We investigated how majority-group witnesses' responses to bias are evaluated.
Confronting is perceived to be a more appropriate response than not confronting.
The exception is when the personal costs of confronting are high.
In this case, evaluations of responses to bias are moderated by observers' values.