Exploring the multimedia effect in testing: the role of coherence and item-level analysis

Abstract

Educational tests often combine text and images in items. Research shows that including images in test items can influence response accuracy, termed the Multimedia Effect in Testing. This effect suggests that using pictures in tests can enhance student performance and reduce the perception of item difficulty. As such, the Multimedia Effect in Testing could influence test validity. However, research in this area has produced varied and conflicting results, which may be partly attributed to the functionality of the images used. Besides, many studies only offer test-level data, making it challenging to determine whether the outcomes represent a generic phenomenon or result from averaging mixed outcomes in individual test items. This present study examined whether coherency of pictures in tests influences response accuracy, mental effort and time-on-task at the test level and item level. Item-level analysis showed that the Multimedia Effect in Testing is not universal; only a small subset of items showed significant differences between text-only and text-picture items. The degree of coherence also did not give unambiguous results. In summary, the study highlights the complexity of the Multimedia Effect in Testing, suggesting it is context-dependent, with not all test items benefiting equally from multimedia elements. The findings emphasize the need for a nuanced understanding of how multimedia affects educational testing.