DIGITAL LIBRARY
PERFORMANCE EVALUATION OF AN AUTOMATED TEST GENERATION SYSTEM IN PROGRAMMING TRAINING
University of Library Studies and Information Technologies (BULGARIA)
About this paper:
Appears in: EDULEARN20 Proceedings
Publication year: 2020
Pages: 2058-2067
ISBN: 978-84-09-17979-4
ISSN: 2340-1117
doi: 10.21125/edulearn.2020.0653
Conference name: 12th International Conference on Education and New Learning Technologies
Dates: 6-7 July, 2020
Location: Online Conference
Abstract:
Education is now considered as a service and hence is a part of the Service Industry, following the changes, developments, and innovation opportunities like other sectors of the Service industry. The usage of electronic learning tests has become not only a regular form of assessment but also a tool for self-learning. After reviewing existing e-test generation systems, the problem that has been identified is the lack of integrated automated knowledge verification systems to replace the manual work of creating the test question and tests with automated. In order to support the teaching work on the creation of e-tests in programming, the Automated Test Generation System "TestLab" was developed. The integrated approach in the Testlab provides for generating individual e-tests through templates and quickly generates and verifies exam tests. The introduction of Testlab in the learning process aims to train students to read program code.

In connection with the introduction of the Automated Test Generation System in the training process, a study has been made to take into account the trainees' reactions when working with the Automated Test Generation System and the demonstrated knowledge after working with it. In the study involved 137 students, 1st year full-time and part-time students majoring in Informatics and Computer Science, who were given the opportunity to prepare their own tests using the Automated TestLab Test Generation System and learn by solving of tests. The study group used an accelerated knowledge testing approach, while the control group was trained using the traditional approach. After each test is completed, the students will check their results themselves. Improved scores were reported for the number of correctly resolved test questions and time to resolve each test.
Keywords:
Evaluation, effectiveness, study, automated test generation system.