BOOK: CREDIBILITY, DEPENDABILITY, Confirmability, RELIABILITY
The SAGE Encyclopedia of qualitative research methods
CREDIBILITY, DEPENDABILITY, Confirmability, RELIABILITY
Reference:
Jensen, Devon, (2008). Credibility. In L. M. Given (Ed.), The SAGE Encyclopedia of qualitative research methods (pp. 890-890). Thousand Oaks, CA: SAGE.
CREDIBILITY
One of the responsibilities of any qualitative researcher is to create a high level of consistency in the article. For example, the readers and research participants should see why a particular research model was used and why the participants were selected for the study. The data analysis process should also reveal a believable link between what the participants expressed and the themes and codes that emerge. The accuracy of this process for both the readers and participants creates a measure of credibility to the research project. As such, credibility can be defined as the methodological procedures and sources used to establish a high level of harmony between the participants’ expressions and the researcher's interpretations of them.
The basic notion with credibility is that both the readers and participants must be able to look at the research design and have it make sense to them. Questions for the researcher to consider in relation to credibility include the following: Were the appropriate participants selected for the topic? Was the appropriate data collection methodology used? Were participant responses open, complete, and truthful?
Here is an example of how these items of credibility could be addressed when putting together a research study on the devaluing of nurses by doctors using a survey instrument and focus groups. The study would lack credibility if nurses were the only participants. It would be more credible by including nurses, registered nurses, and doctors. A closed question survey would lack credibility because the researcher is defining the context through the survey items rather than allowing the participants to define them. A survey instrument with both closed and open-ended items would be more credible. Credibility of the study would be lacking if a focus group of two doctors and two nurses was the only means for the participants to discuss the topic because nurses might not provide open and truthful information in the presence of doctors. In this focus group setting, nurses also would not be able to provide complete information because of the power dynamic that exists between doctors and nurses. Thus, this methodology for collecting the data would have low credibility because it has a very narrow means of illuminating the context under study. The credibility of the study could be enhanced by having a larger focus group, introducing private interviews with the participants, and then providing opportunities for follow-up interviews as necessary.
The researcher can use the following methodological procedures to increase credibility:
Time: Establish enough contact with the participants and the context to get the information one needs.
Angles: Look at the data from different perspectives and viewpoints to get a holistic picture of the environment.
Colleagues: Use support networks knowledgeable in the area to review and critique the research and data analysis findings.
Triangulation: Seek out multiple sources of data and use multiple data-gathering techniques.
Member checks: Use the participants to make sure that the data analysis is accurate and consistent with their beliefs and perceptions of the context being studied.
Jensen, Devon, (2008). Confirmability. In L. M. Given (Ed.), The SAGE Encyclopedia of qualitative research methods (pp. 890-890). Thousand Oaks, CA: SAGE.
DevonJensen
Confirmability
In qualitative research, the actions and perceptions of participants are analyzed for their expressions of meaning within a given context. Consistent with the practices of the selected qualitative methodology used, the researcher then interprets the participant expressions through a coding or meaning-making process. In this coding process, the researcher is looking for messages that are consistent with, confirm, or expand on current knowledge and theory. From these insights, the researcher is then able to make statements about the context under study. In so doing, additional processes must be incorporated into the research design that verifies the truthfulness or meaning being asserted in the study. This is called confirmability.
Confirmability is often equated with reliability and objectivity in quantitative research. Reliability and objectivity are measures of the accuracy of the truth or meaning being expressed in the study. The epistemological function of this process is to suggest that truth and meaning are reliable only to the point where they can be verified as more than just a singular event peculiar to that specific research endeavor and researcher. This is essential because it is an academic process that moves the research beyond a one-time event into a framework where meaning and truth can be used to build on, expand, or create theory.
Confirmability is an accurate means through which to verify the two basic goals of qualitative research: (1) to understand a phenomenon from the perspective of the research participants and (2) to understand the meanings people give to their experiences. Confirmability is concerned with providing evidence that the researcher's interpretations of participants’ constructions are rooted in the participants’ constructions and also that data analysis and the resulting findings and conclusions can be verified as reflective of and grounded in the participants’ perceptions. In essence, confirmability can be expressed as the degree to which the results of the study are based on the research purpose and not altered due to researcher bias.
Although confirmability does not deny that each researcher will bring a unique perspective to the study, it requires that the researcher account for any biases by being up front and open about them and use the appropriate qualitative methodological practices to respond to those biases. For example, a researcher using discourse analysis can have multiple coders of the same data to establish a measure of the consistency in the coding of themes. The researcher can also make the research process as transparent as possible by clearly describing how data were collected and analyzed and possibly offering examples of the coding process in the final document. Confirmability can also be expressed through an audit trail where an independent reviewer is allowed to verify the research process and interpretations of the data as consistent on both the literature and methodological levels. Selected participants can also be asked to review some of the coding and meaning-making process to determine whether the researcher's interpretations are consistent with their perceptions.
DevonJensen
Jensen, Devon, (2008). Dependability. In L. M. Given (Ed.), The SAGE Encyclopedia of qualitative research methods (pp. 890-890). Thousand Oaks, CA: SAGE.
DEPENDABILITY
One of the challenges of working within a qualitative context is the variability of the environment. Through extensive literature reviews and experience in the context, a researcher can create a pretty good theoretical understanding of what the environment will be like and then design appropriate methodologies for studying it. Once the researcher is out in the field, he or she may find it to be quite different from what was expected. This could affect research procedures such as what types of interview questions are asked and how many interviews are conducted. Dependability in a qualitative study recognizes that the research context is evolving and that it cannot be completely understood a priori as a singular moment in time. Dependability accounts for these issues through relevant methodologies.
A catalyst for research is the desire to have the study affect theory and/or practice on a broad scale. This means that the results should be consistently linked to revealed data and that the findings should be an accurate expression of the meanings intended by the participants. For this to happen, there must be a research infrastructure to support a repetition or replication of the study that will have similar results. This condition, which is equivalent to reliability in quantitative research, requires that the researcher supply adequate and relevant methodological information to enable others to replicate the study. If a study design is so unique and specific that it cannot be replicated, the research will have limited impact beyond the context of the study and the dependability of the study design will be affected. An example of this would be interviewing the last two clients of a social services program that is ending. It would be very hard to repeat this study because there are no more clients and the program is completed.
Dependability also addresses the fact that the research context is open to change and variation. The researcher must be conscious of change and must track all of the nuances that differ from the design in the proposal. As part of this, the researcher should track the alterations to the research design made necessary by the changing context. This could include changes in methodology such as increasing the number of interviews required, tracking nonverbal cues as well as spoken text, including document analysis, increasing intercoder reliability by having more coders, and/or increasing contact time in the environment from 1 week to 2 weeks. Tracking this process is called an inquiry audit. An external agent will review the researcher's fieldnotes and log book to ensure that the various changes in the research design have both methodological and theoretical foundations and are linked to the revealed data. The transparency and relevancy of this process will increase the dependability of the study.
DevonJensen
RELIABILITY
Reliability, in the field of research, is broadly described as the dependability, consistency, and/or repeatability of a project's data collection, interpretation, and/or analysis. Reliability is viewed very differently in qualitative research from how it is viewed in quantitative research. In the quantitative domain, reliability is specifically characterized as the extent to which multiple researchers arrive at similar results when they engage in the same study using identical procedures. In these conditions, differences in results are described as measurement error. Therefore, from a quantitative perspective, reliability is specifically defined, sought, and measured, and it is accepted as an essential indicator of a study's quality (along with measures of validity and generalizability).
In contrast, because of the paradigmatic and methodological diversity of approaches that comprise the field, reliability has not been described with such uniformity in qualitative research. Whereas many qualitative researchers describe parallel concepts such as credibility, dependability, confirmability, and consistency as appropriate qualitative correlates to reliability, others avoid the purposeful quest for reliability altogether. Those who overtly seek credibility and dependability often assert that such aims support the rigor of qualitative work and ensure that studies avoid “haphazard” subjectivity. Three of the commonly cited indicators of credibility and dependability are methodological coherence (the appropriate and thorough collection, analysis, and interpretation of data), researcher responsiveness (the early and ongoing verification of findings and analyses with study participants), and audit trails (a transparent description of all procedures and issues relative to the research project). Such strategies are commonly employed by qualitative researchers to demonstrate systematic attention to reliability-related issues.
On the other hand, some have asserted that purposeful attempts to demonstrate reliability are counterintuitive to much of the work that emanates from the qualitative domain. They point to the interpretive subjective nature of qualitative work as a defining hallmark of the field—one that can be undermined by rigid reliability concerns. At the heart of this position is the notion of reflexivity. Whereas quantitative researchers (and some qualitative researchers) attempt to minimize—indeed eliminate—researcher effects so as to maintain objectivity, most qualitative researchers embrace the notion of reflexivity—the idea that researchers' backgrounds, interests, skills, and biases necessarily play unique roles in the framing of studies and in the collection, analysis, and interpretation of data. Researchers are seen as visible, biased integral players in the process. This depiction of “researcher as instrument” in the project flows naturally with the claim that the richness and meaningfulness of qualitative research is largely dependent on its creativity and originality. Rather than seeking to standardize interview/testing procedures so that any researcher (who is detached and neutral) might gain the same results, the unique identities of both researchers and research participants are transparently identified and purposefully centered. Repeatability, from this perspective, is neither desired nor possible.
Therefore, it is evident that reliability, like many other concepts in qualitative research, is best approached on a case-by-case basis. Although many specific steps can be taken to support the credibility of one's research, such efforts should not compromise the deeper methodological and paradigmatic meanings that underpin this work.
PeterMiller
Comments
Post a Comment