Well done Vernellia! Yes, that is what weights are expected to do.
Just to clarify details: it is not the mean what is used for assessment evaluation. The current evaluation method just picks one assessment that represents all given assessments best (by comparing the differences in the assessment form fields). This assessment is then used as a reference for evaluating all the others.
If a teacher's assessment has weight, say, 4 while other assessments have weight 1, the evaluation algorithm thinks that four independent reviewers agreed on a single assessment. Therefore the chance is high that this teacher's assessment is picked as the reference.