I used the Zero-centered raw scores for the online MaxDiff Analyzer. While Im exploring this tool to analyze my MaxDiff Data. I do have a couple of questions that I couldn't get clarity on while going through the documentation.

1. What is the difference between the rescaled and Probability of choice? Which one should be used ? If the probability of choice suggests that the probability for item 1 is 89%. IS it safe to say that the respondent would choose Item 1 as "Best" when referenced against otehr items 89% of the times?

2. I do have to segment the data by 6 categories. How can I test significant differences amongst scores? Again, should i be using probabilities here or scores?

3. Will TURF be a good option to see how groups of these items perform? or can anyone recommend a different approach?

New to Sawtooth and still trying to get my basics right! Any help would be appreciated. Thanks in advance!