A closer look at split visual attention in system- and self-paced instruction in multimedia learning

Two experiments examined visual attention distribution in learning from text and pictures. Participants watched a 16-step multimedia instruction on the formation of lightning. In Experiment 1 (N = 90) the instruction was system-paced (fast, medium, slow pace), while it was self-paced in Experiment 2 (N = 31). In both experiments the text modality was varied (written, spoken). During learning, the participants’ eye movements were recorded. Results from both experiments revealed that learners spent more time studying the visualizations with spoken text than those with written text. In written text conditions learners consistently started reading before alternating between text and visualization; moreover, they spent more time reading the text than inspecting the visualizations. While in Experiment 1 additional time that was made available in conditions with a slow or medium instruction pace was spent inspecting visualizations, in Experiment 2 longer learning times resulted from reading the text more intensively. With respect to learning outcomes (retention, transfer, and visual memory) Experiment 1 revealed an effect of text modality for visual memory only. In Experiment 2 no modality effects were found. Instruction pace was hardly related to learning outcomes. Overall, the results confirm prior findings suggesting that the distribution of visual attention in multimedia learning is largely guided by the text.     

Learning by drawing: When is it worth the time and effort?

Two experiments compared the effects of learning by drawing to studying instructor-provided visuals on learning outcomes, learning time, and cognitive load. College students studied a text on the human circulatory system and completed comprehension and transfer tests. In Experiment 1 (N = 107), students studied the text with provided visuals (provided visuals) or generated their own drawings from the text with text-based support (verbally-supported drawing) or without support (unsupported drawing). Results showed that while the verbally-supported drawing condition spent significantly more time and experienced significantly higher cognitive load than the provided visuals condition, there were no differences across the three conditions in learning outcomes. In Experiment 2 (N = 85), students studied the text with provided visuals (provided visuals) or generated drawings from the text with provided visuals as feedback (visually-supported drawing). Results showed that the visually-supported drawing condition spent significantly more time and experienced significantly higher cognitive load than the provided visuals condition but also performed significantly better than the provided visuals condition on the comprehension test. These findings suggest generating drawings prior to studying provided visuals is worth the time and effort.     

How a picture facilitates the process of learning from text: Evidence for scaffolding

Three experiments were conducted to study on a more fine-grained level how processing a picture facilitates learning from text. In Experiment 1 (N = 85), results from a drawing task revealed that the global spatial structure of a pulley system picture was extracted even from its brief inspection (for 600 ms, 2 s). In Experiment 2 (N = 105), students who initially inspected the pulley system picture (for 600 ms, 2 s, or self-paced) had better comprehension of the system's functions and made more eye movements in line with the system's global spatial structure when listening to text than students who listened to text only. In Experiment 3 (N = 39), students who first saw the picture (for 2 s) processed written text of the pulley system's spatial structure more efficiently than students who read text only. Results suggest that global spatial information extracted from the picture was used as a mental scaffold to facilitate mental model construction.     

How repeated studying and testing affects multimedia learning: Evidence for adaptation to task demands

Two biases can occur in multimedia learning: overconfidence and over-reliance on text processing. The present research sought to identify these biases and to investigate whether they can be reduced, and hence learning fostered, when studying and testing are repeated. In 2 experiments (Exp.1: N = 79, Exp.2: N = 52), students learned either with text only or with text and pictures (multimedia) about how the toilet flush works, gave judgments-of-learning (JOLs), were tested on the learning contents; afterwards this study-test cycle was repeated. Results from both experiments revealed stronger overconfidence due to multimedia in both study-test cycles (JOLs higher than learning outcomes). Eye movement data showed a relative increase in attention on the picture versus text from cycle 1 to cycle 2; this relative increase in attention was related to better learning outcomes. Repeated studying and testing thus helped to reduce over-reliance on text processing in multimedia learning, fostering performance.     

Do drawing tasks improve monitoring and control during learning from text?

In two experiments it was investigated how drawing as a monitoring task affects self-regulated learning and cognitive load. To this end, participants (Exp. 1: N = 73, Exp. 2: N = 69) were randomly assigned to one of two conditions. In the experimental condition, students were asked to read an expository text on the formation of polar lights consisting of five paragraphs, whereby, after each paragraph, they had to create a drawing of the text's content. In the control condition, students read the same text, but performed no drawing task. In both conditions, students had to give judgments of learning (JoLs) after each paragraph and after reading the whole text as well as rate their cognitive load. Then, they were asked to select paragraphs for restudy. In Experiment 1, participants continued with an assessment of their learning outcomes immediately after their restudy selection, whereas in Experiment 2 they were first given the opportunity to actually restudy the selected paragraphs before working on the posttest. Results of both experiments indicate that JoLs rather than cognitive load predicted posttest performance. Moreover, students in the drawing condition compared with the control condition exhibited more accurate (relative) monitoring in Experiment 1 in that their JoLs were more strongly related to performance. Moreover, JoLs predicted students' restudy decisions in both experiments; however, this effect was by-and-large independent of whether they had to draw. Overall, results hint towards the potential of drawing to support metacognitive monitoring.     

Investigating gaze behavior during processing of inconsistent text-picture information: Evidence for text-picture integration

In two experiments, eye tracking was used to investigate whether learners construct a mental representation during learning that integrates information from text and pictures. The experimental groups received inconsistent text-picture information on one or two pages of the learning materials. The control groups received only consistent text-picture information. It was expected that learners of the experimental groups should have difficulties in integrating text-picture information when faced with the inconsistencies. This should be reflected in their gaze behavior. Experiment 1 (N = 51) and Experiment 2 (N = 45) confirmed that assumption for several eye tracking variables. Regarding learning outcomes, only in Experiment 1 worse performance of the experimental group was observed. Furthermore, Experiment 2 revealed that the majority of learners did not remember the inconsistency between text and picture when asked for it after learning. In sum, the results add to our understanding about the cognitive processes underlying multimedia learning.     

Effects of creating video-based modeling examples on learning and transfer

Two experiments investigated whether acting as a peer model for a video-based modeling example, which entails studying a text with the intention to explain it to others and then actually explaining it on video, would foster learning and transfer. In both experiments, novices were instructed to study a text, either with the intention of being able to complete a test (condition A), or being able to explain the content to others (condition B and C). Moreover, students in condition C actually had to explain the text by creating a webcam-video. In Experiment 1 (N = 76 secondary education students) there was no effect of study intention on learning (A = B), but explaining during video creation significantly fostered transfer performance (C > B; C > A). In Experiment 2 (N = 95 university students), study intention did have an effect on learning (C > A; B > A), but only actual video creation significantly fostered transfer performance (C > A).     

Signaling text-picture relations in multimedia learning: A comprehensive meta-analysis

The signaling principle recommends emphasizing relevant aspects of a multimedia message by means of signals (e.g., color coding). We determined the effectiveness of signals that highlight correspondences between text and pictures as well as possible boundary conditions by means of a meta-analysis. To this end, 58 potentially relevant articles were identified in a comprehensive search. After rating the studies based on inclusion criteria and correcting for biases, 27 studies were included in the meta-analysis yielding 45 pairwise comparisons with N = 2464 participants. Domain-specific prior knowledge, pacing of the materials, pictorial format, mapping requirements, and distinctiveness of signals were coded as moderators. For transfer and comprehension performance a positive small-to-medium effect size (r = .17, 95% CI [0.11, 0.22]) favoring signaled multimedia material was found, which was moderated by prior knowledge. The findings support the effectiveness of the signaling principle in particular for learners with low prior knowledge.     

Cognitive styles and instructional design in university learning

Changes in conceptualization and measurement of the verbalizer–visualizer dimension led us to re-examine the hypothesis that students learn best when instructional material matches their cognitive style. First-year psychology university students (n = 41) studied information on three personality theories presented in text only, text+picture, or text+schematic diagram format, demonstrated recall and comprehension of each theory, and completed an adapted cognitive styles questionnaire. Spatial and object visual scale scores were not correlated, but the latter showed a significant though relatively weak negative correlation with verbal scale scores. Recall could be predicted from students' verbal and object visual scores when presentation format matched these cognitive styles. All three styles significantly predicted students' comprehension, but only when they matched the presentation format. The results support the distinction between spatial and object visual styles, and provide evidence that learning outcomes improve when instructional material is matched to students' cognitive styles.     

Differential Effects of Salient Visual Events on Memory-Guided Attention in Adults and Children

Both salient visual events and scene-based memories can influence attention, but it is unclear how they interact in children and adults. In Experiment 1, children (N = 27; ages 7–12) were faster to discriminate targets when they appeared at the same versus different location as they had previously learned or as a salient visual event. In contrast, adults (N = 30; ages 18–31) responded faster only when cued by visual events. While Experiment 2 confirmed that adults (N = 27) can use memories to orient attention, Experiment 3 showed that, even in the absence of visual events, the effects of memories on attention were larger in children (N = 27) versus adults (N = 28). These findings suggest that memories may be a robust source of influence on children's attention.     

Learning by writing explanations: Is explaining to a fictitious student more effective than self-explaining?

Research has demonstrated that oral explaining to a fictitious student improves learning. Whether these findings replicate, when students are writing explanations, and whether instructional explaining is more effective than other explaining strategies, such as self-explaining, is unclear. In two experiments, we compared written instructional explaining to written self-explaining, and also included written retrieval and a baseline control condition. In Experiment 1 (N = 147, between-participants-design, laboratory experiment), we obtained no effect of explaining. In Experiment 2 (N = 50, within-participants-design, field-experiment), only self-explaining was more effective than our control conditions for attaining transfer. Self-explaining was more effective than instructional explaining. A cumulating meta-analysis on students’ learning revealed a small effect of instructional explaining on conceptual knowledge (g = 0.22), which was moderated by the modality of explaining (oral explaining > written explaining). These findings indicate that students who write explanations are better off self-explaining than explaining to a fictitious student.     

Training self-assessment and task-selection skills: A cognitive approach to improving self-regulated learning

For self-regulated learning to be effective, students need to be able to accurately assess their own performance on a learning task and use this assessment for the selection of a new learning task. Evidence suggests, however, that students have difficulties with accurate self-assessment and task selection, which may explain the poor learning outcomes often found with self-regulated learning. In experiment 1, the hypothesis was investigated and confirmed that observing a human model engaging in self-assessment, task selection, or both could be effective for secondary education students’ (N = 80) acquisition of self-assessment and task-selection skills. Experiment 2 investigated and confirmed the hypothesis that secondary education students’ (N = 90) acquisition of self-assessment and task-selection skills, either through examples or through practice, would enhance the effectiveness of self-regulated learning. It can be concluded that self-assessment and task-selection skills indeed play an important role in self-regulated learning and that training these skills can significantly increase the amount of knowledge students can gain from self-regulated learning in which they choose their own learning tasks.     

More accurate than assumed: Learners’ metacognitive beliefs about the effectiveness of retrieval practice

We doubt the prevailing interpretation of lower Judgments of Learning (JOLs) for testing over rereading to reflect learners' favoritism of an ineffective activity. We argue that JOLs for testing are biased due to a negative feedback effect. In three preregistered experiments (N_final = 306), we eliminated the feedback effect by asking students to only imagine learning with the described activities (rereading/testing) after reading a text and by capturing offline-JOLs (off-JOLs = being decoupled from the current learning experience) as a function of an imaginary final test delay (5 min/1 week/2 weeks). In 5-min conditions, off-JOLs consistently reflected no differences between rereading and testing; in 1-week and 2-week conditions, two (of three) experiments demonstrated an advantage of testing over rereading. These results are consistent with actual learning outcomes in an experiment using the same text and activities (Rummer et al., 2017, Exp. 1). Learners’ metacognitive judgments resembled actual learning outcomes more accurately than suggested by previous research.     

Complexity of visual learning material moderates the effects of instructor's beat gestures and head nods in video lectures

Instructors often show rhythmic movements with speech in video lectures. However, the effects of these movements have not been tested in video lectures with an instructor and visual learning material. Results of Experiment 1 showed students in the beat gestures and head nods conditions showed significantly less attention to the visual learning material, more attention to the instructor, whereas in the beat gestures + head nods condition drew significantly more attention to both. In Experiment 2 we tested whether the complexity of the visual learning material moderated the effects of rhythmic movements on students’ learning. When the presentation was simple, students in the head nods + beat gestures condition showed more positive interaction, and significantly better recall and transfer. The results have implications for video designing: if an instructor and simple learning material are visible, she is encouraged to show beat gestures and head nods to mark important information.     

Visual processing of pictures and letters in alphabet books and the implications for letter learning

Alphabet books as studied in this research typically highlight one letter per page combined with a depiction of a word that begins with the letter (e.g., a bear illustrates B). This study tests whether children’s letter knowledge improves as a result of alphabet book sharing and how the visual processing of pictures and letters affects learning from repeated alphabet book readings. The study is designed as a randomized control trial in which participants were assigned to one of two experimental groups (N = 30) or a control group (N = 15). Half of the experimental group received version A in which the letters A–L were illustrated with anthropomorphic figures and the letters M–Z with objects. Half received version B in which the letters A–L were illustrated with objects and the letters M–Z with anthropomorphic figures. Mean age of the children was 57.6 months (SD = 3.6). While sharing the alphabet book we registered children’s eye movements in the first and fourth (last) reading session. Alphabet book reading stimulated letter knowledge although the make-up of the alphabet book moderated the effects. Relatively more visual attention to pictures of anthropomorphic figures interfered with learning letters from alphabet book sharing. Visual attention to letters also predicted letter knowledge and learning. Only a small part of each letter attracted children’s attention and the briefer their fixations on this distinctive area, the more letters they knew at the pretest and were able to learn from alphabet book sharing.     

What makes a matrix so effective? An empirical test of the relative benefits of signaling, extraction, and localization

What type of display helps students learn the most and why? This study investigated how displays differing in terms of signaling, extraction, and localization impact learning. In Experiment 1, 72 students were assigned randomly to one cell of a 4 × 2 design. Students studied a standard text, a text with key ideas extracted, an outline that localized ideas topically, and a matrix that localized ideas topically and categorically. One version of the displays signaled the displays’ organization and one version did not. The matrix display proved best for facilitating fact and relationship learning because of its ability to localize related information within topics and categories. Simply signaling or extracting text ideas was not helpful. Experiment 2 demonstrated that not all matrices are created equal because they can vary in terms of how information is localized. About 54 students were assigned randomly to one cell of a 2 × 2 design that varied localization of matrix topics and categories. Students studied matrices high or low in topical organization and high or low in categorical organization. Results confirmed that a high, natural ordering of matrix topics is necessary to highlight relationships and bolster relationship and fact learning.     

What is the effect of talking heads in educational videos with different types of narrated slides?

Video-based learning plays an increasingly important role and thus the optimal design of video-based learning materials attracts the attention of scientists and practitioners alike. In this context, producers of educational videos often include a talking head in their videos, although theory (e.g., cognitive theory of multimedia learning) also suggests potential disadvantages for this format. Since talking heads attract a lot of visual attention, further empirical research is necessary to investigate whether a talking head can hinder learning, especially presented next to graphic-based learning content. To address this research gap, we conducted two online experiments to investigate the effects of a talking head in educational videos with narrated slides (short slideshow lectures) on learning outcomes (i.e., factual knowledge acquisition) and participants’ subjective ratings of the learning material (e.g., perceived learning). In Experiment 1 (N = 96), we varied whether the instructor’s talking head was present or absent in the videos as a between-factor, and whether the visual content on the slides was graphic-based (pictures, diagrams, maps) or text-based (bullet points) as a within-factor (slide type). In Experiment 2 (N = 184), we additionally varied as a between-factor whether the contents appeared sequentially or statically all at once (presentation type). Our results showed that the talking head did not affect learning outcomes, regardless of slide type and presentation type of the videos suggesting that the inclusion of a talking head offers neither clear advantages nor disadvantages. Potential explanations for the findings and directions for future research are discussed.     

Benefits of emotional design in multimedia instruction

Emotional design of multimedia instruction involves making the essential elements in the lesson's graphics more appealing, such as by rendering them with human-like features and with distinct, appealing colors (Um, Plass, Hayward, & Homer, 2012). College students received an 8-slide multimedia lesson on how a virus causes a cold for 5 min (Experiment 1) or for as long as they wanted (Experiment 2). For the control group, the graphics consisted of simple black-and-white drawings in which the host cell was represented as a large circle, and the virus was represented as a small circle with small spikes on the outside and a rectangle on the inside. For the enhanced group, the graphics were redrawn to render the host cell as a red face with expressive eyes (registering surprise, fear, and sickness at various stages in the process), and the virus as a blue face with fierce eyes and with a green dot at the end of each of the blue tentacles surrounding the virus face. The enhanced group performed better than the control group on a subsequent learning test (d = 0.69 in Experiment 1, d = 0.65 in Experiment 2) and gave higher effort ratings in Experiment 1 (d = 0.65) but not in Experiment 2 (d = −0.10). The findings are generally consistent with the cognitive affective theory of learning with media, and point to the importance of incorporating motivation into cognitive theories of multimedia learning.     

Drawing pictures during learning from scientific text: testing the generative drawing effect and the prognostic drawing effect

Does using a learner-generated drawing strategy (i.e., drawing pictures during reading) foster students' engagement in generative learning during reading? In two experiments, 8th-grade students (Exp. 1: N = 48; Exp. 2: N = 164) read a scientific text explaining the biological process of influenza and then took two learning outcome tests. In Experiment 1, students who were asked to draw pictures during reading (learner-generated drawing group), scored higher than students who only read (control group) on a multiple-choice comprehension test (d = 0.85) and on a drawing test (d = 1.15). In Experiment 2, students in the learner-generated drawing group scored significantly higher than the control group on both a multiple-choice comprehension test (d = 0.52) and on a drawing test (d = 1.89), but students who received author-generated pictures in addition to drawing or author-generated pictures only did not. Additionally, the drawing-accuracy scores during reading correlated with comprehension test scores (r = .623, r = .470) and drawing scores (r = .620, r = .615) in each experiment, respectively. These results provide further evidence for the generative drawing effect and the prognostic drawing effect, thereby confirming the benefits of the learner-generated drawing strategy.