by Michiel Bliemer » Tue Aug 17, 2021 6:29 pm
Preference-space model:
V = b1 * x1 + ... + bcost * cost
WTP is w = -b1/bcost
95% confidence interval: [w - 1.96 * se(w), w + 1.96 * se(w)], where standard error se(we) needs to be computed using the Delta method.
WTP-space model:
V = lambda * (b1 * x1 + ... - cost)
WTP is w = b1
95% confidence interval: [b1 - 1.96 * se(b1), b1 + 1.96 * se(b1)], where standard error se(b1) comes directly out of model estimation.
As you can see from the above, determining confidence intervals in WTP space models is straightforward, it just requires parameter estimate b1 and its corresponding standard errir, much easier than in preference-space models.
Michiel