Text this: Rater accuracy and training group effects in Expert- and Supervisor-based monitoring systems