For our research we use Intraclass correlations (ICC's), but I can't figure out wether I need to use the Single Measures-value or the Average Measures-value. Our study:

10 children with a rare metabolic disease wore an activity tracker (kind of fitbit) for two weeks, with the first week being the test and the second week being the "retest". We already found the device to be valid in a previous study, but now we want to determine the test-retest reliability of the activity measured by the device (so is week no.1 comparable to week no.2, and weekend 1 VS weekend 2, and the 1st mondaynight VS 2nd mondaynight, etc.).

We used: SPSS > Analyze > Scale > Reliability Analysis > Settings: ICC, Two-way mixed, Absolute Agreement. Are these settings correct? And if so, do we need the single-measures or average-measures one. I really can't find anyone who knows this for this specific study-setup. Thanks so much in advance!