Thematic analysis - How many codes do you need?
Вставка
- Опубліковано 12 гру 2021
- Thematic analysis involves coding your data, and to develop strong and valid themes, you will ideally need to have a lot of codes. In this video, I explain this in more detail, explaing the overall purpose of coding first and then reflecting on 3 main reasons why it is better to have More rather than Less codes.
#qualitativeresearch #academia #researchpaper
➡️ Feel free to ask me questions in the comments, and if you feel that you may require a more guided assistance, you can book a lesson with me through the following link drkriukow.com/qualitative-res...
➡️ My self-study course on Qualitative Data Analysis www.udemy.com/course/how-to-a...
➡️ My self-study course on using NVivo 12 www.udemy.com/course/from-zer...
Visit my website and explore the different ways in which I can support you and your study! drkriukow.com/my-services/
Dear Dr. Kriukow, I already bought your course about qualitative analysis, it’s really helpful. I have one question, in intial coding, (line by line coding) do you recommend doing it in Word office or in NVIVO? Based on your experience . Thanks
This was such a phenomenal help! I panicked when coding a survey response regarding healthcare workers' experiences wearing PPE during the early phases of COVID-19. I thought I was coding too descriptively, and was unsure if grouping descriptive codes into analytic themes was a future step to take, or if I needed to redo the way I was coding this response. Thank you Dr. Kriukow!
this very helpfull, love the way you describe the content Dr. Kriukow😊
Thank you!
Doing data analysis so this was great timing 👌
Happy to hear this !
Very informative. At one stage i ended with so many codes and i though i was way off the line but soon realised that everything is useful and some bits will soon fall off. Thank you for the great video
Exactly, as I always say - too many codes is less of a problem than not enough codes!
I'm actually coding data for the first time my data is surveys and interview transcripts thank you after watching this video I'm going back and adding more codes I don't think I have enough thank you so much your videos are helping me through this process greatly. how much do you charge for a zoom meeting I didn't see the link
I found this really useful, thanks. One question though, what are your thoughts on double coding data? I.e., data that fits into two separate codes.
no issue with this at all, this happens All the time. If you looked at any given interview transcript that I coded, there will be absolutely loads of such overlaps and this is a completely common thing
Do researchers ever use excel? I don't have nvivo, but use excel and find it works very well. I have a column for the code, name of participant and then at the end I assign themes. Do you know of any publications or guidance for usi g excel? Or would this be classed as manual coding? Thank you for your content
This is not manual coding. Not sure about publications, but I do come across Many people who use Excel. As long as it feels ok to you and you feel comfortable using a given software, you're fine. I've recently came across someone using Mendeley for data analysis! Believe it or not. All of these are equally good as long as they help you answer the research questions!
Hi! Is it possible to include Farm Environment and Financial Difficulties as themes on farmer’s lived experiences? Thank you in advance
Both sound good to me, but I cannot tell you if it is ok to have these themes, since I have no idea what your data is like
Well almost all the respondents shared experiences regarding those that’s why we come up with that idea but we are not sure if experiences with that could actually fit with their lived experiences
Like, do their farm environment can be included with their lived experiences?
Hello Dr,
I have a question regarding coding of interviews, I want to compare 2 groups of learners based on their proficiency levels...
Should I do 2 projects or 1 in Nvivo software for comparing these 2 groups by visualization??
Regards.
I would always use 1 NVivo project. If this is part of the same research project, everything related to it should be in the same NVivo file
@@qualitativeresearcher
Many thanks, yes, it is one research.
Just it deals with comparing 2 groups.
So, based on this issue, I should prepare separate categories??
Could you pls elaborate to me?
Regards.
Hi, I have a similar project and I am definitely doing 2 projects - it's much easier. The groups are different in my project however, e.g dogs and cats and have had different questions so it is a bit different to yours maybe? Projects can always be merged later should you want to and if it makes sense to do so. Alternatively, you could use the Sets bit in Nvivo to look at the groups separately?
Thank you for the video. I have a question. Say I have coded 5 interviews. And when developing themes I have noticed that some of the codes have been left unused. Can I leave them out? How would I support this in my thesis paper? I would appreciate your answer. Thank you.
of course, in the process of developing themes lots of codes will be merged, changed, OR DELETED. Completely normal, no need to justify this as this a normal part of the process. Codes are there to develop a good understanding of everything in your data, but then once you do develop this understanding, good enough to decide on your themes, everything else will need to go , so do not hesitate to delete lots of your codes
Thanks, but how much is lots of codes?
My school requires you to have your themes before you start coding, based on the literature review. So, once I start coding and if the themes come out different during the field testing stage as I prepare for my proposal, then should this change my dissertation theme as well since the field testing is only like 3 people
not sure what you mean, sorry. I would not change anything based on the themes developed in the pilot study (field testing)
I think you're using deductive coding. This is the disadvantage because you already have predetermined themes. And there are themes that will emerge after gathering the participants' responses.
I am confused now because I now have more than 400 codes from 10 transcripts and I don’t know where to start next. Should I look at those codes from each transcript again or copy all 400 codes in one document and go from it? Hope will get the reply from you 👏
Hi Alex, have your questions been solved? If yes, would you like to share it with me? I have the same question now. Many thanks!
400 codes is a lot. I would try to refine & reduce them if you can - you may find on going through that you have repeats or similarities. I would imagine with this many nodes you only have a couple of examples under each code. By refining and reducing you may see your themes more clearly. Keep in mind the research questions you are trying to answer while doing this.
1) Code the data
2) Organise the codes [reducing and revising and make sense as you go]
3) Develop themes (in relation to the research question)
For a data set of 10 interviews of 10-12 pages each, how many codes would be ideal, if the final paper is supposed to be 15,000 words in all?
There's no such thing as ideal, you just need to be detailed. Based purely on average numbers that I see (but again, this by no means is to say that this is the "correct" or right number), I'd say around a 100 would be good, but if you have anywhere betwene 50 and 100 it woudl still be quite common, Then again, if you have More than this, I would not be surprised either
@Research with Dr Kriukow Thank you so much for taking the time to reply. Your direction is much appreciated. In all the videos on thematic analysis I have viewed, yours are by far the most helpful.
Also I want to ask if every participant answer should only has one code?
no, the codes will be used between the participants, some of them will be used for 1, others may be used for all 5 participants
Do each interview question have specific codes and themes?
For example, I have 2 interview questions for 1 research question. Does this mean that I can create 2 themes with different codes?
Tnaks
themes do not always have to much individual research questions. I may, for example, have a general RQ ("what are the experiences of nurses post pandemic?") and several themes corresponding to it ("challenges they face", "coping strategies", "nurses views", etc). Also, one interview question may result in response that covers answers to several research questions, so this is all pretty flexible
Hello. How many themes should I have on average?
You can have 1, you can have 5 or 10. Really hard to say, but if I was to give you an average, I would say around 5 is what I often have (just themes, this does not include all relevant sub-themes)
I panicked yesterday on this very concern !!! I thought I had too many codes and got lost .
I hope you feel better now :)
"If ever your interview was just only 5 person, How many codes need to do ? Because in our research we've interview only 5 person but I have 35 code for it ,it was ok not?" But I think your already stop answering any questions ,But I hope you've answer or notice my question one day...Thank you and advance!!
Hard to tell what number, as I explained - I would say that 35 is a pretty low number, but this depends on how detailed your transcripts were. I would normally expect abit more, however, maybe 70 or up to a 100 - HOWEVER, as I said, every dataset is different! So it is Very hard to generalize like this
So, what is the answer then? Sorry, but how much is too much?
There is no specific, fix numbers. I would say that for a final thematic framework, a total of 50 sub-themes and themes combined is quite extreme and I would avoid going above it. I have seen studies with 15 in total, and 20 in total, etc. If I were to draw an average, I usually end up having around 30-35 but then again - this will depend on a study
??????😔