
Creswell opens Chapter 10 with this question. In qualitative research there are multiple perspectives, voices, and accounts that figure into the final product, how then is the product judged as acceptable?
Creswell summarizes multiple authors' work in Chapter 10 and discusses various approaches to validation, reliability, and evaluation.
Of course one must be cautious so as not to evaluate qualitative research with a quantitative lens.
Several times Creswell notes that thick description is necessary for all of the approaches.
I particularly appreciated, possibly due to my strength of "strategic," the citation by Creswell on page 204 where Eisner (1991) stated "The researcher compiles bits and pieces of evidence to formulate a "compelling whole." At this stage, the researcher looks for recurring behaviors or actions and considers disconfirming evidence and contrary interpretations.
I believe one of my strengths is the ability to see connections between separate works. Of course in qualitative work this can be both a blessing and a curse! These seeming connections would be the exact fodder for this topic!
Thus there would be a high need for me to use consensual validation so that I might solicit the opinion of others to assure I am not making up things!
Lather (1991) presented four types of validation: triangulation, construct validation, face validation, and catalytic validation. In 1993, Lather presented four additional frames (related to feminist research): ironic, paralogic, rhizomtic, and situated validation (p. 205).
On page 206 Creswell writes "a synthesis of validation perspectives" and this made me wonder if this is what a "meta-analysis" is?
Creswell presents eight validation strategies:
- Prolonged engagement in the field
- Triangulation
- Peer review
- Refining hypotheses until all cases fit
- Clarifying researcher bias from the outset
- Member checking (the most critical technique for establishing credibility-from the famous Lincoln & Guba)
- Rich, thick descriptions
- External Audits
Creswell recommends engaging in at least two of these strategies in any study.
Evaluation Criteria:
Howe and Eisenhardt suggest five standards to be applied to all research:
- Assess a study in terms of whether the research questions drive the data collection and analysis rather than the reverse
- Examine the extent to which the data collection and analysis techniques are competently applied in a technical sense
- Are the researchers' assumptions made explicit
- Does the study have overall warrant
- Is there value for informing and improving practice and in protecting confidentiality
Overall I am surprised at the thinness of this chapter. After having spent 200 pages in the teaching of how to gather data, how to design a study and so forth, I am surprised that only 20 pages are dedicated to validity. I surmise that Creswell believes if the researcher spends considerable time in the data from the first 200 pages, the study will be well done enough that validating and evaluating is a more smooth process!