# Anchored MaxDiff and adjusting the scale to 0-1 probability

Hi all,

My data is a conjoint MaxDiff (BWS profile case). I had a question in the survey rating an anchor (death) to the items (levels of dimensions) based on the on-the-fly calculations.

Now my understanding is that the anchor could be valued as 0, which is one end of the intended scale. But I want the other end of the range to be bounded at 1. That means the addition of the values of all best levels from every the dimension should be 1
(6 dimensions in total & 31 levels, 5 dimensions have 5 levels, and one have 6 levels)How can I assure this in my results?

There is an idea of a linear transformation of the values of the levels, would this work here? And how can I apply it to the anchored results as the anchor (Death) is equal to 0?

Note: Death is considered a scenario that is initially composed of 6 items (one of each dimension), but in the anchoring model, it is included in the model and thus compared to each of the 31 items not the scenarios composed of them.

Explanation of the linear transformation:
let's say my cases are described by 2 Dimensions "Cost & Distance" & each of the dimensions have 3 levels:
Cost: 1-10\$ 2-50\$ 3-100\$
Distance: 1-10 Km 2-5Km 3-1 Km
Anchor: stay where you are (supposed 0 cost & 0 distance but it is not described this way)
The best possible case is the combination of 10\$ & 10 Km, or (1,1) & its value is Vb. similarly, the worst case is (3,3) & its value is Vw.
If we assume that Vl is the value of any of the 6 levels resulted from any of the 3 models (Logit, LC, HB), then the linear transformation is calculated as:
[Vl - (Vb/2)]/Vw (in case of no anchor, but what if we have an anchor replaces the Vw)
With modifications to suite the estimates of each the estimation method
So the aim is to have the anchor valued at 0, the best case at 1 & all the other combinations scaled in between (-1 & +1)

Regards'