Abstract
This research aimed to investigate animation-based achievement tests according to the item format, psychometric features, students’ performance, and gender. The study sample consisted of 52 fifth-grade students in Samsun/Turkey in 2017-2018. Measures of the research were open-ended (OE), animation-based open-ended (AOE), multiple-choice (MC), and animation-based multiple-choice (AMC) achievement tests developed for the unit “Motions of the Earth” of the subject area “Earth and Universe” of the science course. Tests were compared to item difficulty, item discrimination, internal consistency levels, and students’ performance and gender. Additionally, predicting students’ science grades was tested by animation-based open-ended and multiple-choice tests. Paired sample t-tests, Pearson’s correlation coefficient, Friedman test, Wilcoxon test, two-way mixed ANOVA tests, and multiple regression analysis were applied to analyze research data. Research results show no significant difference between animation-based and traditional tests' validity and reliability levels. Research result also shows that students’ test scores were significantly higher at animation-based tests than traditional tests. Besides, students’ test scores differed according to the animation type used in tests. However, gender was not a significant variable on students' test scores. Finally, despite using animation as a significant predictor of Science grades, just animation based multiple-choice test has significantly predicted students’ Science grades. Future research can investigate variables that affect students' participation in animation-based tests and their opinions on animation-based tests. The finding regarding the animation type variable can be investigated in-depth in terms of the effect levels of the visual and auditory elements added to the animations by including an equal number of items on the test.