>The link you provided still doesn't answer the question of union vs non-union at companies with equivalent market dominance.

I'm sure if you came up with a meaningful measure of market dominance you could find the data to test your little theory.