The effects of changing leader power: A test of “situational engineering”