Dear all,
I understood that it is necessary to conduct scaling (using zero-centered diffs) on part worth utilities in order to aggregate individual level utilities (=part worth utilities) to sample level utilities. Thus one is looking for a way to normalize the values for a subsequent meaningful averaging across the sampleĀ“s individuals. Intuitively I thought this would include the utilities of ONE ATTRIBUTE across the to be aggregated sample of ALL INDIVIDUALS. Now it seems that ALL ATTRIBUTES are considered for ONE INDIVIDUAL when using the zero-centered diffs.
So my assumption here is that the aggregation process of individual level to sample level utilities consists of (1) Estimation of individual utilities; (2) Scaling / Normalization (zero-centered diffs); (3) Averaging. Furthermore the SSI Web Manual says: "This [zero-centered diffs] normalization transforms the raw utilities for each respondent to a scale wherein the average difference between best and worst levels across attributes is equal to 100."
In a nutshell I donĀ“t get how the zero-centered diffs scaling enables the averaging, since it considers all attributes for one individual instead of one attribute for the entire sample. Can anyone shed some light on this please? I probably have an error in reasoning at this point.
Thank you very much!
Best,
Chris