To the best of my knowledge, it is the most appropriate to carry out the SEM modeling on the basis of constructs, which were measured through indicators measured on the same sample.
To my knowledge, this is not possible - However, if you find a solution, plaese write me. It needs to be measured on the same sample, so that each reply for both concepts can be attributed to one person.
My opinion is that the indicators should be measured for the same sample, the same time period and under the same circumstances, if we want to conduct as much adequate modeling process as possible,