In my quantitative study I have chosen a sample of 200 people for a questionnaire survey, I also want to a qualitative study for the same topic, so what should be my interview sample for that? Are there any set criterias for PhD sampling?
Hi Henna,
This is just personal preference, but as a 'rule-of-thumb' I have over the years always aimed for between 5-10% of my quantitative sample - this gives a good ratio, i think. However, if we are talking large sample sizes (i.e. in the thousands), common sense should prevail. As others have mentioned, this also depends on your methodology and the accompanying 'type' of interview.
Hope this helps.
David
Any number (from 1 to many) can be justified in qualitative research. You can also start from 1 and add more until you decide that this is enough (than any additional interviews will not add much to what you already have).
There is not hard and fast rules for the qualitative sampling however you have to choose the sample which can triangulate your quantitative data. Once you start your qualitative data collection, if respondent's response comes similar you can stop the data collection that means sample to saturation.
I'm not sure I understand exactly what you are asking. I am interpreting your question as there are two contemplated studies, one quantitative and one qualitative. If that's the case why are you doing two studies? What is the sampling frame from which the 200 was drawn? Was the sample selected according to sampling error considerations? Selecting a sample for the qualitative study should follow those rules too. However, if you do a separate study you'll have a large amount of interviews to complete. Perhaps reviewing your sampling technique is a place to start.
Another way to go for the qualitative study would be to obtain a ramdom sample from the 200 pool of the quantative study sufficient to avoid sampling error. Generally 10% is sufficient.
However, the real question is what result are you seeking? Either procedure will produce only reliability of the results. It will not produce evidence of validity that is more important. For validity a correlation with results of a measure with established validity is necessary.
Hi, I would really start with qualitative research using either focus groups or a sample of the size of a focus group for in-depth interviews to investigate and explore possible outcomes to a subject of interest. From there, I may proceed to quantitative analysis using your set sample or more to look for causal relationships. This process could be triangulated with content analysis technique as well.
Validity and reliability tests exist to either approach but in different forms.
Hello, Henna.
I'm glad you're asking this question. It seems to me that in social science research, you want a large enough sample that will yield the information you are seeking, but it needs to be limited, too. Be as parsimonious as you can be as you assemble your sample, out of ethical considerations of respect for your subjects.
It is not obvious to me from the way your question is phrased whether the timing and spacing of your investigations is intended as a mixed method study or as two independent investigations. Generally speaking, the answer to "how many" depends on which type of qualitative method design you have in mind. In any type of qualitative research design I have ever seen, random sampling really doesn't make much sense. For example, in case studies, you must select your cases, and there needs to be some driving force that guides you to select compelling cases, not random selection. For a phenomenological study, you would need subjects who have experienced the central phenomenon in question, and random selection might fail you here, too. For grounded theory studies, you would select future study subjects based on the emergent data, continuing to interview until you encounter theoretical saturation.
Hi Henna,
This is just personal preference, but as a 'rule-of-thumb' I have over the years always aimed for between 5-10% of my quantitative sample - this gives a good ratio, i think. However, if we are talking large sample sizes (i.e. in the thousands), common sense should prevail. As others have mentioned, this also depends on your methodology and the accompanying 'type' of interview.
Hope this helps.
David
Hi Henna,
Based on your question, sounds like you are doing Concurrent Triangulation Design (Creswell, 2009) / Convergent Parallel Design (Creswell & Plano Clark, 2011) i.e. performing quantitative & qualitative data collection at the same time. Think there is no fixed rule how many interview samples you should obtain for the qualitative data collection. Depending on your research objective, you can choose minimum 1 for rigorous interviews until the informant's inputs reached saturation or choose as many as you can until their interview comments also reached saturation / no more new discovery. However, this mixed method design requires a lot of your effort on both designs' data collection & analyses. Wishing you all the best.
Regards,
Fung
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size