Pregled bibliografske jedinice broj: 1094050
Computer-based question and exam evaluation in summative knowledge assessment
Computer-based question and exam evaluation in summative knowledge assessment // 43rd MIPRO 2020 International Convention - Proceedings / Skala, Karolj (ur.).
Rijeka: Institute of Electrical and Electronics Engineers (IEEE), 2020. str. 1820-1825 doi:10.23919/mipro48935.2020.9245326 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 1094050 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Computer-based question and exam evaluation in summative
knowledge assessment
Autori
Tomić, Siniša ; Paunović, Vlatka ; Bosnić, Ivana
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
43rd MIPRO 2020 International Convention - Proceedings
/ Skala, Karolj - Rijeka : Institute of Electrical and Electronics Engineers (IEEE), 2020, 1820-1825
Skup
43rd International Convention on Information and Communication Technology, Electronics and Microelectronics (MIPRO 2020)
Mjesto i datum
Opatija, Hrvatska, 28.09.2020. - 02.10.2020
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
e-Learning , Classical Test Theory , Item Response Theory , automated exam evaluation , automated hybrid exam verification
Sažetak
This paper describes a system for automatic multiple-choice pen- and-paper exam generation, which includes evaluation of students’ answers and evaluation of the quality of exams and questions. During the design phase examiner defines the number of questions, area for each question, and the number of variations of the exam. In the evaluation phase, the system automatically evaluates the students’ answers. In the verification phase, the system uses Classical test theory (CTT) concepts to provide feedback to the examiner highlighting all questions identified as too easy or too difficult for students. When using the KR-20 formula to verify more than 100 exams and 5, 000 students, results show that the reliability coefficients are higher than 80% on average, which is considered very high.
Izvorni jezik
Engleski
Znanstvena područja
Računarstvo
POVEZANOST RADA
Ustanove:
Fakultet elektrotehnike i računarstva, Zagreb