I want to investigate the reliability of test-retest measurement of some cognitive functions (working memory, reaction time) by computer software in healthy subjects. Subjects will be tested on two occasions with 10 days apart in the same conditions and location. The tests will be done on computers and data will be extracted after for analysis.
By reading the available literature I have found several possibilities for statistical analysis of data. There are examples were simple correlation was done (r) for first and second set of data. Also paired t test was done. Some authors used ICC (1,1) method. And there is Bland-Altman plot. I am not sure which one to use. I know there is Bland Altman modification for repeated measures, since it was originally used for comparing two methods.
I would greatly appreciate any suggestion on this issue.
ICC and Bland Altman are suggested. Pearson correlation coefficient and t test have only indirect explanations and could mislead.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Computer Software