Raters who rate a note with a high harassment-abuse tag-consensus model note intercept will get an increased penalty to their raterAgreeRatio.
Work in progress changes to user enrollment: treat earnedOutNoAcknowledge the same as earnedOutAcknowledged going forward, and start counting the number of times earned out going forward.
Re-calibrate low diligence threshold so that it catches the same percent of notes prior to the harassment-abuse change.