A paper from the 2013 STEM Annual Conference.
The automatic marking of short-answer free-text questions presents an opportunity to assess and provide personalised feedback to large classes with savings of cost and better consistency than human markers and a more authentic assessment experience for students than is obtained from the use of multiple-choice questions alone. Short-answer free-text e- assessment questions have been in regular use at the Open University Science Faculty since 2006. A high marking accuracy has been retained following a move to pattern-based answer matching and student engagement is good. A similar question type has been added to the Moodle quiz engine. However questions of this type are underused by academics. This paper discusses some reasons for this including the need for a large number of marked student responses for use in developing the answer matching.