CODEWITHSUNDEEP
rmarginal-effectsr-marginaleffects
Joe Loffredo
Joe Loffredo

Reputation: 1

Setting user-specified values in plot_comparisions from marginal effects in R

I am using the marginaleffects package in R to analyze some results from a conjoint experiment.

I've created a model object that is in the form of model = svyglm(dv ~ attribute1*respondent_characteristic_1\*respondent_characteristic_2 + attribute2 + ...) .

My goal is to use the plot_comparisions function to analyze the difference in predicted probability of selecting a candidate if attribute_1 changes from one level to another.

I have code in the form of:

plot_comparisons(model,variables = "attribute1",condition = c('respondent_characteristic_1', 'respondent_characteristic_2'))

This produces a plot essentially what I want, but with one problem. It sets all other attributes from the conjoint to its modal value. I want to calculate these comparisions with each attribute set at its baseline value. I tried using newdata to set this, but I think condition overrides that. Any suggestions how to do plot_comparisions but with setting other variables at user-defined values rather than at their mean/modal values?

Upvotes: 0

Views: 38

Answers (0)

Related Questions