Given the constraints of COVID-19, our IRB is limiting data collection that can be completed in-person. Thus I am looking to utilize teleadministration and comparing that to previously collected data. With that said, this means we have two separate groups that we are comparing. Is there a way to evaluate agreement between administration modality similar to that of an ICC or Bland-Altman plot?
I was thinking if one can adapt Limits of agreement (LoA) estimate to compare the two method agreement. i.e., the interval within which a proportion of the differences between measurements of the two modality lie.
Given the limit of information you provided, and also about different subjects. Alternative is a non parametric approach, [comparing the results for each administrative modality] given that the responses maybe on likert scale, (ie., If you need to compare the 5-point scales one at a time, then non-parametric statistics) and perhabs the average of several Likert scale items which should gives you a scale that "approaches" interval-scale properties. You can use regular parametric statistics, and with three or more groups, ANOVA.
See this discuss, perhabs it might be similar case. https://www.google.com/url?sa=t&source=web&rct=j&url=https://www.researchgate.net/post/What_is_the_most_suitable_statistical_test_for_ordinal_data_eg_Likert_scales&ved=2ahUKEwiJ9PXnztzsAhWhnVwKHTzFAj0QFjALegQIMBAC&usg=AOvVaw0p9DgbQ_8Nx-7nWuh1C5XG
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics