Create new variables which contain the preference shares for the alternatives in a MaxDiff latent class analysis, MaxDiff hierarchical Bayes or MaxDiff ensemble output. The shares are computed from the individual-level coefficients generated by the MaxDiff analysis.
Example
The script produces a SUMMARY table with the new variables:
Technical details
The formula used to obtain the share for alternative \(j\) of an attribute is:
\[S_j = \frac{e^{\beta_j}}{\sum_je^{\beta_j}}\]
where the sum ranges over all of the alternatives, and \(\beta_j\) is the individual-level coefficient for alternative \(j\).