In order to have any effect upon educational theory or practice educational research studies must be rigorous and present results that are acceptable to other educators and researchers (Merriam, 1998). To accomplish this task, studies must be of high quality and results must be trustworthy and dependable. Validity and reliability have been the traditional standards used in quantitative research studies to judge quality. Because their assumptions about reality differ from those of quantitative researchers, qualitative researchers have argued for the use of alternative criteria in judging research quality (Creswell, 1998; Trochim, 2006).
Claiming that they provide a better reflection of underlying assumptions found in qualitative research, Guba and Lincoln (1985, as cited in Creswell, 1998 and Trochim, 2006) propose that the alternative terms ≴credibility”, “transferability”, “dependability”, and “confirmability” be used in place of those commonly found in discussions of quantitative research. Credibility is analogous to “internal validity” and refers to establishing results that are credible or believable from a participant's perspective (Trochim, 2006). Analogous to “external validity” is the term transferability, which involves the ability of results to be generalized or transferred to another context or setting.
“Reliability”, as a quantitative criteria, refers to the ability of research results to be replicated or repeated and is based upon the assumption of the existence of a single reality (Merriam, 1998). In contrast, Merriam contends that with qualitative research the focus is on understanding and explaining the world as others have experienced it and it is assumed that multiple realities exist. Thus, there is no benchmark upon which to measure repeatability or upon which to establish traditional reliability. Guba and Lincoln (1985, as cited in Creswell, 1998 and Trochim, 2006) alternatively propose the term dependability, which is intended to be analogous to reliability. Rather than focusing on the ability of others to replicate results, the intent becomes having others concur that the results are sensible and consistent with the collected data (Merriam, 1998). According to Trochim (2006):
The idea of dependability, on the other hand, emphasizes the need for the researcher to account for the ever- changing context within which research occurs. The research is responsible for describing the changes that occur in the setting and how these changes affected the way the research approached the study.
To end with, confirmability, the qualitative equivalent of objectivity, is similar to dependability, in that it refers to the ability of results to be confirmed or substantiated by others.
Credibility, transferability, and dependability for qualitative research can be established using “triangulation”, “member check”, and “rich, thick description”. Triangulation refers to the use of multiple sources and methods for collecting data (Creswell, 1998; Leedy & Ormrod, 2001; Merriam, 1998). Sources of data for triangulation can include student interviews, teacher and student journals, assignment scores , as well as questionnaires. Triangulation sheds light upon common themes found in different sources (Creswell, 1998) and strengthens dependability and credibility (Merriam, 1998). Member check involves sharing the researcher':s interpretations of data obtained from various sources with the participants from which they were gathered and determining if they feel that the results are credible (Creswell, 1998; Leedy & Ormrod, 2001; Merriam, 1998). Finally, rich, thick description involves writing out detailed descriptions of the participants and setting under study (Creswell, 1998; Merriam, 1998). The purpose here is to address transferability and allow the reader to determine if the finds can be transferred to other contexts.
When using questionnaires, validity and reliability can be accomplished through using pre-existing questionnaires that have construct validity, internal consistency, and reliability pre-established by its designers. Construct validity for questionnaires are generally established through using confirmatory factor analyses, which test how close input correlations could be reproduced. Predictive validity can be shown by demonstrating scale correlation with final grades. Finally, internal consistency and reliability are demonstrated through measurement of ≴squared correlation between observed scores and true scores” (Yu, 2001). Results should generate moderate to high coefficient alphas. If the questionnaire that will be used is modified then reliability and validity will need to be reestablished. To accomplish this task, a pilot study should be conducted in which data obtained from the modified questionnaire is analyzed and compared to the reliability and validity data of the original instrument. Furthermore, data obtained from the modified questionnaire should be triangulated with data from other observations and a member check. A member check involves sharing and discussing the results of a questionnaire or survey with selected participants who have completed it.
Creswell, J.W. (1998). Qualitative inquiry and research design: Choosing among five traditions.
Thousand Oaks: Sage Publications.
Leedy, P.D. & Ormrod, J.E. (2001). Practical research: Planning and design (7th ed.). Upper
Saddle River, New Jersy: Merrill Prentice Hall.
Merriam, S.B. (1998). Qualitative research and case study applications in education.
San Francisco: Jossey-Bass Publishers.
Trochim, W.M. (2006). The research methods knowledge base (2nd ed.). Retrieved from
Yu, C.H. (2001). An introduction to computing and interpreting Cronbach Coefficient Alpha in
SAS. Retrieved from http://www2.sas.com/proceedings/sugi26/p246-26.pdf.