I am using the anchored MDS utilities from the Sawtooth estimation module. In it, I get utility of anchor point as zero and the utility value of other attributes. But they are not centered to zero for each respondent.
So, when I calculate the choice likelihood using logit, should I zero center the utilities (excluding anchor point) or just use them as such.
In other words, when using logit analysis in MaxDiff, is it 100% necessary for the utilities to be zero-centered for each respondent?