Qualitative methods in physical therapy research: a form of disciplined inquiry.

Qualitative Methods in Physical Therapy Research: A Form of Disciplined Inquiry We must continue to build on our modest success of the last several years in fostering more research. And we must continue to strive to increase the scientific rigor, strengthen the theoretical underpinnings, and improve the social importance of research in physical therapy. [1]

These were the recommendations given by Eugene Michels, the American Physical Therapy Association's recently retired Associate Executive Vice President for Research and Education, to those who will lead the physical therapy profession into the future in its research efforts.

To achieve the goals advocated by Michels, we will have to use a broad range of research methodologies. Physical therapy as a field of study contains phenomena, events, problems, people, and processes that provide rich resources for different types of inquiry. A broad array of quantitative and qualitative research methods will enable us to understand the complexity and richness of both the clinical and educational environments in physical therapy and contribute to the development of a theoretical knowledge base. The discussion and use of qualitative methods in physical therapy research is continuing to develop and grow. [2-6] As we expand our knowledge base through the application of various research methodologies, it is important that we understand the research methodology we use and what makes the use of certain research methods disciplined.

Shulman has eloquently written about the importance of understanding method in research:

Method is the attribute which distinguishes research from mere observation and speculation. When adversaries argue about the nature of the world or the best approach to some particular human endeavor, we typically find ourselves evaluating their respective claims through examining the methods they used to reach their conclusions. There are few subjects that generate as much passion among scientists as arguments over method. This is not surprising, since scholars who can agree on matters of method can pursue research questions in a parallel fashion and then argue over the results of their respective investigations. However, if they do not agree even on some matters of research method, then their findings are likely to be commensurable. There will be no way to properly compare one inquiry with the other. [7]

Disciplined inquiry has been defined by Cronbach and Suppes as follows: "Disciplined inquiry has a quality that distinguishes it from other sources of opinion and belief. The disciplined inquiry is conducted and reported in such a way that the argument can be painstakingly examined." [8] Examination of the methods used to reach a stated conclusion is a critical element in critiquing research. Disciplined inquiry or research may occur through the use of quite different research methods that are based on different philosophies. The use of these different methods may be considered disciplined if the method can be examined.

To use qualitative methods in our research, we must thoroughly understand this methodology. To do this, we should have an understanding of the following: the philosophy underlying qualitative research, the application of qualitative methods (the variety of approaches, shared characteristics, and research techniques), and the applicability of reliability and validity concerns in qualitative research. This article will address each of these issues in hopes of enriching our understanding and application of qualitative research as a form of disciplined inquiry.

Research Models--Philosophical Base

Most of us are familiar with the basic assumptions underlying the most common research model, a natural science approach, often called the "scientific method." The philosophy underlying this conventional research approach has historically been called an empirical, or "logical positivist," approach and is a search for facts or causes of phenomena. The experimental or quasi-experimental design emphasizes empirical observations that lend themselves to analysis by mathematical tools. This approach, which involves quantitative research methods, also was embraced initially by social science researchers, starting with the work of theorists such as Emile Durkheim. [9,10]

Disappointment arose, however, when this research model did not adequately inform researchers about the complexity of human behavior and social interaction. This disappointment prompted German philosophers and historians to make a distinction between natural and human science. They argued that the methods of human science should take a phenomenological or hermeneutical approach. The term "hermeneutical" is derived from hermeneutikos', the Greek word for interpretation. In this approach, which frequently involves qualitative research methods, the researcher is committed to understanding human behavior from the individual's perspective. The philosophy underlying the phenomenological approach, or "interpretivist" position is that human beings are self-interpretative and that meaning, from the individual's perspective, has an essential place in the characterization of human behavior. What people believe to be true may be more important than objective reality, because people act on what they believe. [9-11]

Modern-day theorists of social science have criticized the natural science approach based on logical positivism, and some refer now to the post-positivist era in research in which they call for the use of multiple methods. [12,13] For example, Campbell states

Logical positivism failed to recognize that even at best, experimental research is equivocal and ambiguous in its relation both to the real physical processes involved and to scientific theory; and that attention to this equivocality calls for the use of multiple methods, none of them definitional, triangulating on causal processes which are imperfectly exemplified in our experimental treatments and measurement processes. [12]

Still others argue that positivist thinking remains the dominant culture in research [10,11] In reality, to argue the choice of a positivist approach (quantitative methods) versus an interpretivist approach (qualitative methods) is inexact and potentially nonproductive. The researcher, given a thorough understanding of the research methodology, should have free choice to select whatever method or combination of methods is best suited to addressing the identified problem situation.

Application of Qualitative

Methods

Variety of Approaches

Qualitative research comprises a multitude of approaches. [11,14] These approaches are viewed and evaluated by different criteria. For example, naturalistic inquiry, described by Lincoln and Guba, [10] focuses on having the researcher maintain complete separation from the positivist philosophy. They want naturalistic methods to remain natural; that is, the investigator avoids manipulating the research outcomes, and multiple realities are allowed. [10] Eisner advocates an artistically oriented research approach for educational evaluation called "educational connoisseurship and criticism." n15 He defines connoisseurship as the art of appreciation and criticism as the art of disclosure. Connoisseurship requires an awareness and understanding of the phenomena observed, and criticism involves description, interpretation, and evaluation. n15 Still other qualitative approaches use more of a sociological approach involving field-based methods such as interviews and observation in formulating case studies. [16]

Ethnography is perhaps the most common term used when talking about qualitative research. The literal definition of ethnography is "writing about nations." Ethnography, as a systematic research approach, was first used by anthropologists to study cultures. True ethnographies are analytic descriptions or reconstructions of intact cultural scenes and groups. The ethnographer presents the shared beliefs, practices, artifacts, folk knowledge, and behavior of a certain group of people. Ethnography can also take different forms depending on the type of ethnography (eg, holistic, interpretative, generic, or quasi). For example, generic and quasi ethnography refer to the use of ethnographic methods in data collection and analysis, but the study may focus on a small subset of a larger cultural system (a "microethnography") or a short observation period. [14,17]

Shared Characteristics

Even though there are different approaches within the qualitative research tradition, there are shared characteristics that stem from the underlying philosophy of phenomenology. These shared characteristics may be better understood by contrasting them to the general characteristics of the natural science approach to research (Tab. 1). Qualitative research evolved from the philosophical base of phenomenology, where an interpretive approach to research is encouraged. The setting is the natural setting. Whatever happens in this natural setting, or context, is critical to the research and may lead to expansion of a flexible, dynamic research process and more than one interpretation of events. In contrast, quantitative methods are often used in the natural science approach, where the focus is frequently to generate and verify objective knowledge. A predefined design, control of possible confounding variables, and a laboratory setting allow the researcher to reduce outside interference and make verifiable inferences about systematic observations. The characteristics in Table 1 have been contrasted starkly to focus the reader, but realize that many of these characteristics could be seen on a continuum.

The shared characteristics in the qualitative research tradition have underlying concepts that guide their use. What follows here is a discussion of several of those general characteristics.

1. Research is done in a natural setting. The researcher collects data by spending time in the field or at the setting under study. The researcher wants data collected in context, that is, in the natural setting. The focus is on the understanding of human behavior from the individual's or insider's perspective. Anthropologists use the term "emic" when describing this insider perspective. They believe cultural behavior should be studied and categorized from processes of the people studied and that meaning is subject-generated, or emic, not researcher-imposed, or "etic." [16]

2. The researcher is the instrument. The researcher is the primary data-gathering tool. The researcher can be both observer and participant, depending on the investigative setting. Direct, personal contact with people is an essential part of the research. Because the researcher is the data-collection instrument, identifying and stating the sources of potential bias and subjectivity in the researcher's own make-up is critical to the quality of the work. [16,17]

3. The research is both descriptive and interpretive. Interpretation is critical to the research process. The researcher continually strives to make sense and meaning out of the research setting. Narrative description can be done in both quantitative and qualitative work, but it is interpretation that makes the research meaningful. [18] While in the field, the researcher uses methods such as long-term participation, careful recording of field notes, interviews, collecting artifacts, and analytic reflection to obtain data that are both descriptive and interpretive. [19] "Thick description" is a term used to represent the idea that the description is rich and allows meaningful interpretation, not that the description is thick in terms of volume. [17]

4. The research process is systematic, yet flexible. Research is undertaken with a plan or some guiding principles. The researcher usually enters the field with conceptual ides. These concepts come from the researcher's own experience, a review of the literature, and in some instances, preliminary observations. The concepts are organized into a conceptual framework that is used to guide the direction of data gathering and analysis. Thus, the framework provides the context for the research questions, and it justifies and organizaes the approach to the study's purpose. The framework specifies who and what will be studied and what the presumed relationships among these variables may be. A framework may be represented graphically, in narrative form, or as a combination of both. The Figure provides an overview of questions and issues to consider in planning an observational study, where the conceptual framework is central to the study. [20] Once the problem situation is identified, the conceptual framework can be formulated. The framework is "grounded," or has an underlying structure created from several sources. This initial underlying structure provides the researcher with a strong foundation for posing research questions or hypotheses. Next, the researcher creates a draft conceptual framework including identification of the dimensions (eg, individuals, interactions, settings) that are essential to answering the posed questions and other more practical considerations (Figure). Initial revision of the framework occurs after preliminary observations and as the research proceeds when researcher insights and new data may be included in a revised framework. [17,19,21]

5. The research works on a generative, not a verificative, dimension. The goal of the research is usually not verification of set, predetermined, or a priori hypotheses but rather generation of concepts that emerge from the data. Inference and speculation are essential components throughout the research process. If there is hypothesis formulation, it is likely to be successive as the researcher formulates and tests successive explanations for what is observed. The research also can be guided by research questions of interest that may change as the data collection proceeds. The cognitive processes of induction and deduction should be in constant interaction in qualitative research. [17,19,21]

6. Data collection and data analysis are interdependent. Analysis of the data does not occur following data collection; rather, some analysis proceeds throughout the data-collection phase. Both data collection and data analysis can inform each other. For example, some analysis of data may enable the researcher to return to the field to collect new data, validate findings, or refine some aspect of the research. [9,16,17]

7. Data analysis involves a process of higher level synthesis. A major dilemma for qualitative researchers is the amount of data collected. Data analysis, therefore, must include procedures for data reduction. Although some of the data may be quantifiable, much of it is verbal and must be interpreted and reduced. The researcher uses analytic processes such as comparing, contrasting, or aggregating data to try and find patterns or themes that interpret the findings at a higher conceptual level. Frequently, data analysis will involve some form of content analysis where coding schemes are developed from the data, from categories in the conceptual framework, or from theory. [17,21]

8. Theory may have a multidimensional role. Theory helps describe and explain what the researcher finds by transcending the data. Theories are conceptually based and allow researchers to transfer the meaning of their fndings to other settings. Goetz and LeCompte discuss three ways theories can assist in the development of qualitative research problems or questions. [17] An established theory can be used as the source for generating a research question or designing the conceptual framework. A second use of theory is how it can influence the questions a researcher poses. For example, if a researcher has a background in psychology or sociology, he or she would be likely to apply theories from these disciplines when posing questions and interpreting the data. A third use of theory is explanation of some aspect of the phenomena that emerge from the data. In other words, a theory first thought to be useful and part of the conceptual framework may not be as helpful as the data are collected and interpreted. Thus, the original theory is discarded and another theory or related aspects of several theories are presented to account for the phenomena observed. Another possible role of theory in qualitative research is to develop grounded theory. Here the researcher works more inductively and develops theory from systematic analysis of the qualitative data. Grounded theory is a systematic research approach for the collection and analysis of qualitative data. The purpose is to generate explanatory theory that furthers our understanding of social or psychological phenomena. [22]

Qualitative Research Techniques

Participant observation. Participant observation is the primary tool the qualitative researcher uses. Some people describe participant observation as a schizophrenic act. The participant observer comes to a social situation to engage in ongoing activities and to observe the people and physical aspects of the situation. The researcher, therefore, tries to understand meanings of a group of people by simultaneously participating in the activities and observing the other participants and reflecting on what is seen and heard. [23]

The setting often dictates how much of a participant or observer a researcher can be. For example, in some settings it may not be possible to participate, only to observe. The first obstacle the researcher must face is access to the research site and then establishing trust and rapport with those who are being observed. Interpersonal skills are essential to this type of field work. Field notes serve as the written account of what the researcher sees, hears, experiences, and thinks. Raw or rough field notes are what the researcher records while in the field. They include dialogue or what was said, as well as the researcher's observations and perceptions. Researcher comments or potential interpretations are identified as such, and care must be taken not to confuse raw data with interpretations. The ideal situation is to have the researcher sit down immediately following an observation and redo the rough set of field notes, filling in more detail from immediate recall. The researcher must provide as much accurate detail as possible, reconstruct relevant dialogue, and keep focused on the topic or research questions. Table 2 provides some additional guidelines for the construction of field notes. [24]

Interviews. Interviews are conversations with a purpose. There are several different interview formats that can be used in qualitative research. Three general categories of interviews are listed and defined in Table 3. [25] A new researcher may elect to use the most structured interview approach--a scheduled, standardized interview. In this form of interview, there is no variation in questions, sequence, or follow-up questions, so the researcher may devote full attention to the mechanics of the interview.

The researcher's goals are to establish rapport with the interviewee and to discover his or her experience. Audiotaping interviews is a great help to the researcher, because it allows time to be attentive to the interviewee and to probe for further explanation, as necessary. In some settings, a researcher may cultivate a close relationship with one or two individuals who may serve as key informants, or the observer's observers. These individuals may be interviewed several times throughout the course of data collection. [23,24]

Artifacts. Human beings make and use things. Artifacts can provide evidence for the researcher that indicates people's feelings, experiences, and knowledge. These materials can be the manifestation of the beliefs and behaviors that make up a culture. These materials can include things such as policies and procedures, memorandums or letters, documents, bulletin boards, and diaries. The researcher is usually involved in locating the artifacts and in identifying, analyzing, and evaluating the material. The researcher will want to know who uses the artifact and how, under what conditions, and for what purpose it is used. [17] Qualitative researchers may also use tools such as photographs and videotapes to collect nonverbal data or data that provide an accurate representation of the sequence of events. [24]

Reliability and Validity

Concerns

Some scholars have criticized qualitative research methods for being unreliable, resulting in research that lacks validity and generalizability. Ignoring these basic tenets of research poses a threat to the credibility of qualitative research as a form of disciplined inquiry. Applying the concepts of reliability and validity to qualitative work, however, requires us to use techniques that are somewhat different from those used in a natural science approach.

Reliability

Reliability, or the extent to which a study can be replicated, generally requires the researcher to use the same method to obtain similar results. This replication poses a problem for qualitative researchers, where the emphasis is on human behavior in the natural setting and differences in a researcher's background and changing circumstances may provide for multiple concepts of reality. Both the kind of data and the research process limit the researcher's ability to use techniques that would control or manipulate the natural occurrence of events. A unique situation is difficult to reconstruct. [16,17]

Reliability in qualitative measurement is addressed through the use of low-inference descriptors, that is, verbatim accounts of interviews, direct quotations in field notes, and mechanical methods of recording data. This type of rich data is generally considered to be the most credible. The use of multiple researchers and establishment of interobserver, interinterviewer, interrecorder, or interanalyst reliability in the data-collection and analysis phases also guards against threats to reliability. Reliability is relevant and important for qualitative measurement because it does help ensure replicability and provide a necessary prerequisite for validity. Peer examination and corroboration of findings by other researchers who work in simular settings can be helpful in verifying that a single researcher remained consistent in the data-collection and analysis process. Repeated observations or interviews can be used to lend credence to stability of the data, if that is an issue. The researcher also should carefully delineate the method so other researchers may use the original work as an "operating manual" when they replicate the study. [17,26]

Validity

Establishing validity requires the researcher to demonstrate that what was observed and measured is believable and credible. Dawson suggests three general validation methods: 1) confirm or disconfirm findings through the use of a systematic coding and analysis process and the use of triangulation methods, 2) use strategies to rule out specific sources of invalidity, and 3) create research conditions under which valid information will be collected. [27]

In the first validation method, the emphasis is on the researcher's deliberate attempt to find negative evidence or cases as the investigation proceeds. This evidence, in turn, can be used to refine the research questions or raise the data analysis to a higher level that includes negative, or "divergent," findings. Triangulation, frequently defined as the use of multiple data-collection methods, is more than that. Triangulation can involve the use of multiple methods of data collection, data sources, researchers, or theories. The logic in triangulation is that no single method or meaning ever adequately addresses the problem. The use of triangulation should prevent the researcher from accepting too readily the validity of initial impressions. [17,19,25]

Dawson's second validation method, rule out sources of invalidity, is similar to the concepts of internal and external validity. Several of the threats to internal validity, posed by Campbell and Stanley for experimental design, [28] are applicable to qualitative methods but with somewhat different interpretations (Tab. 4). For example, concepts such as the effect of history or mortality are considered part of the normal process in a natural setting; therefore, strategies are not used to rule out these effects. Potential threats from an observer effect, that is possible bias or distortion of data collected through participant observation and interview data, can be addressed by a thorough data-collect on process, including sufficient time spent in the field, establishing trust with the individuals being observed, use of multiple data sources, and the collection of confirming and disconfirming evidence (Tab. 4).

Threats to external validity, or how representative or generalizable the study is, also can be addressed in qualitative research. Generalization in qualitative research is limited by the extent to which comparable studies with comparable samples can be found and analyzed for similarities. If the researcher establishes that the components of the study such as the units of analysis, concepts generated, sample characteristics, and setting are sufficiently described and defined, then the study has comparability (ie, other researchers could use the study as a basis for comparison). A second characteristic for external validity, translatability, requires that the researcher use methods, definitions, and a theoretical framework that are explicitly identified and understood by other researchers and could be translated to another study. [17]

Dawson's third validation method, creation of favorable conditions for collection of valid data, assumes that the researcher will establish sufficient access to the field site, spend sufficient time in the field, and establish trusting relationships with those individuals from whom the data are gathered. [17,27]

Many concepts related to qualitative methodology have been covered rapidly and in insufficient detail in this article. The short annotated bibliography shown in the Appendix should provide the new qualitative researcher with additional information about using qualitative research methods.

Conclusion

The use of qualitative methods in physical therapy research offers researchers the opportunity to better understand the context of clinical practice and professional education. As a developing profession working toward defining and validating a theoretical body of knowledge, application of a variety of research methods provides us with the opportunity to establish a broad-based, integrated knowledge base. Thorough researcher knowledge and understanding of the methodology applied in our research will lead to disciplined inquiry. This article presented background information that I hope will lead to disciplined inquiry in the application of qualitative methods. We must remember the advice of Shulman when selecting our research methods:

Selecting the method most appropriate for a particular disciplined inquiry is one of the most important, and difficult, responsibilities of a researcher. The choice requires an act of judgment, grounded both in knowledge of methodology and the substantive area of investigation. [7]

Acknowledgments

I am indebted to Lea Shulman, PhD, and Milbrey McLaughlin, PhD, Stanford University, for teaching me about qualitative research. I also thank Rebecca L Craik, PhD, Beverly J Schmoll, PhD, and Kay F Shepard, PhD, for their significant, useful suggestions in the preparation of this manuscript.

References

[1] Michels E: Physical therapy research: The crucial challenge for the future. Progress Report of the American Physical Therapy Association 17(6):6, 1988

[2] Scully RM, Shepard KF: Clinical teaching in physical therapy education: An ethnographic study. Phys Ther 63:349-358, 1983

[3] Schmoll BJ: Ethnographic inquiry in clinical settings. Phys Ther 67:1895-1897, 1987

[4] Shepard KF: Qualitative and quantitative research in clinical practice. Phys Ther 67:1891-1894, 1984

[5] Hulme JB, Bach BW, Lewis JW: Communication between physicians and physical therapists. Phys Ther 68:26-31, 1988

[6] Jensen GM: The work of accreditation onsite evaluators: Enhancing the development of a profession. Phys Ther 68:1517-1525, 1988

[7] Shulman L: Disciplines of inquiry in education: An overview. Educational Researcher 10(6):5-12, 1981

[8] Cronbach LJ, Suppes P (eds): Research for Tomorrow's Schools: Disciplined Inquiry for Education. New York, NY, Macmillan Publishing Co, 1969, p 15

[9] Taylor SJ, Bogdan R: Introduction to Qualitative Research Methods: The Search for Meanings, ed 2. New York, NY, John Wiley & Sons Inc, 1984, pp 1-14

[10] Lincoln YS, Guba EG: Naturalistic Inquiry. Beverly Hills, CA, Sage Publications Co, 1985, pp 14-46

[11] Fetterman D: Qualitative approaches in evaluating education. Educational Researcher 17(8):17-22, 1988

[12] Campbell D: Can we be scientific in applied social science? In Conner RF (ed): Evaluation Studies Review Annual. Beverly Hills, CA, Sage Publications Co, 1984, vol 9, pp 1-40

[13] Howe K: Two dogmas of educational research. Educational Researcher 14(8):10-18, 1985

[14] Jacobs E: Clarifying qualitative research: A focus on traditions. Educational Researcher 17(1):16-24, 1988

[15] Eisner E: The Art of Educational Evaluation: A Personal View. New York, NY, Taylor & Francis, New York, 1985

[16] Patton MQ: Qualitative Evaluation Methods. Beverly Hills, CA, Sage Publications Co, 1980

[17] Goetz J, LeCompte M: Ethnography and Qualitative Design in Educational Research. New York, NY, Academic Press Inc, 1984

[18] Guba EG, Lincoln YS: Effective Evaluation: Improving the Usefulness of Evaluation Results Through Responsive and Naturalistic Approaches. San Francisco, CA, Jossey-Bass Inc, Publishers, 1981, pp 53-83

[19] Erickson F: Qualitative methods in research on teaching. In Wittrock MC (ed): Handbook of Research on Teaching, ed 3. New York, NY, Macmillan Publishing Co, 1986, pp 119-161

[20] Evertson CM, Green J: Observation as inquiry and method. In Wittrock MC (ed): Handbook of Research on Teaching, ed 3. New York, NY, Macmillan Publishing Co, 1986, pp 204, 205

[21] Miles MB, Huberman AM: Qualitative Data Analysis: A Sourcebook of New Methods. Beverly Hills, CA, Sage Publications Co, 1984, pp 17-25

[22] Strauss AL: Qualitative Analysis for Social Scientists. New York, NY, Cambridge University Press, 1987

[23] McCall GJ, Simmons JL (eds): Issues in Participant Observation: A Text and Reader. Reading, MA, Addison-Wesley Publishing Co Inc, 1969

[24] Bogdan RC, Biklen SK: Qualitative Research for Education: An Introduction to Theory and Methods. Boston, MA, Allyn & Bacon Inc, 1982, pp 74-96

[25] Denzin NK: The Research Act: A Theoretical Introduction to Sociological Methods, ed 2. New York, NY, McGraw-Hill Inc, 1978

[26] Goodwin L, Goodwin W: Are reliability and validity "relevant" in qualitative evaluation research? Evaluation and the Health Professions 7:413-426, 1984

[27] Dawson J: Qualitative research findings: What We Do to Improve and Estimate Validity. Read at the Annual Meeting of the American Educational Research Association, New York, NY, April 19-23, 1982

[28] Campbell DT, Stanley JC: Experimental and Quasi-Experimental Designs for Research. Palo Alto, CA, Houghton Mifflin Co, 1966

G Jensen, PhD, PT, is Assistant Professor, Division of Physical Therapy, School of Health Related Professions, The University of Alabama at Birmingham, Birmingham, AL 35294 (USA).

This article is adapted from a presentation at the American Physical Therapy Association-Canadian Physiotherapy Association Joint Congress, Las Vegas, NV, June 12-16, 1988.

This article was submitted July 21, 1988; was with the author for revision for 11 weeks; and was accepted February 2, 1989.