Suppose you are going to watch video lectures on how to prevent COVID-19. Is it better to view the lectures passively, or to view them and generate your own explanations of what you are learning? Research on video lectures suggests that learners show enhanced attention and better learning performance when watching a video lecture using the learning strategy of generating explanations (Fiorella et al., 2020; Pi et al., 2021).
Learning from video lectures has become a prevalent learning activity in both formal and informal educational settings. However, when passively viewing video lectures, many learners struggle to actively make sense of the learning material (i.e., selecting incoming information, organizing it, and integrating it with prior knowledge). One approach that helps learners to actively make sense of the learning material and optimize their learning from video lectures is to generate explanations as they learn (Fiorella & Mayer, 2016). Generating explanations is an active learning strategy that involves constructing statements that clarify the meaning of new knowledge by relating it to prior knowledge, promoting a deeper understanding of the learning material (Pi et al., 2021).
Given the rapid growth of online learning, it is becoming more and more common for learners to engage in online activity simultaneously with their fellow learners (Pi et al., 2020). For example, learners often simultaneously watch the same video lecture via massive open online courses (MOOCs). Studies on peer presence suggest that the mere presence of a peer or simply being aware of the co-learner’s presence can facilitate or inhibit learning performance (Belletier et al., 2019; Skuballa et al., 2019). However, relatively little research has been carried out on the social environment in which learning from video lectures takes place, and even less on the interaction of social environment and learning strategies on learning from video lectures.
The effects of learner-generated explanation on learning from video lectures
Generative learning theory emphasizes that learner-generated explanations facilitate learning (Chi, 2000). According to the theory, unlike other generative learning strategies (e.g., summarizing, retrieval activities), learner-generated explanation has the goal of making new material personally relevant (Chi, 2000; Fiorella & Mayer, 2016). Therefore, learners are able to fill in missing information, monitor their understanding, and regulate fusions of new information with prior knowledge when discrepancies or deficiencies are detected (Chi, 2000).
Learner-generated explanation has been studied for three decades now. One of the first studies in this area was conducted by Chi et al. (1989), who found that learners spontaneously generated explanations for themselves while reading a book. More recent research documented a positive relation between learner-generated explanation and learning outcomes (Lachner et al., 2021; Lawson & Mayer, 2021). However, much of the research on the effects of learner-generated explanation has focused on traditional classrooms, text, and animation as learning contexts (Lawson & Mayer, 2021; Roelle & Renkl, 2019), and there is a scarcity of research on the effects of learner-generated explanation in the context of video lectures.
Researchers have only recently begun to identify learning strategies that are more effective than passive viewing when learning from video lectures. To our knowledge, only two studies have examined the effects of learner-generated explanation on learning from video lectures (Fiorella et al., 2020; Pi et al., 2021). Both studies confirmed that learner-generated explanation when learning from video lectures enhanced learning performance. Furthermore, Pi et al. (2021) found that, compared with learners watching instructor-generated explanations, those who engaged in learner-generated explanations showed greater neural oscillations related to working memory and attention (increased theta and alpha band power).
The effects of co-learner presence on learning from video lectures
Research on learning from video lectures with co-learner presence, with both learners engaged in the same online learning activity, has been done in infant (Lytle et al., 2018), school-age (Tricoche et al., 2021), and adult samples (Pi et al., 2020). Evidence in social psychology shows that having a co-learner present who is engaged in the same learning task has considerable positive effects on learning outcomes across a range of learning contexts, including traditional classrooms (Tricoche et al., 2021) and online learning (Skuballa et al., 2019).
The social environment of co-learner presence also has positive effects on learning from video lectures (Huang et al., 2017; Lytle et al., 2018). For example, Li et al. (2014) reported that learners with co-learners who attended the weekly video lectures in one of two MOOCs resulted in higher satisfaction ratings, and both students showed better learning from the video lectures than those without co-learners. In another study, Lytle et al. (2018) reported that learners learned more from video lectures that taught foreign-language sounds when there was a co-learner presence. Compared to learners who learned alone, those who learned with a co-learner exhibited more speech-like vocalizations and were better able to discriminate between the foreign-language sounds.
For more than a century, social psychologists have tried to understand how co-learner presence can help or hinder learning (Zajonc, 1965). According to social presence theory, co-learner presence helps us learn because it primes the learner to adopt the perspectives of others, represent others’ minds, and infer whether and how others understand the incoming information (Jouravlev et al., 2018; Lachner et al., 2021). The learner adjusts explanations and self-regulates based on these mentalizing processes. As a consequence, these self-regulations might enhance learning performance. By contrast, the distraction-conflict theory postulates that co-learner presence triggers a conflict between attention to the ongoing task and attention to the co-learner (Belletier et al., 2019; Skuballa et al., 2019). This attention conflict might lead to cognitive overload, resulting in more mental effort and less learning.
The effects of praise on learning
Praise is a form of positive feedback. It commends the worth of a person, or expresses approval or admiration after the person shows a desired behavior or meets a goal (Al-Ghamdi, 2017). There is a growing body of literature on feedback that recognizes the importance of praise in learning (Al-Ghamdi, 2017; Zhao & Huang, 2020). The literature has shown that an image of an instructor or even a social robot expressing praise can improve learning performance, for example in math and reading comprehension (Davison et al., 2021). Researchers attribute the mechanisms of these benefits to learners’ increased motivation and positive attitude about new learning and further learning tasks (Maclellan, 2005; Takeue et al., 2013).
Drawing on feedback literature, one may speculate that the presence of a co-learner’s image expressing praise would be a beneficial social environment for learning from video lectures. However, the majority of studies on the effects of praise on learning have been conducted in other learning contexts, such as traditional classrooms (Al-Ghamdi, 2017) or a text-picture presentation on a screen (Zhao & Huang, 2020). It is unclear whether praise affects learning from video lectures, in which the brief presentation of information potentially could create more cognitive load.
The current study
Past research has largely examined the separate effects of learning strategies (Fiorella et al., 2020; Pi et al., 2021) and the social environment (Lytle et al., 2018) rather than their joint effects on learning from video lectures. Given the rapid growth of video lectures in formal and informal learning, it is important to establish effective learning strategies and understand ideal social environments in which these strategies can be used to foster deep learning. This information will also be critical in order to identify evidence-based ways to optimize performance in this learning context. The present study contributes to this goal by examining whether different learning strategies and social environments independently and interactively affected learning from a video lecture about infectious diseases.
Regarding learning strategies, we compared learner-generated explanation, in which the learner typed out an explanation after each subtopic was presented in a video lecture, to instructor-generated explanation presented as part of the video. Previous studies have indicated that learner-generated explanation while learning from video lectures was more effective than passive viewing (Fiorella et al., 2020; Pi et al., 2021), and in the current study learners passively viewed the instructor-generated explanation presented on a slide in the video. Regarding the social environment, we focused on three social contexts: the presence of a co-learner who gives praise for the learner’s explanation, the presence of a co-learner who does not give praise, and the absence of a co-learner. Co-learner presence in combination with the image of a teacher (Takeue et al., 2013) or robot (Davison et al., 2021) giving praise has been shown to facilitate learning. Finally, it should be noted that, in past research, video lectures were system-paced rather than self-paced. Because system-based pacing is uncommon in the use of video lectures in educational contexts, all video lectures in the current study were self-paced.
The mechanisms underlying the effects of these learning strategies and social environments remain unclear because past research has focused mainly on learning outcomes rather than learning processes. Therefore, in addition to learning performance, we also tested other outcome measures that might provide information about the learning process. These process-related outcomes were: the quality of explanations; self-reported mental effort; attention allocation on the video lecture, the co-learner, and the learner-generated explanation area based on eye tracking measures; and behavioral patterns in initiating different functions in the video area and the explanation area, assessed by log data. These additional outcome variables might provide clues as to why certain combinations of learning strategy and social environment promote learning more than others.
There were five conditions: (1) learner-generated explanation + co-learner + praise; (2) learner-generated explanation + co-learner; (3) learner-generated explanation + no co-learner; (4) instructor-generated explanation + co-learner; and (5) instructor-generated explanation + no co-learner. By comparing the first two conditions, we tested whether praise influenced learning from video lectures. We compared the second and fourth conditions to assess whether learner-generated explanation was more beneficial than passively viewing an instructor-generated explanation for learning in the presence of a co-learner. By comparing the third and fifth conditions, we examined whether learner-generated explanation was more beneficial for learning than passively viewing the instructor-generated explanation when learning from video lectures in the absence of a co-learner.
We utilized generative learning theory, the social presence hypothesis, the distraction-conflict theory, and previous research to make predictions about the effectiveness of learner-generated explanation, co-learner presence, and praise (Chi, 2000; Davison et al., 2021; Fiorella et al., 2020; Lytle et al., 2018; Roscoe & Chi, 2008). Our general prediction was that learning strategy (learner-generated or instructor-generated explanation) would interact with social context (co-learner, co-learner who gives praise, or no co-learner) to predict each outcome. The following hypotheses were proposed:
Hypothesis 1
The quality of explanations will be the highest in the presence of a co-learner who gives praise, followed by the presence of a co-learner who does not give praise, then finally by the absence of a co-learner.
Hypothesis 2
Learner-generated explanation will benefit learning performance; the benefit will be highest in the presence of a co-learner who gives praise, followed by the presence of a co-learner who does not give praise. Instructor-generated explanation will be less beneficial for learning performance, with lower scores in the presence of a co-learner, and the lowest scores in the absence of a co-learner.
Hypothesis 3
Learner-generated explanation will increase self-reported mental effort; the scores will be highest in the presence of a co-learner who gives praise, followed by the presence of a co-learner who does not give praise. Instructor-generated explanation will produce lower mental effort, with lower scores in the presence of a co-learner, and the lowest scores in the absence of a co-learner.
Hypothesis 4
Learner-generated explanation will enhance attention to the video lectures as assessed by eye movement data; the benefit will be highest in the presence of a co-learner who gives praise, followed by the presence of a co-learner who does not give praise. Instructor-generated explanation will be less beneficial, with lower attention in the presence of a co-learner, and the lowest attention in the absence of a co-learner.
Hypothesis 5
Learner-generated explanation will enhance attention to the co-learner area of the screen, as measured by eye movement data; the greatest enhancement will be in the presence of a co-learner who gives praise, followed by being in the presence of a co-learner who does not give praise, and then the instructor-generated explanation in the presence of a co-learner.
Hypothesis 6
The presence of a co-learner will enhance attention to the learner-generated explanation area of the screen, as measured by eye movement data; the greatest enhancement will be in the condition of the learner-generated explanation with the presence of a co-learner who gives praise, followed by the learner-generated explanation with the presence of a co-learner who does not give praise, and then the learner-generated explanation in the absence of a co-learner.
Hypothesis 7
Learner-generated explanation will enhance behaviors related to explanation adjustments and self-regulation, as measured by log data in the video area and the learner-generated explanations area; the highest number of behaviors related to explanation adjustments and self-regulation will be in the presence of a co-learner who gives praise, followed by in the presence of a co-learner who does not give praise, then in the absence of a co-learner. Instructor-generated explanation will be less helpful in enhancing learning in the presence of a co-learner, with fewer behaviors related to explanation adjustments and self-regulation, and the fewest behaviors related to explanation adjustments and self-regulation will be observed in the absence of a co-learner.