Unequal sample sizes in AMOS

Hi at first my English is not very good, so I am sorry for mistakes. And I am absolut new in AMOS.

I have a question regarding dealing with unequal sample sizes in AMOS.

The predictor variables are datas from Students (5th grade, and 7th grade), the dependent variable is extant of teacher datas.

The sample with the students from 5th grade consists 150 students, the sample of the 7th grade consists of 185 students and the sample of the teachers consists of 300 teachers. So the sample sizes are unequal. I want to do a path analysis between the students and the teacher reports.
I need to keep all the data of the teachers, so I can`t exclude anyone. Is it possible to estimate the missing students data? Or will this skew my calculations?

And another question, I have e mediation in my model, do I need to run a test for the significance of the mediation, or is it enough if the direct path doesn`t get significant anymore? And one of my predictor is a dummy variable (yes, no), does this change any of the steps in the path analysis?

And did I read right, that the control variables can`t influence the exogenously variables? One of my control variables is gender of the students, and I think it will influence the pedictor variables because this are students data?

Sorry for confusing you with my writing and so many questions. I would be so happy if someone could help me. I have books, but they don`t answer this "simple" questions.

Thank you very much!