I have conducted grounded theory research via mixed methods approach. The study was conducted in two phases. First phase was purely grounded and so ended up while developing core categories and base themes. In second phase, I constructed hypothesis from core categories and conducted a survey to validate my qualitative base themes. The comparative analysis of qulitative and quantitative data ended up in emergence of sub-themes. Now to construct final substantive theory, I have created a causal model (via subjective interpretation), showing interdependency and connectivity between sub-themes.
Thanks in Advance!
You can, IF you also have evidence for the causal CONNECTIONS among the categories. This would pretty much need to be qualitative evidence from the interviews, showing that participants themselves saw these connections (this is what Anselm Strauss called "axial coding"). Otherwise, you can present this as a purely hypothetical model. I've attached two papers that each have more information on parts of this argument.
In tradition of Grounded Theory starting with theoretical sampling, means the research purpose is not on generalizing to a population or to other instances, but to build a theory or identify a causal mechanism. Instances are chosen either on their similarity to each other or their difference to each other (literal or theoretical replication). This leads to a differentiation of theory.
Fundament of both approaches is the constant comparison of all data, in order to monitor relevant factors. On this basis, it remains possible that some of the underlying theoretical modellings are applicable to other groups. Any further external generalization however, need to be tested by further research.Therefore a causal connection can be demonstrated on basis of your data and accordingly represents a specific narrow milieu - even though there is potential for further external generalization.
Nicole, I completely agree. Qualitative research CAN establish a causal connection for the specific individuals or settings studied (see the papers I cited above). This is quite different from generalizing a causal conclusion to OTHER populations or settings, a goal for which qualitative researchers have coined the term "transferability." Such transfer depends on understanding HOW the proposed causal mechanism will operate in a new setting, a task for which quantitative methods provide few useful tools, and for which a qualitative understanding of the meanings, processes, and contextual features of the new setting are essential. See the attached paper.
This sounds like what a mixed methods researcher would call a sequential exploratory design (qual --> QUANT), where you develop a model using a qualitative method and then test that model via a quantitative method.
In that case, the issue of generalizability would depend on the sample that you used in your survey.
I am skeptical about claims of causality for qualitative data, even if for specific populations. Joseph suggested it is possible if the participants themselves recognized the causes. But this is far from sufficient to establish causality. If science has taught us anything, it has taught us that we we often presume to be causal in nature is often not the case. Also, what methodological safeguards are there to control for other potential causes? As an example, it is difficult to know whether something is causal if there is no way to establish what might have happened in the cause's absence (control group in quantitative research). There are many other issues i have with inferring causality from self-reported qualitative data.
I didn't mean that causality can only, or definitively, be established if participants recognize a causal relationship. Causality can NEVER be established beyond any possible doubt, even (I would say especially) in quantitative research, which at best can show only a correlation. Establishing a causal relationship requires evidence that rules out other plausible causes. Randomized controlled trials are a relatively strong way of doing this, but have the defect (like all other quantitative methods) of providing no evidence about the presumed causal PROCESSES that created the result. This is a particular strength of qualitative research, which can provide strong evidence of the processes (including the meanings, motivations, and contextual influences) that led to the outcome. For a more detailed argument, see the attached paper, as well as the one on the value of qualitative inquiry for public policy that I uploaded in response to a previous post.
- Badges
- Science topic
- Similar topics
- Biological Science
- Bioecology
- Systems Ecology
- Landscape Ecology
- Connectivity