Abstract
In order to gain reliable scores from constructed-response tests — often known as performance tests — it is critical that human raters use standard methods of evaluating and rating examinee responses. The 13th issue of ETS’s R&D Connections series describes some of those methods and the quality controls that ensure their effectiveness.
Original language | English |
---|---|
Publication status | Published - Feb 2010 |
Keywords
- Automated scoring
- Backscoring
- Constructed-response scoring
- Method validity
- Trend scoring
Disciplines
- Educational Assessment, Evaluation, and Research