Effects of three types of potentially biasing information on symptom severity judgments for major depressive episode