In sum, in Study 2 we introduced a time lag between measurement of the leadership variables and subordinate voice. We found that leader openness (Hypothesis 1) again positively influenced voice, even when controlling for prior level of voice, but transformational leadership (Hypothesis 2) was no longer significantly related to voice. Also,psychological safety again played a mediating role between leader openness and subordinate voice (Hypothesis 3). Finally, we found that performance level moderated the influence of leader behaviors on subsequent voice (Hypothesis 4) in such a way
that the impact of leader behavior was stronger for better-performing subordinates.