choice-metrics.com

Hi there,

On page 112 of the Ngene Manual (version 1.2.0), there is a quote "Example syntax is presented below for the color example, where the first color level has been assigned a prior parameter value of -0.15, the second 0.45 and the final omitted level, a value of zero."

The syntax described is:

U(alt1) = color.dummy[-0.15|0.45] * color

Is it possible to recode the 'final omitted level' to be 1 instead of 0?


The reason I want to do this is that I am using priors from health economics DCEs that are scaled from 0 (worst possible state) to 1 (best possible state).

This is a made up example but let's say I have a pain attribute with 3 levels that represent severe pain, moderate pain, no pain:

Pain.d[0.76 | 0.87]*Pain[0,1,2]

It then assumes that my level 2 (no pain) is 0, which is quite far from the truth (it's probably close to 1).

You cannot change dummy coding, which always has a reference level with zero utility, but you can change the attribute level that is the reference and you can change the order of the levels.

So instead of Pain[0,1,2] you could write Pain[2,1,0], where 0 is now the reference level with zero utility. The last level is always the reference level in Ngene, but you change choose the order of 0, 1, and 2.

Michiel

Thank you! I think that should work