Context: I am currently doing my undergraduate research and I am stuck on my analysis part. The study is a quantitative experiment, I had 17 participants performing a maitlands mobilization Grade 3 for 3 times at a time. What I was looking for is if all individuals could exert the same force consistently across the 3 times (i.e. checking for reliability).

Problem: I want to test the Inter-rater reliability and Intra-rater reliability of these participants.

For intra-rater reliability I decided to use ICC in SPSS - One way random and Inter-rater reliability a Two-way random looking at the single measures for both.

However for Intra-rater reliability some of my values came out negative.

Question: Should I be doing a different method? For example, ICC 3,1

How can I interpret the negative values if they are still useful?

  • 1
    $\begingroup$ please try to write clearer sentences -- the clearer you can write your question, the better you understand the question, and the better you understand the question, the more likely you will be able to give others the information they need to help you. $\endgroup$ – Chill2Macht Mar 24 '17 at 17:38

You definitely want to be using two-way models, not one-way model, since there is meaningful variation between both trials and subjects. Random or mixed effects models (ICC2,1 and ICC3,1) will be the same value so that doesn't matter. If you get a negative value that may be because you made a mistake in setting up the data or due to low between-subjects variance. If you describe your data in more detail and more clearly, I can provide further help.

| cite | improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.