choice-metrics.com

I have a doubt which is as follows:
Suppose, you are working with a choice task involving two unlabeled alternatives, where:
Alt-1 represents an existing technology.
Alt-2 represents an upcoming new technology.
Both alternatives share generic attributes, but for one key attribute, say Z, Alt-2 always has a specific level fixed to 1, while Alt-1 has it fixed to 0. This fixed level is crucial because Attribute Z distinguishes the new technology from old , and is the basis for estimating the willingness-to-pay (WTP) differentiation.

Doubts:
Do you need to explicitly specify a constraint in Ngene during design phase, so that this particular attribute level is fixed at 1 for Alt-2 and at 0 for Alt-1 in every choice task?
How can this command look like? In estimation Alt-2 should then have an ASC?
If no constraints are needed, and the levels for that attribute are allowed to alternate between the two alternatives, how will this affect model estimation? Will it not undermine the distinction between two alts, especially in analyzing WTP for Alt-2 over Alt-1, with socio-demographic variables, later?

An attribute that has a fixed level for one alternative and another fixed level for the second alternative is essentially part of the label of the alternatives. In that case, it is captured in the constant. Of course you can still show the fixed attribute levels to respondents, but in your utility function you only have a constant that captures the combined effect of these fixed attribute levels as well as the labels "existing technology" and "new technology". You will not be able to disentangle the effect of the label and this attribute unless you allow the attribute levels to vary.

Michiel

Thank you. This clears the doubt I had.