I often find it helpful to have a table at the beginning of this section which summarizes the themes and sub-themes. This can be particularly helpful to the reader if the numbers of theme and sub-themes is lengthy, or if their organization is complex.

Not necessarily a separate summary. What I find useful is to summarise the themes as I present the findings, by building a coherent narrative of how they emerged and related to one another.

Dear Khaled,

From my own perspective, it is important and necessary to provide the readers with at least a brief explanation about themes and sub-themes, otherwise, it might cause some confusion for the reader to understand what exactly you mean by each theme. For example, if you have come up with a theme called "Trust", which impacts participation of a user in online communities, it is not clear if by "Trust" you mean trust in the platform, trust in other users, or trust in the shared content. Besides, different researchers define trust in various ways based on the concept and context of the research. Therefore, it might be a good idea to make themes and sub-themes clear for your reader even with providing one sentence for each of them. In most cases, to save some space, you could provide all the themes and sub-themes in a table and provide a short explanation for each of them.

I would provide the summaries to elaborate the intended meaning behind the themes identified. To save space, I have found theme mapping useful as an addition to written summaries. See Thomas, G. (2009) How to do Your Research Project, London: Sage for examples of theme maps used to organise and present themes in graphical form.

I often find it helpful to have a table at the beginning of this section which summarizes the themes and sub-themes. This can be particularly helpful to the reader if the numbers of theme and sub-themes is lengthy, or if their organization is complex.

Yes, summarizing for each theme and sub-theme is important for the findings in terms of correlating each theme and sub theme with each other to provide a good narrative. Atleast one sentence summary for the ones whose findings are not as valuable as the rest.

Using ATLAS ti software, its easy for you to analysis and summarize themes and sub themes

I would include a table describing and summarizing each theme/subtheme in the codebook in Appendices. It is an important finding and I think it also helps with the narrative. In my findings chapter, I would provide my interpretation of the findings, along with representative quotes.

Yes, it's very important. By summarize the findings into the tables both of themes and sub themes, it will help you and the reader easily understand what is the content in your themes and sub themes. I usually also put the meaningful quotation to help me recognize the findings when I'm doing the analysis, so I won't look back into the whole transcript again and again.

Yes, as you need to have a clear introduction for your theme/sub-themes and tell the reader what is it about, you need to summarise what you did describe

Yes, each theme and its sub-theme should have a summary. Atleast how many out of how many said what should be summarized for each sub-theme for the reader to have a holistic understanding of the data.

This is a slight aside, but I have found the concordancing software linked below to be very useful for initial explorations of digitally formatted text. It allows you to quickly explore your data where you want to identify particular words, phrases and frequencies as starting points for more elaborate explorations.

See the website here for links to the software:

https://www.laurenceanthony.net/software/antconc/

I am using MQXQDA and it has been really helpful. It has helped me locate phrases, word frequencies, and easily create a word cloud. Thank you, Dr. Adrian Twissell

Khaled Alostath Using phrases, word frequencies, and word clouds might be a good way to start for some forms of analysis, but these are all descriptive tools which are unlikely to be useful in creating the kinds of interpretive goals that are associated with creating themes.

Well David, David L Morgan , I did suggest it only as a starting point. It's just another tool - helpful, or discard-able, as you choose. When you have a lot of transcripts to look at, it can help you focus, or not, as the case may be :)

I do know that it can be quite difficult to deal with large numbers of transcripts, but I still worry that these particular tools will lead to descriptive summaries, rather than interpretive themes.

I also find it useful to have a table or schematic diagram as a more visual presentation of the themes. And it is always necessary for me to have a description of each theme with sample data extracts

David L Morgan I understand. I am aware of its function, and I am not using these tools in my thesis. However, in which cases should or is it preferable to use words frequencies, word cloud, etc.?

Adrian Twissell

Khaled Alostath

Many researchers are interested in prevalence. For example Ken Hyland has written much about the way certain grammatical structures are used in academic writing in relation to identity; for example first person pronoun. This type of research often draws from large corpora of texts and one way to explore them, for prevalence/frequency of specific textual clauses is to use a concordancing tool such as the one I mentioned.

This is not to suggest, however, that these researchers stop there. I agree with David L Morgan in the insistence that interpretive themes represent the eventual goal of qualitative research of the type mentioned in this question.

I personally do not have any use for word clouds etc. because they don't contain any interpretive information. Indeed, they are barely even descriptive because they are so de-contextualized indicators. Perhaps they make useful pictures for those who know nothing about a topic and the things people say when they discuss that topic, but why would someone with no knowledge about the topic be reading your paper?

Thanks for your response David L Morgan . If you are comparing 'word cloud' with concordancing then I would argue that there is no comparison. Word clouds are merely visual representations, whereas concordancing tools go much further in identifying quantitative data for those who wish to use it. As I stressed previously, concordancing is just a starting point, should you wish to use this approach; the reading of narrative following simple identifications of textual features is generally more enlightening. Alternatively, explorations of narrative leading to the generation of themes, a thematic analysis, may be usefully validated through a corpus analysis using concordancing software of the type I cited. And therefore you gain both interpretive information and an indication of prevalence, if this is desired in relation to your context.

Adrian Twissell If you collect your own data, and thus hear what participants have to save before beginning analysis, then I cannot imagine how a concordance could be "enlightening." And I certainly would not call its products "narrative," since they are so thoroughly decontextualized

Further, themes are meaningful interpretations of the data, not strings of words. So the idea of "validating" them through corpus analysis does not supply any "interpretive" information. Lists yes, interpretations, no.

In short, people are always looking for short cuts so that the machine will do the work of reading the data and interpreting it. Word clouds are the worst, and concordances little better.

Maybe this approach is useful if you scrape thousands of tweets or other "big" but thin data, but I can't see how it could possibly be useful if you collect in-depth data.