how can I measure repeatability of several measurements on 4 different elements carried out by 10 users?. Each user measured each of the 4 elements 100 times, obtaining at the end 1000 measurements for each of the elements.
When you write "repeatability" do you mean that you are measuring individuals/unit at several time points? What are you looking to report?
I want to report the consistency to achieve identical results across multiple tests. In this case is a computational tool that perform the measurements controlled by the user. The measurement unit is meters (elevation), the four different element or cases are 4 topographic profiles. and the users are 10 students.
The ultimate goal is to detect if this tool have both a high accuracy and high repeatability. Accuracy is not a problem, i can use the standard distribution of the measurements for each profile or element. But i have no idea how to measure repeatability in this specific case.
I think you should consider intra-class correlation (ICC) for this problem.
Julius:
I think that for your design you must look for repeatability and reproducibility. You can follow the procedure explained in this links that may help you:
http://www.engineeredsoftware.com/papers/msa_rr.pdf
http://www.redalyc.org/articulo.oa?id=84903579 (paper in Spanish).
If you perform a two-way anova, this test will give you more information interpreting the hypothesis test: the reproducibility is OK if there are no-significant differences between topographic profiles and between users.
ICC is OK but only to conclude that two sets of measures have some degree of concordance and don't quantify the performance.
Gook luck!
The final way I did it is here. I measured the repeatability of my software TerraceM .
Article TerraceM: A MATLAB (R) tool to analyze marine and lacustrine...
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping
- Maps