I am working on scale development in behavioral finance by undertaking a mixed-method approach using the exploratory sequential design. The phenomenon has diverse meanings in existing literature (some measuring it in terms of behavior while others use combinations of dimensions such as knowledge and access). I am unclear about its definition, so I want to explore the perspectives about the concept and what components participants feel it includes by taking a phenomenological approach.
My notation for research design is qual→ QUAN→ QUAN. Please guide me as if my approach is right. Do I need to go in so much depth as my main aim is to develop and validate the scale and not undertake a qualitative approach? I just need to take a qualitative viewpoint to support my framework or to guide the initial items bank and dimensions I created using a literature review. Second, if not then is it still remains a mixed-method design? and if yes, then guide me as to how much the sample size should be? I referred to Creswell & Poth's (2018) Qualitative inquiry research design, which says 3-10 (Dukes, 1984) or 5-25 (Polkinghorne, 1989). I am confused, so please if you have any concrete reference, suggest here.
Thanks in advance to the reviewers!
There’s a lot to unpack. Is your ultimate goal to develop a survey? What population? Admittedly, a phenomenological analysis seems out of place, especially if you want to reach a very heterogeneous sample. Without further knowledge, if you have a scale in mind start with it first and then use cognitive interviews to research it. If you need to develop a scale, then use thematic analysis to start.
sample size for all qualitative research design are collect the samples until the data saturation achieved approximately 30 - 40.
Exploring "the perspectives about the concept and what components participants feel it includes" is a typical goal for what are known in mixed methods research as exploratory sequential designs (qual --> QUAN). Since almost all qualitative data explores feelings about phenomena, this does does not require an explicit phenomenological method.
Focus groups are a common tool for the qualitative portion of this design, because they allow you compare a variety of different perspectives on the topic. If you already know that there a diverse set of possible responses, then I would plan for 5-6 focus groups. If you use individual interviews, saturation for this task its less demanding than for a complete qualitative understanding of a topic, so you might plan for 12-15 as a starting point.
I did not suggest a sample size because I don’t know how/what diverse means. The more diverse, more heterogeneous, the larger the sample size might have to be. Everything depends on your purpose ands aims (plus if you’re a part of a team). Still, phenomenological would be misplaced.
David Coker I have considerable experience with developing questionnaires from qualitative input, and I generally recommend 3-6 focus groups or 10-15 individual interviews for this task -- depending on how complex the topic is and how diverse the survey population will be.
But I still consider saturation to be the key consideration. If you are continue hearing new things, then you should be doing more interviews.
David L Morgan I think you’re probably correct … but I always recommend move being rules of thumb. Every case is different.
If we don’t rely on prior experience we’ll get “rules of thumb” like 30-40.
Qualitative data collection involves in depth interactions with the participants.
Larger sample size means less quality interactions, 3- 10 sample size is enough. However, the maximum is 25.
The poster would be well advised to develop a sample size dependent on purpose and aims. A very homogeneous group could be 6-15+, while a multisite study with a very heterogeneous population could be 12-30+ (depends on how many different sites). The "+" means you are open to following the data. The key might be saturation, but please use a method to define and report what you mean; surface-level is easy, while depth and adequacy cannot always be forecast ahead of time.
Francis, J. J., Johnston, M., Robertson, C., Glidewell, L., Entwistle, V., Eccles, M. P., & Grimshaw, J. M. (2010). What is an adequate sample size? Operationalising data saturation for theory-based interview studies. Psychology and health, 25(10), 1229-1245.
Mason, M. (2010, August). Sample size and saturation in PhD studies using qualitative interviews. In Forum qualitative Sozialforschung/Forum: qualitative social research (Vol. 11, No. 3).
Braun, V., & Clarke, V. (2021). To saturate or not to saturate? Questioning data saturation as a useful concept for thematic analysis and sample-size rationales. Qualitative research in sport, exercise and health, 13(2), 201-216.
Hennink, M., & Kaiser, B. N. (2022). Sample sizes for saturation in qualitative research: A systematic review of empirical tests. Social science & medicine, 292, 114523.
David Coker I disagree with your estimates for sample size, which might be appropriate for a full-scale, stand-alone qualitative study, but are too high for a survey development process. There is a reason why such studies are summarized qual --> QUAN in mixed methods, because the "small quan" is given the specific purpose of creating items that meet the needs of quantitative study that drives the design as a whole.
Fair enough. We will have to disagree. I still will say a very heterogeneous sample such as many age groups and cross-cultural will be far different than a homogeneous sample. Another issue is long questionnaires are also problematic with small samples (researching 10 questions is different than 100 with 7 factors, etc.). Even for small qual. I think the limitations of treating a heterogeneous group as small would lack representativeness. Each group might have 1 or 2 spokespersons! Without knowing the poster’s purpose, we don’t know what he is trying to accomplish.
- Badges
- Science topic