In my perspective, numbers in qualitative studies really mean nothing. And you are right, numbers can be misleading.

Remembering that explication of the social phenomenon being studied is the primary consideration in qualitative research, I think that there is certainly a place for numerical/frequency data within (at least some)qualitative studies. Typically these numerical data are based upon descriptive statistics rather than predictive statistics and they can be employed to help contextually-based "thick description". For example, in some of our work we interview parents about their experiences with children having reading problems in schools and how this impacts the family. In addition to documenting the lived expereinces with cyclical thematic analysis, we also do mention how many of the parents view certain problems and how many actually discussed said themes. While this helps consumers of our research get a better feel for how often such themes and concerns arise, it also accomplishes another goal when trying to advance qualitative research within a discipline(s) that are not familar with or oreinted to qualitative research. I believe this was the reason several people created the "method of instances" in conversation analysis. For the uninformed, it enables them to see that the qualitative results might occur more frequntly than just once in a while.

As to the idea that numerical data might be misleading, that should never be the case if the qualitative researcher always employs Contextually-based intrepretation...That is what thick description is all about. Hope this helps.

Jack

Fairly important depending on the purpose. For example, if you’re seeking funding or assistance, or aid, etc. from government or private agencies, they will often want it quantified and understand it better that way. You can explain the variability (qualitative and quantitative), which is also important. Be clear on why you coded and lumped certain qualitative feedback in various categories (even if it was, “respondent uncomfortable” or “respondent very enthusiastic” discussing the issues, etc.). What do you want to communicate, to who, why, what is their time limit or attention span, etc. ... I’ve always found it useful. People respond to it. You can accommodate the qualitative variability, emotive diversity of responses, squishiness, etc. but be clear with definitions, examples and analogies. Hope this helps.

I think that it depends on the type of qualitative analysis. If you are conducting a qualitative content analysis, then including frequencies (and an interpretation of them) is quite appropriate. Hill et al. 2005 offer a framework for interpreting frequencies in qualitative analysis that I find useful. I am including a link to one of my publications that uses the frame as an example. Hope that it is helpful. Article We Who Are Dark...The Black Community According To Black Americans

It depends on the type of analysis. Importantly, one should not assume a participant not saying something means it does not exist, only that saying something means it does. You may want to read some articles on quantizing and quantitizing data.

If several participants indicated similar idea/experience/understanding, it helps to show commonality of that theme. However, those experiences / understandings they refer to may have different context and details, which is also important to mention. I agree with Jack above, it helps in providing "thick description". I also agree with Cheryl that if a participant doesn't mention something wouldn't necessarily mean it didn't exist. It may be true for that particular person, within her/his life story and circumstances.

Hi Jacob,

I think it would be fine. I believe it would be used as a descriptive measure more than anything. If they were unreported, then I'd have curiosity regarding just how many people were interviewed, and how many of them had similar answers. So, in my opinion, it's fine.

Here's an item that provides some good food for thought:

https://books.google.com/books?id=bOLFDQAAQBAJ&printsec=frontcover&dq=creswell+qualitative&hl=en&sa=X&ved=0ahUKEwiFwpuApKLbAhURvVkKHRWQApMQuwUIMDAA#v=onepage&q=creswell%20qualitative&f=false

Have a great day!

--Adrian

The following papers/views may further help, namely: numbers can complement and enhance narratives/improve transparency/support findings (see: Black, 1994; Sandelowski, 2001; Neale et al, 2014). More specifically, the use of numbers in conjunction with qualitative methods and data is an effective strategy for qualitative researchers when such the numbers are used as a complement for the research……Nevertheless, incorporating numbers in qualitative research has some advantages (for example, adequately presentation/enhancing clarity/support of evidences/systematic check for evidences, etc.); as well as some problems (such as inaccuracy/misleading/ misrepresentation/replacing the actual description of evidences, etc.)……(see: Sandelowski, 2001; Maxwell, 2010, Hannah and Lautsch, 2011; Neale et al, 2014).

• Black, N. (1994) Editorial: Why we need qualitative research, Journal of epidemiology and community health, 48, 5, pp. 425-426.
• Hannah, D. R. and Lautsch, B. A. (2011) Counting in qualitative research: Why to conduct it, when to avoid it, and when to closet it, Journal of Management Inquiry, 20, 1, pp. 14-22.
• Maxwell, J. A. (2010) Using Numbers in Qualitative Research, Qualitative Inquiry, 16, 6, pp. 475-482.
• Neale, J., Miller, P. and West, R. (2014) Editorial: Reporting quantitative information in qualitative research: guidance for authors and reviewers, Addiction, 109, 2, pp. 175-176.
• Sandelowski, M. (2001) Real qualitative researchers do not count: the use of numbers in qualitative research, Research in nursing and health, 24, 3, pp. 230-240.

Thank you all for your comments. My hesitation to use frequencies corresponds with your answers. Indeed, frequencies give a feel of the prevalence of a certain type of response, of a content that appeared in the taxt. However, that is exactly the problem. Sampling in qualitative research is not randon (usually), and thus the frequencies represent noting more than appearance of a content within the sample, and the aforementioned "feel" may be merely a source for wrong inferrences. As Maxwell (2010, p. 479) notes:

" Numbers can lead to the inference (by either the researcher or the audience) of greater generality for the conclusions than is justified, by slighting the specific context within which this conclusion is drawn. A particular setting or sample may be unrepresentative, and a facile reading of quantitative results may lead a reader to ignore this limitation. Qualitative research is intrinsically local, and any claims for the generality of its conclusions rely on a different kind of argument from that of quantitative research, what Yin (1994) called analytic rather than statistical generalization. "

isnt that content analysis? Content analysis is a form (but the least rigorous form) of qualitative analysis

I think that reporting actual numbers of participants who did or did not say something implies a level of precision that cannot be supported in most qualitative data. Even as "descriptive statistics," the small size of most qualitative samples would yield enormous standard errors for these counts, making them almost meaningless in any numerical sense.

In contrast, I think it can be useful to give broader indications of how prevalent something was -- such as saying that something was only mentioned by "a few" participants or that "nearly all" the participants mentioned something else. This kind of statement helps establish that there was a pattern in the data, which is what we are usually seeking.

I'm a little confused. If it includes numbers isn't it a mixed methodology, of which there are ample uses. If I want the richness of qualitative data, I want to use the qualitative method. If I also want to show the subset of respondents who may not understand the topic by the definition, and therefore further explain their qualitative data, is that not meaningful?

It completely depends upon the nature of the problem being studied. I have used it very meaningfully in few of my studies.

Gary Sizemore it is not a mixed-method if you are merely providing prevalences for qualitative responses without intending to use those prevalences to answer a research question (not having such a uestion in the first place).

David L Morgan the use of precise vs. imprecise indications of prevalence is contestable. Nevertheless, I agree that it is problematic to use numbers. The problem is a multifarious one. It is not only that prevalences may be misleading, as Maxwell (2010) notes, but that themes cross categories, and are the resercher's rather than the participants' construction. As such, it would be very difficult to count, and indeed meaningless in relation to a qualitative research question.

Numbers are not required in qualitative studies, but depending on the context they can make sense. To me it seems important that they belong to an explicit argument. It means you have to indicate why you quote these numbers and what and how you want them to be understood. There is no explicit pardigm as in a quantitative study, therefore you have to provide a context for understanding.

I have recently finished my reserch study which was qualitative and the data I collected using in-depth interviews. I went thruhg all the interviews and followd few authentic books written on qualitative research study. Using numbers is not neccessary either, however; giving the meaning to the participnats observations and their experiences they tell in the interviews is core prorgam of qualitative research study. That is why this is very fleible in nature rather fixed and static.

I tned to agree with your comments above Jacob. Many years ago when I was just beginning my research career, the very distinguished anthropologist Sir Godfrey Lienhardt cam to give a presentation (he was being honoured by a nearby university). A young research scientist asked him about the validity of small samples. Sir Godfrey invited the questioner to look around him and reflect on which might be more valuable - what 25,000 people replicate, or what one person with insight innovates? Connoisseurship in research involves acquiring the ability to sense the differences that make a difference, in context, as Gregory Bateson argued.

Number has less value definitely and even the sharing of one participant might be much more important and logical than the responses from 10 to 20 participants who are in wrong direction.

Number/count does matter in qualitative studies. It indents to explore the information instead of number/count.

Frequency of responses from participants maximize the indentation of the information in FGD.