I hope this is an easy one:
I have a price variable with 5 levels and specified it as log-linear after assessing count analysis and part-worth results. The design matrix translates my 1, 2, 3, 4, 5 into -0.612, -0.14, 0.101, 0.264, 0.388.
I noticed that the transformed values are zero-centered and have a span of 1 by default. Now my questions are as follows:
1. how did we get to the transformed values from 1, 2, 3, 4, 5.
2. what do I have to pay attention if I were to use estimated parameters for calculations of shares of preferences outside of MBC? Do I have to transform the input variable beforehand?