There are two options for exporting utilities for CBC, ACBC, ACA, or CVA conjoint analysis studies to a .CSV file:
Raw
The original utility weights as computed by the score estimation method are exported. Depending on the response error (in the case of CBC or ACBC) or the scale-use tendencies of respondents (in the case of ACA or CVA), each respondent's utilities can have a different "scale factor," meaning that the utilities may be uniformly larger or smaller in magnitude between respondents.
Raw utilities have appropriate scaling for conducting market simulations, but we recommend the normalized Zero-Centered Diffs procedure if the purpose is to compare respondents or groups of respondents on the utility weights or to otherwise use the utility scores in tabulations or statistical procedures where it is assumed that the utility weights are on the same scale across respondents (e.g. regression analysis, cluster analysis, factor analysis, etc.).
Zero-Centered Diffs
These are normalized utility scores where for each respondent a) within each attribute the utilities are made to sum to zero (are zero-centered) and, b) across attributes the differences between best and worst utilities average 100. Zero-centered diffs remove scale factor differences between respondents such that making comparisons between respondents is more robust.
Zero-centered diffs are more appropriate to use when the purpose is to compare respondents or groups of respondents on the utility weights or to otherwise use the utility scores in tabulations or statistical procedures where it is assumed that the utility weights are on the same scale across respondents (e.g. regression analysis, cluster analysis, factor analysis, etc.).