Measuring the cognitive load induced by subtitled audiovisual texts in an educational context
Abstract
Audiovisual aids are nowadays commonly utilised in classroom environments where lecturers are able to supplement verbal instruction with pictures or videos to enhance learning. To make these resources more accessible, subtitles are added to make them easier to understand. Although much research has already been done on the effect of subtitles on cognitive load, there is to date no conclusive evidence of the benefits of subtitles or the hindrances that they may cause. The aim of the study was to provide clear evidence of the effect of subtitles on cognitive load (CL) by looking at their effect on processing different amounts of information, the effect of different types of subtitles (verbatim or edited) and how the composition of subtitled stimuli (containing redundant and non-redundant information) affects CL. Two experiments were conducted. The first was exploratory, to determine the effects of subtitles on CL. The participants (n=64) watched a recorded lecture in one of four presentation modes: 1) audio only, 2) audio and video, 3) audio and video with verbatim subtitles, and 4) audio and video with edited subtitles. No significant differences were found for either the CL experienced or the performance between the presentation modes. The second experiment was more comprehensive than the first and included the recording of eye-tracking data and personal data (such as English proficiency, working memory capacity, etc.). The participants (n=23) watched four recorded lectures, randomly presented in one of the four presentation modes (the same as in the first experiment). The results indicated no significant difference for either CL or performance between the presentation modes. However, a linear mixed effect model indicated that the participants focused longer (higher CL) on the verbatim subtitles then on the edited subtitles (+23.41 ms). Significant differences were also found with the CL of subtitles, where edited subtitles imposed 52% less cognitive load than verbatim subtitles, but were 24% less likely to be processed in the presence of redundant information. A significant difference was also found regarding the processing of subtitles in the presence of redundant information, as edited subtitles are 24% less likely to be processed while in the presence of redundant information, compared to the verbatim subtitles. Edited subtitles were also found to be 45% more likely to be processed than verbatim subtitles. This study seems to indicate that subtitles do not have a significant effect on either CL or performance, but that the difference is rather between different types of subtitles and how they are composed (the amount of redundant information included).
Collections
- Humanities [2671]