I measured a Psychometric parameter for 60 participants.
I collected the data thrice, so I acquired 3 values for that parameter, of which I calculated the mean value.
Since the data was relative in nature, it was necessary to check its repeatability/reliability.
Hence, I invited all the 60 participants again and re-did the study and measured the parameter again for 3 trials. And again, calculated the mean.
Both the measurements were done by only 1 individual for all the participants and his role does not affect the data in question.
I was not sure about which ICC model should I use, I referred to one of the articles by Koo and Li (2016). Based on what I understood, for test-retest reliability they suggest using ICC(3) with two-way mixed model and absolute agreement.
But for me I am not interested to check rater effect. I wish to check the reliability/repeatability of my data, if it is remaining similar or is it changing. So I am not sure if I should use ICC(2) model instead,as in 2-way random model with absolute agreement.
Can you please advise me on how I should go ahead? Your help would be highly appreciated. Thank you!
Best regards,
Parth Shah,
Parth Shah, I wrote an article about intraclass correlation coefficients about four years ago, and I have put a copy onto RG under my name. I think you should be able to find the information you want there. The Koo and Li (2016) article is good, but I might approach things from an angle that is more meaningful for you.
Here is the "official" reference to my article in its journal in case you need it:
Trevethan, R. (2017). Intraclass correlation coefficients: Clearing the air, extending some cautions, and making some requests. Health Services and Outcomes Research Methodology, 17, 127–143. https://doi.org/10.1007/s10742-016-0156-6
All the best for your research.
Thanks a lot Sir for sharing the link. I will go through it and get back to you in case if I need any further guidance or advice. I really appreciate your help of sharing the link for relevant article.
Best regards,
Parth.
Parth Shah, you're welcome, Parth. Yes, do feel free to get back.
Please note that different researchers, and SPSS, use slightly different "terminology". In my article, I try to shed light on that, and for test-retest situations I state that the 2-way mixed model is most appropriate. Refer to Table 1 in my article.
If I understand your situation correctly, yes, I think you should use absolute as your type of ICC. Furthermore, you could validly use averaged measures as your form because you took three readings from each participant on each occasion.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis