- In interviews, must reoccurrence of response be from EVERY SINGLE QUESTION in interview guide for saturation to occur?!
No Umar, saturation occurs when you are not generating any new categories in your model when you add new data.
If you are carrying out semi-structured interviews your interviewees may answer different questions to the ones that you have posed so it would certainly not relate at the individual question level in the way you have suggested.
No. Saturation is about whether you're getting any new information from participants. There will be themes which you identify within the data that reoccur with time, but individual little comments that aren't relevant to any other themes will always crop up. There is no way to ensure that everyone is saying roughly the same thing across interviews - that will never happen. The thing you are looking for, the area of saturation, will be the dominant themes that reoccur across numerous interviews. When you reach a point where no new themes emerge, where most of what they are saying relates to an existing theme, or where new themes do not emerge, that's when you've reached saturation.
For example, a participant may say 10 things and 9 of them relate to existing themes but 1 does not. That doesn't mean you're NOT at a saturation point. That just means that they said something which no other participants have echoed or commented on. It may not be relevant anyway.
There is a difference between "data saturation," where you are not generating any new codes, versus "theoretical saturation," where you are not generating any new insights into your analysis. Theoretical saturation primarily applies to grounded theory, so if you are not using that approach you can concentrate on the simpler process of data or code saturation.
I have attached a useful article by Guest et al. on data saturation.
Like Prof Morgan rightly said, theoretical saturation is the basis and not questions-wise basis.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude