampling, reliability, validity, and ethical issues. The language they use to present and discuss results applies many of the familiar lines of the hypothetico-deductive model.
Random sampling of the studied population (regardless of specific research tools to be used) provides assurance that the qualitative researcher has obtained a representative and unbiased sample of research subjects (Lindlof 1995, p. 121). Since researchers want to claim that their findings can be generalized to the selected population, it is critical that the demographic characteristics of the sample match those of the population defined in the study. For example, Lowry and Towles (1989), in their study of the portrayal of sex and its consequences in afternoon soap operas, randomly sampled episodes of soap operas from each TV network in order to be able to draw conclusions about soap opera content in general.
Using an approach similar to that of quantitative research, qualitative researchers in this tradition examine variables that relate to specific behaviors, traits, or characteristics that are narrowly defined in as specific a manner as possible. They manipulate the independent variables and measure the outcome of the experiment in terms of dependent variables, those defined behaviors, traits, or characteristics thought to exist in relationship to the independent variables. Before they conduct experiments, researchers form a hypothesis about the relationship between the variables. Data interpretation then centers on determining whether the hypotheses are negated by the results; researchers do not generally examine data for themes or issues unrelated to the predetermined focuses of the study. Based on a review of relevant literature, Chavez (1985), for example, hypothesized that writers treat men and women differently in comic strips. She then collected a sample of comic strips, coded the gender, settings, and activities of the characters, and concluded that her hypothesis was supported.
This process of hypothesis formation and testing often is less formal, however, even among those striving to maintain a scientific approach to their research, and it can take many different forms. In the study of the realism of aggression on television by Potter et al. (1995), for example, the authors laid out a set of premises about what would constitute a realistic (similar to the real world levels of violence in numbers and context) portrayal of aggression. They then collected a sample of television programming, coded the various acts of aggression, and compared the numbers and types of aggression in their sample to the premises they had developed.
Even those researchers who do not convert their data into numerical form often go to great lengths to assure that their findings are valid and reliable. For researchers at this end of the qualitative continuum, validity means that the concepts they examine are those intended to be examined and not confounded or overlapping with other concepts not intended to be included. A study is reliable if researchers find, or could expect to find, the same, or very similar, results when they conduct the study again For example, Waitzkin (1990) provides criteria to establish reliability and validity of qualitative data: (1) discourse should be selected through a sampling procedure, preferably a randomized technique; (2) recordings of sampled discourse should be available for review by other observers; (3) standardized rules of transcription should be used; (4) the reliability of transcription should be assessed by multiple observers; (5) procedures of interpretation should be decided in advance, should be validated in relation to theory, and should address both content and structure of texts; (6) the reliability of applying interpretive procedures should be assessed by multiple observers; (7) a summary and excerpts from transcripts should accompany the interpretation, but full transcripts should also be available for review; and (8) texts and interpretations should convey the variability of content and structure across sampled texts (pp. 480–482).
Waitzkin’s criteria (1990) emphasize three specific concerns associated with reliability and validity from a scientific perspective. First, the collective decisions and interpretations of multiple researchers would be closer to an ‘‘objective’’ reality than a presumably more biased perspective of a single individual. Second, it is important to consider the whole body of available data as the basis for interpretation to avoid making general statements that reflect only a subset of the data; the emphasis is on what is common throughout the data, not on that which is unusual, distinctive, or unique. Third, written transcripts must be publicly available for verification.
Qualitative researchers working in this tradition
use a system of coding to categorize videotaped
or observed behaviors, written responses to
survey questions, verbal responses to an interviewer,
or other data (Wimmer and Dominick 1997).
Once labeled, the observed behaviors can be counted,
sorted, and analyzed statistically. Coding schema
can be standardized typologies that are used by
other researchers, or they can be developed in
light of a specific research question or body of
data. Chavez (1985), mentioned above, developed
a typology of settings and activities (e.g., child care
or working in an office) for the cartoon characters
based on what she found in her comic strip data
set. To aid in the analysis of transcript data, specialized
computer software programs, such as
NUD*IST, Ethnograph, or NVivo (a new program
that integrates text, image, sound, and video data
Ethical issues at this end of the continuum, similar to those in quantitative research, focus on methodological procedures, in particular honesty and thoroughness in data collection and analysis. Authors often elaborately spell out their research procedures in their publications, making the procedures and data available for scrutiny in order to justify claims or conclusions they draw. In scienceoriented qualitative research, authors stay behind the scenes, portraying themselves as trustworthy and credible through their disembodied discussion of methods without showing in their texts their own involvement or self-interest.
Those engaging in scientific approaches to qualitative research usually adhere closely to the writing style used by quantitative researchers. A passive voice shadows the presence of the author and obscures the ‘‘I’’ of the researcher (Gergen 1994). Statements such as ‘‘It was found that . . .’’ and ‘‘The data revealed that . . .’’ reinforce the notion of neutral authors who have discovered preexisting ideas, and who, without contaminating the material with their own perspectives, then pass it along for readers to receive as knowledge. Of course, researchers who see their work as scientific often acknowledge that the author is not a blank slate without values and beliefs, but their use of a disinterested, passive voice remains a sign of how important they view the ideal of distance and objectivity, even if it is not fully attainable.
Middle-Ground Approaches to Research
Between science and art, one finds a sprawling middle ground of qualitative researchers, who seek to analyze events, find patterns, and create models from their data (Neuman 1997). Here, researchers do not adhere rigidly to the rules of empiricism; but they are not likely to experiment with narrative, poetic, or literary forms of writing either. In the middle, researchers seek some combination of scientific rigor and artistic imagination. How these two goals intersect differs for various researchers and often connects to the author’s specific location relative to art and science on the qualitative continuum.
Middle-ground researchers use a variety of methodologies to gather data for analysis, including unstructured or semistructured interviewing (Fontana and Frey 1994; Mishler 1986), focus groups (Kitzinger 1994), participant observation or fieldwork (Lincoln and Guba 1985; Lofland and Lofland 1995), textual analysis (Reinharz 1992), and analysis of narrative (M. M. Gergen 1992; Riessman 1993). While there are many ways to go about selecting a sample, those in the middle ground of qualitative research often use purposeful sampling (Miles and Huberman 1984), in which they try to obtain cases that are information rich, or a ‘‘snowball’’ approach (Reinharz 1992), in which they ask current participants to identify other possible participants.
One of the most useful and widely applied strategies associated with the middle ground is an approach, developed by Glaser and Strauss (1967), called grounded theory. In this approach, researchers emphasize the generation of categories and theory from systematic coding and analysis of qualitative data (often transcripts) (Charmaz 1990; Glaser and Strauss 1967; Janesick 1994). This method of inductive reasoning differs from traditional social science in which researchers use previously established theory and test it deductively with data to see whether the theory can be sustained.
Methodological concerns are very important to grounded theory researchers, who often hold to the belief that if you apply a valid and systematic methodological approach, you’ll get closer to an accurate representation of what’s actually going on. In analyzing the data, some adhere rigidly to formal steps in grounded theory research—data notes, sort and classify, open coding, axial coding, selective coding, with memo writing occurring throughout the process (Charmaz 1990; Neuman 1997; Strauss and Corbin 1994). The more the researcher adheres to systematic analysis, the greater the likelihood of using computer programs to assist in coding. Other theorists, in the middle of the continuum, who think of themselves as grounded theorists, ‘‘eyeball’’ the data in less systematic ways, perhaps not even completely transcribing or coding data into categories. Yet they too seek patterns among the data they have collected, though they view the process as more subjective and intuitive than scientific and objective.
While some middle-ground researchers may place less emphasis on scientific precision, they usually adhere to criteria or guidelines concerning the processes of data analysis, though these rules may vary widely (see, for example, Charmaz 1997; Glaser 1978; Strauss and Corbin 1997). Tompkins (1994), for example, refers to representativeness, consistency (for public and private texts), and recalcitrance (sanction by research participants or a similar group) as standards for evaluating data used in qualitative research (see also Fitch 1994). Working closer to the interpretive pole, Lather (1986) argues that validity in openly ideological research can be established through four guidelines: triangulation of multiple data sources, methods, and theoretical perspective; assessment of construct validity through use of systematized reflexivity between the daily experiences of people and the theoretical constructs being developed; establishment of face validity through sharing analysis with respondents and refining analysis with their input; and determination of catalytic validity, that is the potential for bringing about positive change and social transformation (p. 67). These different criteria are similar insofar as they provide standards for bridging researcher and participant perspectives, so that findings reflect the meanings of the people whose lives were examined.
Ethical issues for those in the middle group focus on research practices such as covert research, deception, informed consent, anonymity, confidentiality, and revealing knowledge about the less powerful. These issues then lead to ethical questions about what should be studied, how, and by whom (see Lofland and Lofland 1995).
Most middle-ground researchers note the positionality of participants, such as race, class, and sexual orientation, in order to avoid obscuring these factors. For example, Ellingson and Buzzanell (1999) studied a group of white, middleclass, heterosexual breast cancer survivors in a small Midwestern city. They acknowledged that these demographic characteristics impacted the results of the study; a more racially mixed group, or a group composed of lesbians, for instance, most likely would have produced a different set of findings.
Just as those in the middle ground acknowledge the positionality of participants, they also sometimes acknowledge their standpoint, personal background, politics, and interests in the topic (Collins 1991). Insofar as they see knowledge as ‘‘constructed’’ rather than discovered, middleground researchers discuss their personal perspectives or political commitments as an acknowledgment that all knowledge is generated from a specific social position and reflects the perspectives of the researchers involved. Researchers in the middle ground may try to decrease the power disparity between themselves and the people they study (DeVault 1990; Ellingson and Buzzanell 1999). They generally refer to those in the study as research participants or informants rather than subjects, indicating a degree of respect for the people whose lives are being studied (Reinharz 1992).
In the middle ground, researchers study a variety of topics, including complex issues that are difficult or impossible to address with quantitative methodology, such as understanding hierarchy in groups (Whyte  1993) or awareness contexts in death (Glaser and Strauss 1964). Some seek to make visible previously invisible aspects of the lives of women and other groups underrepresented in traditional social scientific research such as ethnic and racial minorities, gays and lesbians, and people with disabilities (Spitzack and Carter 1989). Others examine groups that are hidden, unknown, or inaccessible to most people, such as mushroom gatherers (Fine 1992) or white supremists (Mitchell 1998). As the topics get more complex and oriented toward meanings, subjectivity, and emotionality, it becomes more difficult to invoke older, more traditional, systematic ‘‘scientific methods’’ and apply them.
Writers in this tradition alter some of the conventions of scientific writing. They may include standpoint statements within the introductory section of articles, indicating their personal interest in the topic. They may use a conventional format but preface the article with vignettes or include excerpts from participant narratives in the discussion of findings or in an appendix to add texture and authenticity to the work. To acknowledge their presence in the work, authors may write in the first person. Nevertheless, researchers in this tradition usually privilege theory generation, typicality, and generalization to a wider world over evocative storytelling, concrete experience, and multiple perspectives that include participants’ voices and interpretations. They tend to write realist tales in an authorial, omnipotent voice. Snippets of fieldwork data then represent participants’ stories, primarily valued for illustrating general concepts, patterns, and themes (see Van Maanen 1988).
Research as Artistic Endeavor
During the last two decades, many qualitative researchers have moved toward an emphasis on the artistic aspects of qualitative work (Wolcott 1995). Working from an orientation that blends the practices and emphases of social science with the aesthetic sensibility and expressive forms of art, these researchers seek to tell stories that show experience as lived in all its bodily, cognitive, emotional, and spiritual aspects. The goal is to practice an artful, poetic, and empathic social science in which readers can keep in their minds and feel in their bodies the complexities of concrete moments of lived experience. These writers want readers to be able to put themselves in the place of others, within a culture of experience that enlarges their social awareness and empathy. Their goals include: evoking emotional experience in readers (Ellis 1997); giving voice to stories and groups of people traditionally left out of social scientific inquiry (DeVault 1990); producing writing of high literary/artistic quality (Richardson in press); and improving readers’, participants’, and authors’ lives (see Denzin 1997; Fine 1994).
According to Bochner, Ellis, and their colleagues (Bochner 1994; Bochner et al. 1998; Ellis 1997), the interpretive, narrative, autoethnographic project has the following distinguishing features: the author usually writes in the first person, making herself or himself the object of research (Jackson 1989; Tedlock 1991); the narrative text focuses on generalization within a single case extended over time (Geertz 1973); the text is presented as a story replete with a narrator, characterization, and plot line, akin to forms of writing associated with the novel or biography; the story often discloses hidden details of private life and highlights emotional experience; the ebb and flow of relationship experience is depicted in an episodic form that dramatizes the motion of connected lives across the curve of time; a reflexive connection exists between the lives of participants and researchers that must be explored; and the relationships between writers and readers of the texts is one of involvement and participation.
Rather than believing in the presence of an external, unconstructed truth, researchers on this end of the continuum embrace narrative truth (Spence 1982), which means that the experience as described is believable, lifelike, and possible. Through narrative we learn to understand the meanings and significance of the past as incomplete, tentative, and revisable according to contingencies of present life circumstances (Crites 1971). In this research, authors are concerned about issues of validity, reliability, and generalizability, but these issues may be redefined to meet the goals of the research.
Ellis and Bochner (in press), for example, define validity to mean that the work resonates with readers and evokes in them a feeling that the experience has verisimilitude. A story is valid if it stimulates readers to enter the experience described or to feel and think about their own, whether in memory, anticipated, or lived. Validity might be judged by whether it offers readers assistance in communicating with others different from themselves, or a way to improve the lives of participants and readers or even the author’s own. Since writers always create their personal narrative from a situated location, trying to make their present, imagined future, and remembered past cohere, orthodox reliability does not exist in narrative research. But reliability checks are still important. Researchers often take their work back to participants and give them a chance to comment, add materials, change their minds, and offer their interpretations. Since we all participate in a limited number of cultures and institutions, lives are typical and generalizable, as well as particular. A story’s generalizability is constantly being tested by readers as they determine whether the story speaks to them about their own experiences or about the experiences of others they know. Likewise, does it tell them about people or lives about which they are unfamiliar? Does a work bring ‘‘felt’’ news from one world to another and provide opportunities for the reader to have vicarious experience of the things told (Stake 1994)?
Interpretive research reflects the messiness of lived experience and emotions. Unlike researchers at the scientific end and in the middle ground, artful or interpretive researchers do not look for common denominators and generalities, but instead examine experience that is unique, particular, moving, and possible. They embrace their own subjectivity not to acknowledge bias but to celebrate positionality and their particular construction of the world. While some of these writers employ traditional analysis in their work, examining stories for concepts, patterns, and themes, others argue for the importance of thinking with a story, not just about a story. Thinking with a story means to allow yourself to resonate with the story, reflect on it, become a part of it (see Frank 1995b). Others argue that theory is embedded in the story (Ellis 1995b), that all good stories make a theoretical point.
Arguments about methods are not nearly as prevalent here as in the other two groups. Methods articles are much more likely to emphasize flexibility and emergence than to offer strict rules for how to proceed (Ellis et al. 1997). One of the most discussed methodological issues is whether researchers should attempt to follow the rules of traditional ethnographic methods or whether narrative research should be approached more like writing fiction. One’s position on that question most likely intersects with one’s stance regarding the role of criteria in evaluating narrative writing. In the former, traditional ethnographic criteria might be more commonplace; in the latter, narratives might be judged by their usefulness, the compassion they encourage, and the dialogue they promote (Ellis and Bochner in press). Seeking to position interpretive texts as an intersection of social science and art form, Richardson (in press) discusses five criteria she uses to evaluate interpretive ethnography: (1) Substantive contribution: Does the work contribute to an understanding of the text? (2) Aesthetic merit: Is the text artistically shaped, complex, and satisfying? (3) Reflexivity: Does the writer reflect on the production of the text so that the reader can make judgments about the point of view employed in the text? (4) Impactfulness: Does this work affect me and move me to respond or act? (5) Expression of Reality: Does this work seem a credible account of the real?
Rather than method and criteria, most articles about interpretive/artful ethnography grapple with issues of writing. For narrative researchers, writing (the form) is inseparable from the process of data interpretation (the content). As Richardson (in press) phrases it, writing in interpretive qualitative research is not a ‘‘mopping-up’’ activity at the end of a research project but an integral part of data analysis; authors write to find out what they have to say about their data and their experiences. The voice of the author is as central to the text as the voices of those ‘‘being studied.’’ Writers experiment with approaches and writing conventions to dislodge assumptions about science and incorporate the artistic into representations of data.
Interpretive research embraces a range of processes and approaches, including biographical method (Denzin 1989), observation of participation (Tedlock 1991), ethnography (Van Maanen 1988), autoethnography (Ellingson 1998; Reed-Danahay 1997), interactive interviewing (Ellis et al. 1997), systematic sociological introspection (Ellis 1991), co-constructed methods (Bochner and Ellis 1992), personal experience methods (Clandinin and Connelly 1994), narrative analysis (Bochner 1994), and feminist methods (Reinharz 1992).
Examples of creative representation include layered accounts (Ronai 1995), ethnographic fiction (Angrosino 1998), personal essays (Krieger 1991), impressionist tales (Van Maanen 1988), coconstructed narratives (Bochner and Ellis 1992), poetic representation of data (Austin 1996; Glesne 1997; Richardson 1992, 1994), writing stories (Richardson 1997), ethnographic performance or ethnodrama (Jones 1998; Mienczakowski 1996), and polyvocal texts (Lather and Smithies 1997).
A particularly controversial narrative writing practice and form is autoethnography, which is an autobiographical genre of writing and research that displays multiple layers of consciousness as it connects the personal to the cultural. Several autoethnographic genres currently exist side by side (Ellis and Bochner in press): (1) Reflexive ethnographies focus on a culture or subculture, but authors use their own experiences in the culture reflexively to bend back on self and look more deeply at self-other interactions. Reflexive ethnographers ideally use all their senses, their bodies, moment, feeling, and their whole being to learn about the other (Jackson 1989). (2) In native ethnographies, researchers who are natives of cultures that have been marginalized or exoticized by others write about and interpret their own cultures for others. (3) In personal narratives, social scientists take on the dual identities of academic and personal selves to tell autobiographical stories about some aspect of their daily life.
In all these forms of qualitative writing, narrative researchers use fiction-writing techniques such as dramatic recall, dialogue, lashback, strong imagery, scene setting, character development, interior monologue, suspense, and a dramatic plot line that is developed through the specific actions of specific characters with specific bodies doing specific things. They ask readers to relive the events with the writer and then to reflect on their emotional responses and life experiences, and on the moral questions and concerns that arise. Yet, the work differs from fiction in that the writing and publishing conventions used arise out of social science traditions, and in that the work often, though not always, has more of an overt analytic purpose and more of an analytic frame than fiction has.
Ethical concerns include matters of how we go about doing our research and what we owe those who become characters in and readers of our stories (Ellis 1995a). How do we deal with issues of confidentiality when including people, such as family members, who can be easily identified? What do we do if characters in our stories disagree with our interpretations or want us to omit material about them? How do we include multiple voices and perspectives within our stories? What is the role of the traditional authorial voice of the author? How do we stay true to our participants yet not deceive our readers (Josselson 1996; see also Chase 1996)? Is there value in working from an ethic of care, empathy, and responsibility, rather than informed consent, and is that ever possible in the world of research (Collins 1991; Denzin 1997)? How do we make our projects therapeutic for ourselves as well as our participants and readers? What do we want the world to be? How can we contribute to making it that way and, in the process, become better human beings?
Qualitative methods is a rich and varied set of approaches to research. Journals such as Qualitative Inquiry, Qualitative Sociology, Symbolic Interaction, and the Journal of Contemporary Ethnography, as well as a number of research annuals such as Studies in Symbolic Interaction and Cultural Studies: A Research Annual, along with subfield-specific journals, such as Qualitative Health Research, showcase examples of qualitative research in sociology and provide forums for discussion of methodological issues. The joy (and sometimes the frustration!) of qualitative methods is the promotion and valuing of a wide spectrum of methods, none of which should be viewed as the only way to conduct research in sociology. It might be more comforting if there was one set of rules to follow, but that comfort would come with the tragic price of closemindedness, silencing of voices, and narrowing of vision. We agree with Rorty (1982), who says we ought to learn to live with differences without feeling we have to resolve them.
edu.learnsoc.org Copyright 2010 - 2012 © All Rights Reserved
|Home | About | Contact | Links|