So I may be totally wrong but if you suddenly found that the model had issues with some depth.
Would you:
- examine the underlying assumptions and check if something was missed in the original study
- Slap a quick fix to make it more conservative by drawing a line in the middle of two other lines
I am not saying #2 is stupid, as it is better to be alive than dead, but by putting quick fixes you may sometimes be at risk of missing something more fundamental ?
For example, my understanding is that people use different GFs but I have not met many people who can explain to me why they use this set of values and not something else.
This being said, I am a fairly pragmatic person and won’t lose too much sleep over that
Note: I didn’t read the original research and I do not know how the GFs were decided to be a good solution compared to any other alternative. Also my experience is very limited ...