Dear Prateek,

Sample size in qualitative data analysis doesn't work in quite the same way. Individual subjects are more complex (e.g. an interview rather than a set of variable values) and it depends how complex an analysis you are trying to create (e.g. as measured by the number of themes or categories identified within the data).

What you are often looking for is saturation - where adding new data does not improve the explanations of the themes or the categories or add any new ones. That is when you should stop.

Hi Dr Nidhi, question is regarding "qualitative research"

bests

... as a rule of thumb, a one hour detailed interview can produce about 25 pages of transcribed information. This implies that with eight interviews you have 200 pages of information to fragment. Therefore, at masters course work or below I recommend 8 one hour detailed interviews, if less detailed and less than one hour I recommend 15 interviews. For Master by research work and above, I will go with Peter Samuels' answer ... until you reach saturation.

A number of issues can affect sample size in qualitative research; however, the guiding principle should be the concept of saturation. This has been explored in detail by a number of authors but is still hotly debated, and some say little understood. A sample of PhD studies using qualitative approaches, and qualitative interviews as the method of data collection was taken from theses.com and contents analysed for their sample sizes. Five hundred and sixty studies were identified that fitted the inclusion criteria. Results showed that the mean sample size was 31; however, the distribution was non-random, with a statistically significant proportion of studies, presenting sample sizes that were multiples of ten. These results are discussed in relation to saturation. They suggest a pre-meditated approach that is not wholly congruent with the principles of qualitative research.

The following link has more details

http://www.qualitative-research.net/index.php/fqs/article/view/1428/3027

Depending on the subject matter of the investigation. Qalitative research reqires detailed interview, ten cases or 10% will be adequate.

It is completely related to research background and target. In this sense, the amounts of interviews or etc. can be considered as the credibility of a research. it means that scholars can not present a concrete argue or discussion according to a low number of interviews for instance. In addition, the total number of research population also should be mentioned. In my view, fifteen could be very low acceptable size for qualitative research under specific circumstances. 

Thanks Maryam, Peter, Kambidima, Muayyad, Christopher, Javed and Teri for your inputs and references.

My take on this is that it depends on what kind of qualitative research you are looking at or your research question. If your aim is to build a theory by starting with data collection then the point at which you stop sampling will be when you reach theoretical saturation or conceptual saturation (Saunders and colleagues, 2012). For a qualitative research that does not have theory building as a core focus but just to offer complimentary data to give more depth to quantitative data or to explore an issue, will depend on what is doable or feasible for you as a researcher and in your context. Patton (2002) believes that qualitative inquiry should typically focus on relatively, small sample - even 1 case can be enough. Thank  you 

Dear Prateek

I agree with the above comments.  I know some journal accepts qualitative papers only if the sample size is 20 or higher.Many do not ask this question.  Recruit and interview until the data contents reach saturation. That's it.

Qualitative research is an umbrella term, which basically means it does not rely on assumptions that are quantitative necessities, such as sample size. I think one of the problems with qualitative research is that it aspires to the "status" of quantitative research by inventing its own catch words, e.g., saturation.

There can be no golden rule. Imagine you want to gauge political astuteness in politicians. If you got one interview with Barack Obama and one with Donald Trump, how much could you glean from just those two? Do you really need many politicians to work it out?

Good one John!. 

John Voyce has the most complete answer here. There are several different forms of qualitative research and each has different assumptions and procedures. As others have said, "saturation" is the key as to when you have collected sufficient data. 

Here's a fairly good article on this topic. 

http://tqr.nova.edu/wp-content/uploads/2015/09/fusch1.pdf 

Thanks John, Violet and Susan for your inputs. Very Useful.

Dear Prateek, 

The idea of qualitative research has to do more with the quality of data you can gather rather than the quantity of data as happens in quantitative studies. Quantitative studies rely on the quantity, the more the better so as to determine the power of the output/outcome and its significance to the population under study which will later be generalized to other populations where the research never took place. 

The above comments are very helpful and well done to the contributors!

Dear Prateek,

Qualitative analyses typically require a smaller sample size the quantitative analyses. Qualitative sample sizes should be large enough to obtain feedback for most or all perceptions. Obtaining most or all of the perceptions will lead to the attainment of saturation. Saturation occurs when adding more participants to the study does not result in additional perspectives or information. Glaser and Strauss (1967) recommend the concept of saturation for achieving an appropriate sample size in qualitative studies. Other guidelines have also been recommended. For an ethnography, Morse (1994) suggests approximately 30 - 50 participants. For grounded theory, Morse (1994) has suggested 30 - 50 interviews, while Creswell (1998) suggests only 20 - 30. And for phenomenological studies, Creswell (1998) recommends five to 25 and Morse (1994) suggests at least six. There are no specific rules when determining an appropriate sample size in qualitative research. Qualitative sample size may best be determined by the time allotted, resources available, and study objectives (Patton, 1990).

You can find useful sourcesbelow

Creswell, J. W. (1998). Qualitative inquiry and research design: Choosing among five traditions. Thousand Oaks, CA: Sage Publications.

Glaser, B. G. & Strauss, A. L. (1967). The discovery of grounded theory: Strategies for qualitative research. Piscataway, New Jersey: Transaction.

Morse, J. M. (1994). Designing funded qualitative research. In Denizin, N. K. & Lincoln, Y. S., Handbook of qualitative research (2nd Ed). Thousand Oaks, CA: Sage

Thanks Peter and Thinh (Special thanks for references). 

I don't  agree with John because if you got one interview with Mr. President Barack Obama and one with Mr. Donald Trump  then having  two interviews only don't satisfy the basics  for qualitative research   and since the question is about RESEARCH then  we have to have variation for a set of observations

Abdulrazzak:

You only need two to have variation - that's basic statistics. In quantitative research they talk about variance, which is mathematically defined. You cannot do that in qualitative research. Obviously there is variation if you compare Trump to Obama, but it is not really quantifiable, unless they talk about how much they would spend on healthcare, etc.

The difference is one of political views which are far too complex to process via some mathematical model imported from the hard sciences. As I stated above, for me the problem with qualitative research is that it seeks legitimacy by looking for catch phrases that sound scientifically convincing. Most of it comes from so-called "grounded theory" which,  I find philosophically unacceptable. Straightforward ethnography is, to my mind, much less pretentious.

The social sciences tried to legitimise themselves by using quantitative methods developed in the hard sciences. This led to Charles Wright Mills' accusation of abstracted empiricism - i.e., obsession with data while forgetting the theory. Qualitative methods are seen as the poor relations of quantitative methods, which is why no government will take measures based on a grounded theory study.: governments want numbers.

Unfortunately, I think the obsession with methods has crept into qualitative research, so it starts to resemble Mills' characterisation of abstracted empiricism.

Dear Prateek

I agree with Peter comment , however, you need to know the limit of small sample size.

You can use G*Power software to calculate sample size.

G*Power software is effective tool to calculate sample size for many ranges of experiments. Also, you can determine effect size and power of the test, G*Power is free to download and easy to use after reading the manual, the download link:

http://download.cnet.com/G-Power/3000-2054_4-10647044.html

The link to download G*Power manual:

http://www.gpower.hhu.de/fileadmin/redaktion/Fakultaeten/Mathematisch-Naturwissenschaftliche_Fakultaet/Psychologie/AAP/gpower/GPowerManual.pdf

Good Luck 

Lately, there have been several attempts to predict what I would call "code saturation," which is the point at which you have located at least one instance of all the codes that are likely to appear.

I myself am rather dubious about this approach, but it definitely becoming part of the literature on saturation.

Thanks Khalid and David for your useful inputs.

1,3,5,7 is ideal for the sample size in qualitative research n its depend on what type the qualitative research that you are going to do

The key point is this: Our qualitative sample must be big enough to assure that we are likely to hear most or all of the perceptions that might be important. If the team wants to be more conservative and reduce the risk of missing the one-in-10 subgroup to less than 1 percent  then a sample of nearly 50 would be needed.

Thank you Kiran for your input.

I was also searching for information on ideal sample size for qualitative research and the two articles recommended by D.L. Morgan above were helpful. I hope you find them useful too.

Dear Scholars/Professors,

I am interested to do my research work on " Social security of employees working in steel industry in india". If I want  to carry out QUALITATIVE research on this topic how should I do that kindly share your valuable views and suggestions. What will be the research questions/ objectives, how should I frame questionnaire and what are the questions would be?

Usually, sampling until saturation is reached is advised 

Prateek, I agree with Peter's comments. One advice: When you see the data is saturated (in-depth interviews), recruit 3-4 more participants, collect data and make sure the data is completely saturated. 

Having said that, there are some journal do not accept manuscripts if the sample size is less than 20. One of them is Psycho Oncology, the official Journal of the American Psychosocial Oncology Society ( impact factor 3.095). If you are aiming to publish your research in a particular journal, you may want to look into this issue now. 

@Adhikari

All your questions depend on the hypothesis of your research, and the problem that your study try to find solutions for it. There is no unique model found for all hypothesis.

Good Luck

• Thank you very much Prof. Khalid

There are no specific rules. Various methods in Qualitative research has its own view about the number of cases/interviews to be done. For example in a case study research design, it may vary from 1-20 or more. When it comes to phenomenology, grounded theory approach, and narrative research, it has been done on the principle of saturation. But in Ethnography, the richness of data is taken into consideration rather than the number of interviews.

The important point is, it is the responsibility of the researcher to ensure whether the collected information has the richness to answer your research questions or not.

The selection criteria will determine the sample size. As these criteria increase, the number of samples should also increase. There are various formulas for calculating the sample size and  books that give a fixed number of samples when the universe is infinite. I use the following book as a good resource. I hope it helps.

Hair, J.F., Ronald, L.T., Rolph, E.A. and William, B. (1998). Multivariate Data Analysis. 5th Edition. USA: Prentice Hall.

This sounds more like a quantitative question nu which sample size is important but also dependant on your aim. But in quantitative you want to know "how much OR how often. In qualitative research we want to know "what". Therefore 10 good deep interviews should be enough.

Guest (2006) argues that sample size of 6-25 participants in a qualitative research will enable the research reach saturation in phenomenological study. Therefore, saturation should be the benchmark within the suggested range of sample size.

The world of qualitative research is another onze, so not another planet. Qualitative research is meant to find out what the problemen is, quantitative for counting. Therefore just do some good deep interviewing until saturation. Therefore more of less 10 good interviews should be enough.

Base your sample on saturation. I'ts the ideal thing to do.

Glaser and Strauss (1967) suggest the concept of saturation for obtaining a proper sample size in qualitative research. Other guidelines have also been suggested. For an ethnography, Morse (1994) recommends around 30 – 50 participants. For grounded theory, Morse (1994) has recommended 30 – 50 interviews, while Creswell (1998) recommends just 20 – 30. And for phenomenological research, Creswell (1998) suggests five to 25 and Morse (1994) recommends at least six. There are no particular rules when arranging a proper sample size in qualitative research.

Thank you Dr Alloubani

I believe that collecting qualitative data depends on the researcher and the required data. However, I'd say 10 to 15 interviews are v.good.

I recently conducted my first qualitative study for a 2nd masters. I had 15 health professionals provide perceptions on a growing issue in mental health. But I was criticized heavily on the method of giving the definition of the concept and asking for their feedback. I'm of the opinion that the scientific community is still bias toward this approach unless it is in conjunction with a quantitative arm. I'm now trying to find the best strategy for preparing an article worthy of publication.

Take a look at the journal Qualitative Health Research, which as a strong ISI impact factor.

There is neither fixed sample size nor ideal sample size. Sample size depends upon several factors. These factors have already been mentioned by the other respondents. As the sample size becomes larger, accuracy of findings becomes more.

I find this is a keenly discussed topic and there is no one size fit all, however, the notion of saturation and the quality of responses must combine to determine the number of respondents or interviewee. for a group of professionals, 10-15, but if it is a social group with different perception, you may need to do more so as to see the patterns.

See

Sample Size in Qualitative Interview Studies: Guided by Information Power Kirsti Malterud et al., 2016

I would say it depends on the research problem and the question asked. In my opinion, a qualitative study such us using Focus Groups should be a prerequisite for quantitative study. The focus groups should enable the researcher to formulate a well defined research problem. Let us not forget the Ice Berg principle and that almost 90% of the research problem is hidden. What's more, the focus groups enable us to formulate Research Hypotheses and to write statements for constructing an instrument (questionnaire). Moreover, Knowing the appropriate sample size is a must in drawing a random (representative) sample from the population. What I mean it is imperative to draw a representative sample in quantitative research but in qualitative research we use Snowball Sampling ( we use referrals) and it is biased. Therefore in qualitative research we don't attempt to draw inference from sample to population because the snowball sample is not a representative sample of the population.

I agree with Abdulrazzak Charbaji .

I like to add that for a sample to be drawn from a finite population , it may be possible to determine the sample size depending upon the factors namely the population size , the desired efficiency of estimates , the cost available for survey , the time available for survey and some others.

However, in the case of a sample to be drawn from an infinite population , there is no fixed suitable sample size. In such situation one needs to follow that as the sample size becomes larger, accuracy of findings becomes more. Of course, for choosing size of the sample one can follow the sample sizes as prescribed in the different phases of clinical trials.

No recommended sample size. It depends on the purpose and nature of research. It also depends on the research philosophy you use. For instance, a few such as 25-30 semi-structured interviews can reach saturation if you study is based on interpretivism . Critical realism may need much more than this, as here you are not just trying to capture the sensation rather you need exploration of the deep mechanisms and structures underlying the sensation. It depends on the analysis as well. For those pursuing discourse analysis, 3-5 interviews will be sufficient. Importantly, how much expertise you have as a researcher to conduct an interview and also the time you are engages in conversation with the interviewee also matter. The purpose anyway should be to collect the data that sufficiently and satisfactorily address the purpose of your study.

Much of what colleagues are saying is helpful but try reading an article by Bryan Marshall et.al., (2013), DOES SAMPLE SIZE MATTER IN QUALITATIVE RESEARCH?: A REVIEW OF QUALITATIVE INTERVIEWS IN IS RESEARCH in the Journal of Computer Information Systems. I found it very useful in answering the kind of question you raised.

This article is also useful in determining adequate sampling size for a qualitative research: Malterud, K., Siersma, V. D., & Guassora, A. D. (2016). Sample Size in Qualitative Interview Studies. Qualitative Health Research, 26(13), 1753–1760. doi:10.1177/1049732315617444

The answer given by Muhammad Usman is recommended.

The answers given by David Bishau and Azizan Morshidi will be, as I understand, beneficial and are recommended.

Factor analysis is quantitative research and that indeed deals with sample size, but depending on your goal and the number of factors that need to be extracted, a few hundred cases will do. Qualitative research has nothing to do with sample sizes, as in qualitative research you try to make the image of the problem field, so how does it look like, what elements so the 'what' of the problem, or what are people's motives ? Qualitative researchs quality depends on the quality of the interaction between subject and questioner, the questions qualities, the way people were selected etc. When we compare the problem with a square, we first need to know the angles,and not some dots in the middle. So select people with an extreme vision and ideas, as you need to know first where the problem starts and where it ends. So slechte, do not apply quantitative research items like a selectiveness and sample size, as statistical needs apply there. Good luck, Bob.

In my opinion, this rating of sample size is arbitrary.The sample size depends on the kind of the material that has to be examined.

Always in qualitative researches, the researcher (s) is/are interested in analysing the diversity of data for the research. So the sample size is not very important, as a large sample size can lead to the accumulation of a lot of data but sometimes can not guarantee the attainment of reliable information for the research. A small sample can produce fulfilling information if it is well selected than relying on a large sample size that is not well- informed.

We have done survey research that includes quantitative and qualitative responses. Our N is over 4,000. For specifically the qualitative questions (which are basically open ended questions asking them to share more about their thoughts, experiences or feelings, what would be a good sample size as we cannot sample 4K :) I can't seem to find any research that points to this. Thanks so much!