The reason and rhyme of qualitative research: why, when, and how to use qualitative methods in the study of adolescent health

By furthering clinicians’ understanding of human factors that influence adolescent health, including social context and the patient perspective, qualitative research enhances the type of knowledge that allows us to more effectively prevent unhealthy behaviors and, ultimately, to heal disease.

Abstract

Qualitative inquiry in adolescent healthMany of the life experiences, understandings, and beliefs that place young people's health at risk are difficult to quantify, yet adolescent health providers must assess and evaluate them. Qualitative research offers tools to examine these powerful forces at work in young lives. It looks beyond diagnostic outcomes to explore context and motivation, the "how" and the "why" of adolescent health risk behaviors. Although qualitative methods are relatively new to medical and public health research, clinicians use qualitative techniques every day: observing, actively interviewing, receptively listening, evaluating narratives, integrating data from diverse sources while recognizing inherent biases, and analyzing the information in a flexible and critical manner. This intuitive use of qualitative techniques informs clinicians how to vary their interactions subtly to best meet the differing needs of their patients. This "art of medicine" combined with biomedical science allows clinicians to move beyond medical problem solving and become healers. The goal of all research is to enhance a knowledge base. By furthering clinicians' understanding of human factors that influence adolescent health, including social context and the patient perspective, qualitative research (1Denzin N.K. Lincoln Y.S. Handbook of Qualitative Research. Sage, Thousand Oaks, CA1994Google Scholar, 2Glesne C. Peshkin A. Becoming Qualitative Researchers. Longman, White Plains, NY1992Google Scholar, 3Marshall C. Rossman G.B. Designing Qualitative Research. 2nd ed. Sage, Thousand Oaks, CA1995Google Scholar, 4Morse J.M. Field P.A. Qualitative Research Methods for Health Professionals. 2nd ed. Sage, Newbury Park, CA1995Google Scholar) enhances the type of knowledge that allows us to more effectively prevent unhealthy behaviors and, ultimately, to heal disease.Pursuing an answer: quantitative and qualitative pathsThe scientific method, observation, hypothesis generation, and hypothesis testing, is the central structure of both quantitative and qualitative inquiry. They share an investigative approach that poses a question, collects and analyzes data, and presents results. Scientific rigor and the integrity of a theoretical framework are critical to both. Although quantitative and qualitative research are often presented as being diametrically opposed, this polarization is both artificial and inaccurate. Quantitative researchers have viewed qualitative research as lacking scientific rigor, and qualitative researchers have believed that quantitative researchers were missing the human element. Recently, however, there is growing recognition that the two fields complement each other and that the greatest strides toward improved health can be made when both methods of inquiry respond to a problem (5Stange K.C. Zyzanski S.J. Integrating qualitative and quantitative research methods.Fam Med. 1989; 21: 448-451PubMed Google Scholar, 6Waitzkin H. On studying the discourse of medical encounters A critique of quantitative and qualitative methods and a proposal for reasonable compromise.Med Care. 1990; 28: 473-488Crossref PubMed Scopus (86) Google Scholar, 7Steckler A. McLeroy K.R. Goodman R.M. et al.Toward integrating qualitative and quantitative methods An introduction.Health Educ Q. 1992; 19: 1-8Crossref PubMed Scopus (360) Google Scholar, 8Porter E.J. The qualitative-quantitative dualism.Image J Nurs Scholar. 1989; 21: 98-102Crossref PubMed Scopus (26) Google Scholar, 9Inui T.S. The virtue of qualitative and quantitative research.Ann Intern Med. 1996; 125: 770-771Crossref PubMed Scopus (52) Google Scholar, 10Pope C. Mays N. Reaching the parts other methods cannot reach An introduction to qualitative methods in health and health services research.BMJ. 1995; 311: 42-45Crossref PubMed Scopus (1615) Google Scholar). No research approach is complete or flawless; quantitative and qualitative methods have different strengths and limitations.A first step toward deciding the appropriate methodological path is to focus on the essential question the research intends to answer. Any study of adolescent health issues tries to answer one or more of the following basic questions: What? When? Whether or not? How many? To whom? How? Why? Quantitative methods offer the best and most verifiable answers to the more concrete questions of "what," "when," "whether or not," and "how many" phenomena occur. In the earliest stages of research, however, "what" to study is sometimes not clear. In that case, qualitative inquiry can be used to uncover problems and generate hypotheses. Quantitative research can answer "to whom," but is limited in its ability to describe people. While it can indicate what happens to people from differing demographic groupings, quantitative inquiry lacks the ability to look at the complex interplay among factors that produce individual choice or behavior. Although it cannot look for trends among large groups, qualitative research is an ideal approach to elucidate how a multitude of factors such as individual experience, peer influence, culture, or belief interact to form people's perspectives and guide their behavior. Most importantly, qualitative methodologies can offer far greater insight into "why" and "how" phenomena occur than can quantitative methods.Next, the researcher must consider the way in which the philosophies, perspectives, and tools of quantitative and qualitative modes of inquiry differ. In general, quantitative methods work with discrete, numerical data, analyzing their frequencies and statistical associations, while qualitative research examines narratives, meanings, and behaviors in social context. Medical and public health research has traditionally been conducted under the positivist paradigm, which assumes that reality is objective and quantifiable. Quantitative researchers stand outside their subject, measuring external manifestations, approaching phenomena from the outside in, using objective, tangible variables. Numerical values are assigned to phenomena and validity is achieved through measuring enough cases that the results become statistically significant. The quantitative approach is ideal for studying biomedical phenomena such as the body's response to a medication, or epidemiological trends such as the increasing prevalence and morbidity of asthma.While quantitative research can tell us much about the incidence and outcomes of disease, it cannot answer how to get patients to use medication when it does not make them feel better, or why, despite improving medical technology, asthma is getting worse. The health and risk behaviors of adolescents are clearly affected by social, psychological, economic, and cultural forces. Successful health care depends as much on the human nature, life experience, beliefs, and motivations of the patient as on offering the best medicine. Based on the assumption that reality is socially constructed, complex, and constantly changing, qualitative research investigates behaviors through the human perceptions, understandings, and beliefs that motivate them. Qualitative investigators seek to understand their subjects from the inside out. Inductive rather than deductive, qualitative methods are designed to study a particular phenomenon, group, or behavior in depth to reach a better understanding of the universal. The qualitative researcher often works with a small sample size, deriving detailed information on participants in their living contexts, gaining "thick descriptions" (11Geertz C. Thick description Toward an interpretive theory of culture.in: Geertz C. The Interpretation of Cultures. Basic Books, New York1973: 3-30Google Scholar) of specific situations or case histories, seeking patterns in complexity rather than a simplifying consensus overview.Qualitative research methods search for answers from the broader social context rather than isolating the subject of study. Factors that might be considered confounders to quantitative methods may be embraced in qualitative research; their interrelationships are explored as part of the complex whole being studied. Qualitative inquiry can generate hypotheses as well as test them, starting with a blank slate and developing theory gradually from the data (12Strauss A.L. Corbin J.M. Basics of Qualitative Research. Sage, Newbury Park, CA1990Google Scholar). Where quantitative methods achieve rigor in part by fixing the hypothesis and the method for testing it at the outset of a study, qualitative inquiry finds richness and depth by following the subject matter where it leads.Collecting qualitative dataThe closer to the natural state in which health research data are collected, the fewer limitations exist in translating the findings into real-life applications, but the greater probability that the collected data may be biased by investigator perspective. For example, behavioral research can be conducted on neurotransmitters in a laboratory environment, where it is highly controlled, but many steps need to be taken before experimental results can be applied to humans. A behaviorist who integrates into a study population, participating fully in all activities, is observing behavior in a more naturalistic setting. However, to pursue such research with minimal alteration of population behaviors requires sensitivity and careful technique. In general, qualitative research is far more naturalistic than quantitative research, but even within qualitative inquiry, the proximity to the naturalistic environment and the degree to which the investigator changes the environment can vary.Most research struggles to move toward the naturalistic setting, but ethics, cost, and even safety often necessitate a stepwise compromising approach. For example, a research pharmacologist first tests a drug in a test tube, knowing the effect may not be predictive of the same effect in humans, but risking no lives. If the test results are as hypothesized and no unforeseen problems arise, the drug will be tested on an animal model and, ultimately, humans. An example more relevant to adolescent research is the gathering of the youth perspective. The written survey collects opinions, but it is less than optimal because people do not form opinions on paper; rather, they do so through conversation or life experience. Collected opinion becomes more natural when transformed into conversational form by means of the interview. Another natural setting is found in a focus group, where questions are asked of groups that create the give-and-take atmosphere in which opinions naturally form. The most natural setting to gather opinions might be through participant observation, by which the researcher can observe real-world groups for interactions that formulate and actions that reflect opinions.Qualitative research methods recognize the primacy of the subject of inquiry. The research question and the nature of the data determine and guide the methods used. Well-established as research tools in other fields of human study such as anthropology, sociology, and education, there is a wide variety of qualitative methods of which this overview can discuss only a few. Qualitative data may consist of field observation in the participants' natural environments or other milieux, oral or written narrative, text, sounds, and/or visuals. Many qualitative studies use more than one type of data and collection technique to enrich and add perspective to the pool of information on their subject of inquiry.When researchers are investigating health risk behaviors and their underlying belief systems and motivations, interviews may be an effective means of data collection (13Crabtree B.F. Miller W.L. A qualitative approach to primary care research The long interview.Fam Med. 1991; 23: 145-151PubMed Google Scholar, 14Spradley J.P. The Ethnographic Interview. Holt, New York1979Google Scholar). Interview protocols may range from precisely worded scripts that are repeated with each participant to semistructured conversations, in which broad initial questions elicit participants' responses, which in turn generate further questions, following the unique flow of ideas that occurs with each participant. Those that follow precisely worded scripts can generate quantifiable data similar to those collected via a survey. Advantages of collecting data by interview rather than survey are the more naturalistic setting and the elimination of literacy as a limitation; a disadvantage is that people may be more comfortable telling the truth to a piece of paper than to a person.Focus groups, a variation on the interview in which a group is asked for responses to and discussion of research questions, may prove useful when examining consensus beliefs or group behaviors such as binge drinking (15Stewart D.W. Shamdasani P.N. Focus Groups. Sage, Newbury Park, CA1990Google Scholar, 16Moore C.M. Group Techniques for Idea Building. Sage Publications, Newbury Park, CA1987Google Scholar, 17Krueger R.A. Focus Groups. Sage, Newbury Park, CA1988Google Scholar). Different types of focus groups use a variety of techniques to balance the free flow of ideas while limiting the biasing effects of group dynamics. For the purpose of adolescent-based research, two techniques are most appropriate. The open focus group uses a facilitator to guide a group of participants through a discussion focused on a specific issue. Nominal Group Technique is a one-question method in which a group generates and prioritizes solutions to the question; it uses a tightly structured format designed to limit group dynamics.Ethnographic study of an adolescent subculture might obtain data by means of participant observation (18Reason P. Participation in Human Inquiry. Sage, Thousand Oaks, CA1994Google Scholar, 19Berg D.N. Smith K.K. The Self in Social Inquiry. Sage, Newbury Park, CA1988Google Scholar), where the researcher observes people and their behaviors from within the social matrix of their worlds. The researcher explores environments, participants, and events from a perspective that can range from marginal observer to full participant. In suburban malls or on inner-city streets, for example, one can examine and experience the rituals of adolescents' lives from within their worlds. With participant observation, the researchers themselves become instruments of data collection, studying the participants' lived experience by observing, feeling, responding to, and noting their own perceptions, as well as more objective observations of study participants and their behaviors.Where the inquiry is concerned with the meaning of illness, relationships, or sexuality, the data might be collected through semistructured interviews, journals, or other participant-generated narratives. The forms that narratives take may be culturally or linguistically based, they may be oral or textual, preexisting or elicited by the researchers. Cultural documents, such as newspapers, magazines, and radio or television broadcasts may be studied (20Rich M. Woods E.R. Goodman E. et al.Aggressors or victims gender and race in music video violence.Pediatrics. 1998; 101: 669-674Crossref PubMed Scopus (67) Google Scholar), as well as more personal or group-specific artifacts, photographs, letters, graffiti, and video or audio recordings. Self-documentation, in the form of diaries, poetry, photos, or videos created by research participants, can give them voice and independence, allowing participants to become collaborators in the research (21Rich M. Lamola S. Chalfen R. Video Intervention/Prevention Assessment (VIA) An innovative methodology for understanding the adolescent illness experience.J Adolesc Health. 1998; 22 (Abstract): 128Abstract Full Text PDF Google Scholar). Whether created independently by the participants or motivated by the researchers, these personal documents can be analyzed for the objective features of participants' lives, worlds, relationships, and experiences, but also for what they reveal of the participants' subjective perceptions, understandings, meaning making, and inner lives.Analyzing qualitative dataBecause the goal of qualitative research is to derive rich descriptions of the world and of human experience, subjects of inquiry which are nonlinear and multidimensional, qualitative data are often analyzed from multiple perspectives using different analytical methods (22Miles M.B. Huberman A.M. Qualitative Data Analysis. 2nd ed. Sage, Thousand Oaks, CA1994Google Scholar, 23Silverman D. Interpreting Qualitative Data. Sage, Thousand Oaks, CA1993Google Scholar, 24Wolcott H.F. Transforming Qualitative Data. Sage, Thousand Oaks, CA1994Google Scholar). Analysis begins with the initiation of data collection and continues and evolves throughout the course of the research project, responding to the nature of the findings as the phenomenon being studied is better understood. Like data collection, analysis often evolves and changes as the study progresses to derive the richest, most complete findings and to improve the internal validity of the study. Some methods of analysis consider data in their pure, raw forms as direct delivery of experiences, thoughts, and behaviors to the researcher's theoretical framework. Contextual analysis looks at the data as a whole, while categorical analysis organizes and codes data by ideas or themes. As with quantitative research, qualitative methods benefit from explicit, well-organized data collection, management, and analytic procedures so that the research process is clear and may be replicated, so that the findings and conclusions are verifiable, and so that scientific fraud can be tracked. Data analysis consists of four linked processes: data reduction, data display, conclusion drawing, and verification (22Miles M.B. Huberman A.M. Qualitative Data Analysis. 2nd ed. Sage, Thousand Oaks, CA1994Google Scholar).Data reduction can occur both in an anticipatory way, by collecting the data within a conceptual framework which focuses and bounds the subject of study, and after the data are collected, by summarizing, coding, and organizing the data by themes or within narratives. The goal of qualitative inquiry is to determine and explain a pattern of relationships that may best be analyzed by grouping the data into conceptual categories, either predetermined or arrived at through the accumulation of data. Theoretical frameworks that guide this grouping of the data can include one or more of the following: •narrative analysis (25Riessman C.K. Narrative Analysis. Qualitative Research Methods. Sage, Newbury Park, CA1993Google Scholar)—understanding human motivations, perceptions, and behavior by interpreting the stories people tell of themselves and their experiences•ethnography (26Hammersley M. Atkinson P. Ethnography. 2nd ed. Routledge, London1995Google Scholar)—studying people through the nature of their social structures and behaviors•medical anthropology (27Sargent C.F. Johnson T.M. Medical Anthropology. Revised ed. Praeger, Westport, CT1996Google Scholar)—analyzing how people understand and respond to health and illness through their filters of culture and belief•phenomenology (28Bergum V. Being a phenomenological researcher.in: Morse J. Qualitative Nursing Research A Contemporary Dialogue. Aspen, Rockville, MD1989: 43-57Google Scholar)—understanding humans through the meanings they attach to experience•grounded theory (29Glaser B.G. Basics of Grounded Theory Analysis. Sociology Press, Mill Valley, CA1992Google Scholar)—theory that is not predetermined but is gradually developed from, evolves with, and is tested against the data collected.Data display codes, organizes, and structures the compressed data into theory-guided synopses, patterns, or matrices so that they can be viewed and arranged into a logical whole. This display and arrangement of data can be done manually or by computer. Manually, the data groupings can be placed on index cards for arrangement into a logical pattern, then transferred into graphic form as a grid or flow sheet with arrows indicating the directions and (sometimes) magnitude of effect. Narratives or other text can be read and analyzed for meanings and motives. Audio or video recordings can be reviewed and coded for elements of research interest. Use of computer software allows for rapid, consistent, and comprehensive analysis of large amounts of data. However, software packages for qualitative analysis are based on theoretical structures of which the researcher must be aware and use appropriately. QSR NUD∗IST (30Richards T.J. Richards L. QSR∗NUDIST. 4th ed. Qualitative Solutions and Research, Victoria, Australia1997Google Scholar) and The Ethnograph (31Seidel J. The Ethnograph. 5th ed. Qualis Research and Associates, Salt Lake City1998Google Scholar) code and organize data in a series of hierarchies (e.g., A is an overarching category which includes the general area of B, under which resides the subset of C), while ATLAS.ti (32Muhr T. ATLAS.ti. Ver. 4.1. Scientific Software Development, Berlin1997Google Scholar) and HyperRESEARCH (33Hesse-Biber C. HyperRESEARCH. Ver. 1.65. ResearchWare, Randolph, MA1998Google Scholar) connect data in a network of links with varying directions, intensity, and temporality of relationships (e.g., A makes B possible, B mitigates C, which can only occur after A is completed).Display of qualitative data for analysis presents a potential increase in "the primary risks to the adolescent participant from behavioral research [which] are potential embarrassment and/or disclosure of sensitive information to others" (34Society for Adolescent MedicineGuidelines for adolescent health research.J Adolesc Health. 1995; 17: 270-276Abstract Full Text PDF PubMed Scopus (74) Google Scholar). Because qualitative research methods obtain such detailed information, from life stories to relational maps to photos and videos, participants risk a loss of personal privacy even with names, hospital numbers, or other traditional identifiers stripped from the research record. The storage, display, and analysis of qualitative data must be accorded the highest security and researchers must respect the privacy of their research participants by keeping all research data and findings strictly confidential.Conclusion drawing is the derivation of findings from the theory-structured patterns of data. Ultimately, meaning is found by discovering universal relationships that can be found in the natural and social worlds. Often several evaluations of the data are required for findings to be identified and defined, and their importance clarified. Model building is based on the evolving domains of data that emerge from the analysis process. Qualitative methods can result in multiple findings, related or unrelated, hypothesized or unexpected, from within the theoretical bounds of the research question.With careful analysis, qualitative inquiry can move beyond associations to causality. "… we get inside the black box: we can understand not just that a particular thing happened, but how and why it happened" (35Huberman A.M. Miles M.B. Data management and analysis methods.in: Denzin N.K. Lincoln Y.S. Handbook of Qualitative Research. Sage, Thousand Oaks, CA1994: 428-444Google Scholar). Because qualitative methods can develop and modify theory as well as test it, they can examine influences and events longitudinally, test proposed causal mechanisms against emerging and evolving theory, and suggest how a temporal series of events may lead to a specific outcome. Although qualitative studies may be prospective in design, assessing causality must be retrospective because this is the only means of viewing the narrative sequence. Analysis is "a retrospective gathering of events into an account that makes the ending reasonable and believable … configur[ing] the events in such a way that the whole story becomes clear" (36Polkinghorne D.E. Narrative Knowing and the Human Sciences. State University of New York Press, Albany, NY1988Google Scholar).Verification is the process whereby interim findings and conclusions are verified and tested against the data, a process for which clear and complete data display is critical. All data collected are compared against each other and against the resultant conclusions to reinforce or contradict, enrich or rule out the researcher's analysis and emerging findings. The process of verification is ongoing, starting with data collection. The researcher adopts a reflective stance from which to learn, respond, and test findings against the incoming data stream, sometimes even checking perceptions and interpretations with study participants. This series of question-and-answer cycles, the iterative process, operates as an ongoing dialogue between data collection and analysis. The application of multiple research approaches, triangulation, is a concept derived from navigation where different bearings taken from different angles can precisely locate and characterize a feature in three dimensions (37Denzin N.K. The Research Act. 3rd ed. Prentice Hall, Englewood Cliffs, NJ1989Google Scholar). Triangulation can include the use of different data sources, researchers, theoretical frameworks, or methodologies directed toward the subject of research inquiry (38Morse J.M. Approaches to qualitative-quantitative methodological triangulation.Nurs Res. 1991; 40: 120-123Crossref PubMed Scopus (837) Google Scholar). The multidimensional findings that result can be used for internal verification and further modification of theory and conclusions. Conclusions drawn must be supported by ongoing iteration. Theories that emerge must be responsive to and modifiable by the analytical process. Ultimately they must explain, interpret, and be able to predict outcomes.Bias, reliability, and validity in qualitative researchScientific rigor is necessary for any research method to understand and accurately represent the phenomena it studies. The first level of scientific rigor is achieved by using a research method appropriate to the subject matter. Few would argue that a tightly controlled laboratory experiment is the ideal method for studying a social phenomenon. The key to achieving rigor in any research is to maximize methods that are most sensitive to the subject studied (39Kirk J. Miller M.L. Reliability and Validity in Qualitative Research. Sage, Beverly Hills1986Google Scholar).Bias is at work in all research. The influence of the researcher, the research question, and the method can never be completely neutralized. Qualitative researchers can recognize and minimize the potential for bias inherent in their methods. Facilitators of focus groups are trained to understand and compensate for their own personal biases, to promote equal participation of all group members, and to decrease the biasing effect of group dynamics created by power differentials, transference, and psychosexual issues. Participant observers are trained to fit in while minimizing their influence over routine occurrences. Acknowledging that subjectivity cannot be completely eliminated, the reflective, questioning process of qualitative investigation actively observes, evaluates, and counterbalances the influences that bias may have on data collection, analysis, or the drawing of conclusions (40Peshkin A. In search of subjectivity One's own.Educ Res. 1988; 17: 17-22Google Scholar).Reliability, the consistency with which a research method measures a phenomenon, is a concept from the physical sciences which becomes more difficult to operationalize when the subject of study, social reality, is in constant flux. Even though the social landscape may change, there are properties of human experience, perception, and response that are stable. In evaluating these properties, qualitative methods can achieve diachronic reliability, the reproducibility of an observation at two different times, and synchronic reliability, agreement of observations from two or more perspectives at the same time (39Kirk J. Miller M.L. Reliability and Validity in Qualitative Research. Sage, Beverly Hills1986Google Scholar). Qualitative methods achieve diachronic reliability through careful record keeping and, if possible, through multiple observations of a repeated event. Pretesting and standardizing instruments and training data collection personnel can significantly improve reliability. Transcripts, photos, and audio or video recordings can be analyzed and reanalyzed at different points in time. Synchronic reliability can be improved by evaluating data collection for areas of congruence among independent observations by different observers, known as interrater reliability (23Silverman D. Interpreting Qualitative Data. Sage, Thousand Oaks, CA1993Google Scholar). Standardized data collection protocols and comparisons of different observations are key tools for maximizing qualitative reliability.Validity, the accuracy with which a research finding represents the phenomenon studied, is improved in quantitative studies by eliminating or controlling for variables that might affect the phenomenon. Since the artificiality of the laboratory setting is inappropriate to social research, and since variables may be part of the process being studied rather than confounders, qualitative researchers must learn as much as possible about the study phenomenon and its environment, evaluating their findings for credibility and consistency. Multiple types of data, collected through different means and analyzed through different theoretical frameworks, are triangulated to situate the phenomenon studied and increase the validity of findings. Hypotheses, conclusion