Structural equation modeling - predictor measured in different group

Dear all,

I am kind of stuk on the following problem:

what if you would like to create a model, with several predictors. only one important predictor is measured in a different sample. The characteristics of this sample are similar.

for example, you would prepare a model to describe a treatment effect. You predict that some proportion of the effect is predicted by a placebo effect. So the eventual treatment effect consists of a placebo effect and the actual threatment. However, the placebo group is the only group in wich the size of the placebo effect is measured. So how can you incorporate this proportion into your model using structural equation modeling?

Do you have any thoughts?

Kind regards,