Homogeneity in IPA studies, and the 'magic' minimum number

Ask here about academic and research issues, like designing studies, recruiting participants, choosing statistics, submitting for publication, etc.
User avatar
Pink
Posts: 380
Joined: Sun Aug 21, 2011 8:36 am
Contact:

Homogeneity in IPA studies, and the 'magic' minimum number

Post by Pink » Fri Mar 14, 2014 3:58 pm

Dear All,

*desperately seeking qualitatively informed bods*

my thesis involves using IPA to interview parents of children who have experienced certain life events, in order to gain a sense of the impact of the life events on their relationships with their children. I planned to recruit 11 participants, and to ensure homoogeneity within the sample said that these would only be parents of children up until the age of 12, operating on the broad assumption that parenting teenagers will bring different challenges. However, I've had very little luck with recruiting, and have only gathered 5 interviews. There is a 6th person (apparently six is the magic 'minimum number' for IPA-is this true?) who would like to be interviewed, but her child is now 22, although she was a child at the time of the life events.

I'm not sure what to do, as time is running out, my supervisor is on leave for the next month, and I don't know enough about IPA (although I am learning fast, but for various reasons I had to change thesis topics in October-at the start of 3rd year, so am not where I would like to be) to understand if this would screw up the research, or whether it could conceivably be accommodated. I have already emailed the university ethics committee to check whether they have any objection to me changing the inclusion criteria and they don't, so this is just about

a) just how homogeneous does IPA have to be? and

b) is 6 really the magical minimum number?

I'd be very grateful for any thoughts/guidance you can offer!

thank you

Pink
Kintsukuroi: 'to repair with gold'. the art of repairing pottery with gold or silver lacquer and understanding that the piece is more beautiful for having been broken.

@Sushipink22

User avatar
enid
Posts: 739
Joined: Wed May 30, 2012 5:03 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by enid » Fri Mar 14, 2014 4:25 pm

As far as I have been taught, there is no magic number as such. Small samples are entirely acceptable, and more than 15 is usually completely inappropriate/unfeasible. I have a friend who did IPA for her thesis and had one participant only. For IPA this is acceptable. Another friend had 11 for an entire PhD thesis. IPA is extremely detailed and complex and I would say that 6 is a really good number. If it is due in in September and you haven't begun analysis yet then I would see this as key, as it really does take a long time to do. You have to transcribe yourself with IPA and the analysis takes a lot longer than say thematic analysis or something.

Personally I would include the 6th person and just discuss all of this in the relevant section, highlighting the fact that numbers were low. You just need to make clear that the data for the 6th person is from that person. I think reasoning for low numbers is the homogenous nature of these samples (IPA usually has quite homogenous samples) due to the niche sample group under study and how lengthy the process is. You use IPA for the novel, and novel samples tend to be small, hence one of thre reasons in practice why IPA samples are small. (Other reason being the level of detail in the analysis/write-up). My friend's (with one participant) thesis was absolutely enormous. & passed viva all fine, etc.

User avatar
moonbeam
Posts: 662
Joined: Fri Apr 09, 2010 11:37 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by moonbeam » Fri Mar 14, 2014 4:32 pm

Hi,

Do you have the Smith, Flowers & Larkin (2009) book? They suggest that 'newcomers' to IPA can use a sample of 3. I'm new to it, but I'm using 6 in order to generate a more detailed information regarding the experiences. I think as long as you've stated that in your thesis, and referenced Smith et al (2009) you should be fine (that's what I have done).

Have you seen the yahoo group? It's really helpful and you can ask questions to members of the group who are/have been using IPA. Michael Larkin himself responds at times which is really helpful - especially for me when I am currently struggling with the epistemology side of things.

In terms of homogeneity, I think the same applies. As long as you can provide an explanation as to why you've used that sample, and the characteristics they posses you should be okay. Smith et al (2009) also provide good info in their book. I've had to use both males & females in my sample, as that would have drastically reduced the amount of people eligible to participate in my study.

Hope everything goes okay, and best of luck!

MB
X

Edit:
I've edited the post to include the link for the yahoo group: http://www.ipa.bbk.ac.uk/discussion-group
Last edited by moonbeam on Fri Mar 14, 2014 5:02 pm, edited 1 time in total.
There may not always be a solution to the problem, but there is always another perspective...

User avatar
Loula
Posts: 1096
Joined: Mon Jun 15, 2009 3:28 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by Loula » Fri Mar 14, 2014 4:40 pm

I've never used IPA (and frankly it looks terrifying!), but just thought I'd add another perspective. Does your university have a minimum sample size considered to be adequate for a DClin? My course is very reluctant for anyone to have less than 6 regardless of the methodology so it's worth checking with your research director.

aspiringpsych
Posts: 120
Joined: Wed Nov 13, 2013 1:45 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by aspiringpsych » Fri Mar 14, 2014 7:27 pm

I completed an IPA research project for my MSc firstly I would say this is a type of analysis which takes a LOT of time do not underestimate the time you will need....also I did a lot of reading researching and advising from my supervisor which told me 6-8 for anyone at MSc level is good, maybe 6 or less if you are undergrad. I really loved doing IPA but it is very time consuming and a rich way to data collect if you get me, so you need to spend a lot of time with the data and immerse yourself in it without sounding too cheesey! so just plan your time well and around 6 is definitely realistic and achieveable! Good luck

User avatar
Pink
Posts: 380
Joined: Sun Aug 21, 2011 8:36 am
Contact:

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by Pink » Sat Mar 15, 2014 10:23 am

Hi Folks,

wow, thank you, I didn't think anyone would reply, and this wealth of knowledge and information is fantastic! thank you all for being so generous. As I have mentioned elsewhere on the board, I am something of a procrastinator *coughs*shuffles feet at the masterly understatement* and have successfully avoided learning even the basics about IPA, so your responses have both informed and terrified me-moving from a state of 'unconscious ignorance' to 'conscious ignorance' is not actually a comfortable process! Thank you all for the good luck wishes, I'm really touched actually. It's nice to feel that people are rooting for you. I don't quite know why it is that I can sit with very high risk/dangerous clinical situations without turning a hair, but the mere thought of putting pen to academic paper turns me into a quivering jelly.

Enid-good to hear re your friend, and very helpful advice about how to write it up/where to reflect on decisions made, thank you.

Moonbeam-thank you so, so much for the book recommendation (ordering on amazon as soon as I've written this!) and the link, very, very, very helpful!! Also for the guidance re sample size, and homogeneity. thank you.

Loula-with you on the terrifying! very good point though and I will email research director today, I really don't want to get a 'major condition' of having to gather more data! Thank you for spotting this potentially enormous hazard-good thinking batman!

aspiringpsych-thank you for sharing your experience and enthusiasm, nice to hear something positive. I'm doing it at doctoral level so hopefully 6 will be enough-I'm going to use moonbeam's helpful quote! I am a bit alarmed by what you and enid have said about how long analysis takes-I've allowed myself a week and that's it! there's always pro plus I suppose..

my thanks again to you and warm wishes,

Pink
Kintsukuroi: 'to repair with gold'. the art of repairing pottery with gold or silver lacquer and understanding that the piece is more beautiful for having been broken.

@Sushipink22

Lancelot
Posts: 256
Joined: Sun Dec 27, 2009 8:07 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by Lancelot » Sat Mar 15, 2014 11:17 am

Hi Pink,

I personally would not go for a 6th and dilute the sample -you have an inclusion criteria for a reason. It sounds like a difficult sample to get so I would just highlight that in the limitations. You could note IPA published literature that has small samples and quote favourable papers on saturation. Also, you probably just need to get on to analyse and write the thesis. The problem as I would see it is not a 6th person but having a thesis (on time). It is surprising how many theses pass with no problem with lower samples than 5!

I hope that is helpful.

User avatar
sarahg
Posts: 614
Joined: Sat Dec 10, 2011 1:46 am

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by sarahg » Sat Mar 15, 2014 3:41 pm

I agree with the points made by others in this thread. It is far better to have depth to your analysis, than breadth. Unlike with statistics, there is no 'one size fits all' or 'magic number' with regards to sample size. As long as you can effectively justify your sample size and sample constraints within your thesis you should be fine :).

User avatar
Pink
Posts: 380
Joined: Sun Aug 21, 2011 8:36 am
Contact:

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by Pink » Tue Mar 18, 2014 8:12 pm

Dear Lancelot and Sarah,

Thank you both for your thoughtful and helpful replies, I really value and appreciate them. Lancelot, I think you have a point about just starting to write the damn thing!

warm wishes,

Pink
Kintsukuroi: 'to repair with gold'. the art of repairing pottery with gold or silver lacquer and understanding that the piece is more beautiful for having been broken.

@Sushipink22

lingua_franca
Posts: 931
Joined: Tue Sep 14, 2010 11:29 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by lingua_franca » Wed Mar 19, 2014 1:06 am

I'm using IPA for my PhD thesis and I'm also a newcomer to it. I second the recommendation of Smith, Flowers, and Larkin's 2009 book - it's a really nice intro. I'm working with different groups of participants, with the smallest group containing five and the largest containing eight. IPA's strength rests solely on the detail of the analysis, not the number of the participants - you can justify a small sample size with this method because its thoroughness precludes working with larger groups.

Good luck with it and have fun. :) It's definitely one of the more interesting research methods I've been introduced to, although aspiringpsych is right, it is time-consuming!
"Suppose a tree fell down, Pooh, when we were underneath it?"
"Suppose it didn't," said Pooh, after careful thought.
Piglet was comforted by this.
- A.A. Milne.

User avatar
moonbeam
Posts: 662
Joined: Fri Apr 09, 2010 11:37 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by moonbeam » Mon Mar 24, 2014 11:28 pm

Pink wrote:I am a bit alarmed by what you and enid have said about how long analysis takes-I've allowed myself a week and that's it! tk
How long have you got? I would definitely stick with 5 participants and get cracking with the transcribing (if you haven't already). It really did take longer than I expected, I think to transcribe 10minutes it took around an hour... and to think some of my interviews were 90 minutes long :shock:

Analysis... give yourself at least a week per participant (if you have chance - I don't and I'm seriously regretting all my procrastination. Dexter really wasn't all that important at the time :oops: ). Get Smith's book... follow the 6 steps in that. The initial noting and the emerging themes (step 2/3) take a long time. Then I have copied their example in the book and relayed this into a table with the line numbers (helps with the clear audit trail aspect of qual - also you can check whether the themes are from you or the participant) as you need to write the line number next to the emerging themes. This helps to pull it all together when comparing the themes across cases (all participants).

I think with me.. I have spent so much time fussing and worrying about the epistemology side of things... when I can tinker with the writing at the end :(

Good luck and feel free to PM me if you have any questions.

MB
X


EDIT: I also thought I'd add that I wished I had time to interview them a second time. Reading through and analysing, I realise that I've missed so many opportunities to follow up questions to identify meaning. I think because I was so new to interviewing in this was (and I was a support worker at the time, so had never been an AP - or interviewed anyone) I feel I stuck to my interview schedule with my at least my first 2 participants. This improved as I gained experience, however I would like to interview them again! Something to consider for the next time I use IPA.
Last edited by moonbeam on Mon Mar 24, 2014 11:47 pm, edited 1 time in total.
There may not always be a solution to the problem, but there is always another perspective...

User avatar
moonbeam
Posts: 662
Joined: Fri Apr 09, 2010 11:37 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by moonbeam » Mon Mar 24, 2014 11:32 pm

lingua_franca wrote: I'm working with different groups of participants, with the smallest group containing five and the largest containing eight.
Wow. That's a hefty amount of transcribing and analysing - I feel like my 6 P's are nothing compared to yours haha!

Good luck with it all and hope it goes well

Mb x
There may not always be a solution to the problem, but there is always another perspective...

User avatar
enid
Posts: 739
Joined: Wed May 30, 2012 5:03 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by enid » Tue Mar 25, 2014 10:42 am

moonbeam wrote:EDIT: I also thought I'd add that I wished I had time to interview them a second time. Reading through and analysing, I realise that I've missed so many opportunities to follow up questions to identify meaning. I think because I was so new to interviewing in this was (and I was a support worker at the time, so had never been an AP - or interviewed anyone) I feel I stuck to my interview schedule with my at least my first 2 participants. This improved as I gained experience, however I would like to interview them again! Something to consider for the next time I use IPA.
I have had the same thoughts re: reflections on missed follow-up probing.

I have to say, for me, coding is taking a full day. So I have 41 participants (not using IPA) and coding takes a day = 41 days. And that is even before I have begun analysis. Qual takes a long time.

edit - thankfully as it's thematic I could pay someone else to do the transcribing.

User avatar
moonbeam
Posts: 662
Joined: Fri Apr 09, 2010 11:37 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by moonbeam » Tue Mar 25, 2014 6:58 pm

enid wrote:
moonbeam wrote:
I have to say, for me, coding is taking a full day. So I have 41 participants (not using IPA) and coding takes a day = 41 days. And that is even before I have begun analysis. Qual takes a long time.
41 participants! wowsers -when you said coding at first I thought you meant analysis haha! I think it took me longer than a day, as it took my ages to re-read it and check I had captured sighs and laughter and ummm and ahhhs etc. It's a good experience though.. although I find me ears hurt from the headphones!

I'll definitely be doing a second interview in the future I think...

Good luck, Enid!
There may not always be a solution to the problem, but there is always another perspective...

User avatar
moonbeam
Posts: 662
Joined: Fri Apr 09, 2010 11:37 pm

Re: Homogeneity in IPA studies, and the 'magic' minimum numb

Post by moonbeam » Tue Mar 25, 2014 6:58 pm

enid wrote:
moonbeam wrote:
I have to say, for me, coding is taking a full day. So I have 41 participants (not using IPA) and coding takes a day = 41 days. And that is even before I have begun analysis. Qual takes a long time.
41 participants! wowsers -when you said coding at first I thought you meant analysis haha! I think it took me longer than a day, as it took my ages to re-read it and check I had captured sighs and laughter and ummm and ahhhs etc. It's a good experience though.. although I find my ears did hurt from the headphones!

I'll definitely be doing a second interview in the future I think...

Good luck, Enid!
There may not always be a solution to the problem, but there is always another perspective...

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest