Binary exogenous response variable (Willingness to pay)

[Nigel Sainsbury]

Hi folks,

New to the forum and first post.

I am designing a survey in which three separate survey instruments will use a 7-point Likert-type scale to observe three types of emotional response to environmental degradation. The survey will also employ the contingent valuation single bounded dichotomous choice method to ask participants if they would be willing to pay to advance environmental conservation. So my exogenous response varuiable will be binary categorical (yes/no) for seven different monetary values. Each participant receives a single monetary amount to answer yes/no. 

I am asking the data whether there is a relationship between the three emotions and willingness to pay. If this were a simple willingness to pay study with exogenous predictors, a logit/probit regression analysis would likely be used. 

Given the endogenous predictors (emotions), I would like to use a SEM in Lavaan to conduct the analysis. While I have seen similar methods using MPlus and Amos in the literature in the last few years, I am yet to come across an analogous approach using Lavaan and have not found answers in this dicsussion group.

I would be grateful to hear advice from the community (and in particular those using this approach in contingent valuation studies) as to the implementation of a model for categorical exogenous response variables. For example, is it as simple as the Lavaan online tutorial describes ("all you need to do is to recode it as a dummy (0/1) variable. Just like you would do in a classic regression model.") to deal with non-normality in the exogenous response variable or are other considerations/steps required.

Many thanks in advance,

Nigel