Attribute Gauge Evaluation when More Than Two Ordinal Ratings
Dec 7, 2016 8:44 AM
Ordinal Rating Attribute Study.jsl
The attribute gauge analysis in JMP currently handles two levels of rating: pass or fail. It computes the probability of a false alarm, the probability of a miss, the escape rate, and the rater bias. This standard analysis has been extended to more than two levels when the ratings have an ordinal interpretation. For example, a unit might be rated into one of three grades, where grade 1 is best, grade 2 is second best, and so on.
Open the data table, then open and run the script. Each rating may entered on a separate row or a count may be used in the Freq analysis role. One column must indicate the true rating and another column indicates the test rating to be evaluated.. Here is a simple example for illustration:
Standard is the true rating and Test is the result for evaluation. Run the script and enter the data columns in the proper analysis roles:
Click OK and the results appear at the bottom of a standard Contingency platform window: