Qian
Huangfu
*a,
Hong
Li
a,
Sanshan
Tang
a,
Jianrong
Wang
b,
Qian
Liu
c and
Guojun
Chen
d
aSouthwest University, Chongqing, 430079, China. E-mail: chemqian16@swu.edu.cn
bGuanDu Experimental School, Yunnan Province 650000, China
cChengdu Shude Xiejin High School, Sichuan Province 610000, China
dChongqing No. 1 Secondary School, Chongqing, 430079, China
First published on 20th June 2022
Although the chemical literature contains many studies of multimedia-based learning and teacher enthusiasm, there is a paucity of research on whether and to what extent teacher enthusiasm in video lectures affects students’ learning, especially in chemistry. In this context, this mixed-method study used eye tracking and quantitative analysis to investigate how a teacher with different levels of enthusiasm influenced students’ learning in video lectures. Junior middle-school students were selected to engage in this eye-tracking research. We set up 35 such students as a group to view an experimental video with a low level of teacher enthusiasm, and 35 others as another group to view another experimental video with a high level of teacher enthusiasm. The essential tool for capturing the students’ visual attention was an EyeLink 1000 Plus eye tracker. The total dwell time, fixation counts, average fixation duration, and transition counts were recorded and analyzed, and the results showed that the teacher enthusiasm in the video lectures had an indirect positive effect on the students’ self-efficacy and learning performance and was negatively associated with cognitive load. In addition, students paid more attention to the teacher with the higher level of enthusiasm.
Teacher enthusiasm (TE) has usually been regarded as an effective way for teachers to deliver teaching information to learners (Shuell, 1996) and a crucial feature of high-quality effective learning (Kunter et al., 2013). TE can be effective for enhancing students’ learning interest (Kunter et al., 2008), motivation (Allen et al., 2006; Frenzel et al., 2009), and outcome (Brophy and Good, 1986). To date, many studies have investigated the definition, dimensions, and role of TE in classroom instruction (Collins, 1978; Brophy and Good, 1986), but few have explored whether TE can play a similar role in video teaching as it does in traditional classroom teaching. In previous studies, some of the learning content presented in video teaching was declarative knowledge (DK) while some was procedural knowledge (PK), and certain empirical studies have shown that the effects of teaching approach on DK and PK in video teaching are different (Hoffler and Leutner, 2007; Fabio and Antonietti, 2012).
The present study used DK of chemistry as the teaching content and collected data through eye movement to explore how TE influenced students’ cognitive load (CL), attention, learning performance (LP), and self-efficacy (SE). The aim was to reveal how different levels of TE affect learners in the process of multimedia learning and provide a theoretical basis for producers of instructional videos.
Fabio and Antonietti (2012) compared the effects of students learning PK and DK in video teaching and traditional teaching; they found that the effects of video teaching of both types of knowledge were better than those of traditional teaching, but the effects were not completely the same, with the learning effects of DK being better. In addition, some studies have explored whether seeing the teacher's face affects students’ academic performance in the process of learning PK and DK in VLs. Van Gog et al. (2014) used a PK “Frog Leap” demonstration as the learning material, while Van Wermeskerken and Van Gog (2017) used a DK molecular structure model to build knowledge-point video teaching, carried out a comparative study of video teaching with or without teachers’ images, and added teachers’ eye guidance to video demonstrations with teachers; the results of Van Gog et al. (2014) showed that learners with teachers were significantly better than learners without teachers, while those of Van Wermeskerken and Van Gog (2017) showed that there was no significant difference in students’ academic performance with or without a teacher.
Therefore, we speculate that the influence of TE on students’ learning in video teaching may produce different results for different types of knowledge. In the present study, DK was selected so that the teacher was continuously present in the teaching video.
The purpose of the present study was to determine whether and how TE in chemistry VLs is related to students’ learning, and so the TE in this study was conveyed through nonverbal behaviors such as rich facial expressions, frequent gestures, varied intonation, and even humor. Numerous studies have shown that there is a strong relationship between (i) a sender's verbal and nonverbal cues of instructional communications and (ii) the emotional responses of the receivers, with the receivers being more likely to be affected cognitively and emotionally by the sender's nonverbal messages (Mehrabian, 1972; Sueyoshi and Hardison, 2005; Batty, 2020). Liew et al. (2017) showed that pedagogical object enthusiasm (including smiling, nodding frequently, a high level of animated gestures, and speaking enthusiastically) significantly enhanced students’ positive emotions, and Arguedas and Daradoumis (2021) showed that affective pedagogical tutor feedback could reduce students’ negative emotions such as boredom and anger; these findings are consistent with emotional response theory (Mottet et al., 2006). Meanwhile, past research results show that the facilitating effects of agent enthusiasm on intrinsic motivation and cognitive results are fully mediated by a learner's positive emotions (Liew et al., 2017). A pedagogical agent who performs both gestures and facial expressions leads to better retention (Schneider et al., 2021), and the instructor's facial expressions influence students’ medium-term recall (Wang et al., 2019). The present study is similar to those of Keller et al. (2016) and Murray (2007), i.e., we focus on displayed TE instead of experienced TE.
TE benefits various student outcomes, such as LP (Keller et al., 2013; Burić, 2019), SE (Tuan et al., 2005), CL (Liew et al., 2017), visual attention (VA) (Chen and Wu, 2015), intrinsic motivation (Patrick et al., 2000; Burić, 2019), and interests (Kim and Schallert, 2014), among others. More concretely, TE is always regarded as an indispensable part of effective teaching and has excellent effects on students’ LP (Becker et al., 2014). The findings of the study by Rosenshine (1970) reflect the substantial empirical support for the positive impact of TE on students’ performance. Keller et al. (2016) used concept maps and paper-and-pencil tests to examine the relationship between TE and students’ LP and obtained the same result as that of Rosenshine (1970), i.e., that there is indeed a significant positive relationship between them. Later, Burić (2019) found that TE fully mediated the relationship between the emotional labor strategy of hiding feelings and students’ outcomes, SE, and intrinsic motivation. Those findings matched those of Patrick et al. (2000), who found that students had greater SE to learn the lecture material and had a higher level of vitality after watching enthusiastic VLs.
Collins (1978) constructed the first reliable criteria for measuring TE, her original intention being to develop TE to be evaluated by other professional observers. Subsequently, many scholars developed new instructions for measuring TE that were adapted from the work of Collins (1978). These instructions are used to assess displayed TE via students’ perceptions, and even though the validity of students’ perceptions is still disputable, many types of research have shown that students’ overall perceptions of instruction provide a reliable measure of classroom processes (Trautwein et al., 2006). Table 1 presents several tools for assessing TE, and for the present study, we chose the one by Murray (1983) to evaluate the level of TE in chemistry VLs, the main reason being that compared with the other tools, the items associated with this one are more suitable for evaluating displayed TE (Frenzel et al., 2009).
Instrument | Scale specifics | Item wording |
---|---|---|
Frenzel et al. (2009) | Four items, five-point rating scale from (1) strongly disagree to (5) strongly agree; α = 0.85 | • Our teacher teaches with enthusiasm |
• Our teacher is humorous during teaching | ||
• Our teacher tries to get students excited about the subject of mathematics | ||
• Our teacher seems to take pleasure in teaching | ||
Kunter et al. (2008) | Three items, four-point rating scale from (1) strongly disagree to (4) strongly agree | •Our teacher seems to really enjoy teaching |
•Our teacher is an enthusiastic teacher | ||
•Our teacher is enthusiastic about his/her subject | ||
Feldman (2007) | Not reported | • The instructor shows interest and enthusiasm in the subject |
• The instructor seems to enjoy teaching | ||
• The teacher communicates a genuine desire to teach students | ||
• The instructor never showed boredom for teaching this class | ||
• The instructor shows energy and excitement | ||
Patrick et al. (2000) | Four items, seven-point rating scale from (1) strongly disagree to (7) strongly agree; α = 0.93 | • The teacher just lights up the room when he/she teaches |
• The teacher is a bit dull | ||
• The teacher has a contagious energy about him/her | ||
• The teacher is full of dynamic energy when he/she teaches | ||
Marsh (IDEA; 1994) | Three items, five-point rating scale from (1) hardly ever to (5) almost always. Item description: “Describe the frequency of your instructor's teaching procedures.” | • Enthusiastic about the subject |
• Spoke with expressiveness | ||
• Dry and dull presentations | ||
Murray (1983) | 11 items, five-point rating scale rating frequency of occurrence from (1) almost never to (5) almost always | • Uses humor |
• Speaks expressively or emphatically | ||
• Shows facial expressions | ||
• Moves about while lecturing | ||
• Reads lecture verbatim from notes | ||
• Shows energy and excitement | ||
• Smiles or laughs | ||
• Gestures with hands and arms | ||
• Shows strong interest in subject | ||
• Avoids eye contact with students | ||
•Speaks softly | ||
Marsh and Ware (1982) | Three items; five-point rating scale | • Was enthusiastic about the subject |
• Has a good sense of humor | ||
• Made learning enjoyable | ||
Marsh (1982) | Four items, five-point rating scale from (1) very poor to (5) Very good. Item description: “As a description of this course/instructor, this statement is…” | • Instructor was enthusiastic about teaching the course |
• Instructor was dynamic and energetic in conducting the course | ||
• Instructor enhanced presentation with the use of humor | ||
• Instructor's style of presentation held your interest during class |
To acquire new knowledge, learners must process incoming information in their working memory, which is integrated with schemas stored in long-term memory. Working memory has a limited capacity, allowing only a limited number of items to be processed at one time (Cowan, 2001), and excessive information processing will cause CL and affect learning effect and efficiency. The core principle of CLT is that teaching design and learning environment should optimize learners’ limited working memory capacity to promote efficient learning. Based on CLT, different types of multimedia materials should reduce CL and make full use of working memory. Therefore, instructional designers should consider how to optimize learning and reduce unnecessary load as much as possible through effective design.
Compared with less-enthusiastic teachers, those with high levels of enthusiasm exhibit more-energetic tones and more facial expressions and gestures. Previous studies have shown that the nonverbal cues of pedagogical agents may distract the learner from salient learning (Clark and Choi, 2005; Woo, 2009; Frechette and Moreno, 2010). The results of Wilson et al. (2018) suggest that VLs involving teachers cost more to understand than do VLs containing only audio narrations by teachers. On the other hand, according to evidence supplied by various eye-tracking studies, the instructor also attracts a large amount of VA from the student (Kizilcec et al., 2014; Wang and Antonenko, 2017). Homer et al. (2008) examined how two versions of a computer-based multimedia presentation—i.e., a video version (including a video of a lecture synchronized with slides) and no video (including only the slides and an audio recording of a lecture)—affected LP; they found that having video as well as PowerPoint slides had a split-attention effect, subsequently increasing the CL. In summary, learners do divide their attention between the actor and the learning material (Louwerse et al., 2008). According to CLT, this may cause visual and cognitive interference to the learning process of learners, occupy the limited resources of working memory, and further distract learners from the processing of learning materials, resulting in additional CL (Clark and Choi, 2007).
However, some studies have shown that more verbal and nonverbal information in pedagogical agents does not increase students’ CL. Liew et al. (2017) compared the CL of students after watching teaching videos of highly enthusiastic or neutral pedagogical agents, and the results showed that there was no significant difference between the two groups of students. Wang et al. (2019) also showed that adding teachers’ facial expressions did not increase students’ CL.
According to the signaling principle of multimedia instruction design, the cues in multimedia learning resources can effectively guide learners to pay attention to important learning contents, to relieve the CL caused by attention delay or information screening in the learning process and improve their academic performance (Mayer and Moreno, 2003). Cues in video teaching include nonhuman ones (color, arrow, flicker, etc.) and human ones (teacher's expression, gaze, posture, etc.). Previous studies have found that social cues provided by humans guide learners’ attention far more than do these nonhuman cues (Johnson et al., 2015). Pi et al. (2017) found that teachers’ pointing gestures could direct learners’ VA to the corresponding teaching content to allocate more working memory resources to relevant learning activities by avoiding unnecessary visual searches on PowerPoint slides, and the results showed that learners’ LP was higher. Wang et al. (2019) investigated the effects of teachers’ facial expressions on students’ LP and CL; they found that the facial expressions of teachers are important cues for students’ memory, which could help teachers to emphasize key teaching points and eliminate uncertainty to help students understand well the information being conveyed. Van Gog et al. (2014) found that seeing the human model's face in the video fosters learning, a possible reason being that the speaker's gaze direction could guide learners’ attention at relevant moments and effectively transfer their attention from the model's face to the demonstration area. Therefore, a highly enthusiastic teacher with more facial expressions and gestures can reduce competition for working memory resources between relevant and irrelevant information and thus reduce unnecessary CL.
Based on the above discussion, a higher level of TE may either (i) distract students and thus increase CL or (ii) avoid adding extraneous CL by providing students with visual cues for learning. Therefore, it is very meaningful and necessary to determine the linkage between TE and CL. In the present study, we used a CL scale to measure students’ subjective CL and an eye-movement index to measure students’ objective CL and study the influence of higher TE on students’ CL in video teaching.
Eye-tracking technology is a crucial research tool in the field of education and is used mainly to track the process of people's eye movements that reflect their representation of mental processing in problem solving. Since Havanki and VandenPlas (2014) introduced eye-tracking technology into chemistry education research formally and regarded it as a useful and valuable tool for interpreting results more scientifically, more and more studies have chosen to use this visual survey tool in recent years, but there are still few studies related to the field of video teaching. With the help of eye-tracking technology, chemistry education researchers can explore learners' visual attention and cognitive processes while completing different tasks (Karch et al., 2019; Tang and Pienta, 2022), problem-solving methods and strategies (Tothova et al., 2021; VandenPlas et al., 2021), confirmation of setup demonstration principles (Nehring and Busch, 2020), understanding of chemical phenomena (Hansen et al., 2019), perceptual navigation patterns (Pande and Chandrasekharan, 2022) and some other aspects.
In addition, eye-tracking is an objective measure used in experimental psychology and in CL research through behavioral measures (Paas et al., 2003). The average fixation duration on relevant information, as well as the transitions between different areas of interest (AOIs)—providing information about learners’ cognitive engagement in information processing—should be considered as indicators of cognitive activity. The average fixation duration refers to the average duration of a fixation point in an AOI, with a longer fixation duration indicating a deeper and more effortful cognitive processing of visual information (Just and Carpenter, 1980; Holmqvist et al., 2011). For example, fixation duration increases as the complexity of texts or grammatical constructs increases (Rayner, 1998; Rayner et al., 2012). Fixation duration may thus be considered as an indicator of effort needed for visual information processing (Hodds et al., 2014; Krejtz et al., 2016; Strohmaier et al., 2020). According to the study of Paas and Vanmerrienboer (1994), the intensity of effort being expended by learners can be considered as the essence for obtaining a reliable estimate of CL. A widely used measure of CL is self-reported mental effort (Paas et al., 2003). In summary, average fixation duration can reflect the effort intensity of students in cognitive processing, and it is a very useful indicator in eye-tracking to confirm CL. Korbach et al. (2018) and Kruger and Doherty (2016) came to similar conclusions. Transition counts are considered to represent the learner's cognitive integration of multiple representations, which is the most critical step in multimedia learning for sense making (Mayer, 2014). High transition counts are seen as a strong cognitive investment in information integration (Schmidt-Weigand et al., 2010; Park et al., 2015), i.e., a higher transition count means that learners are exerting more mental effort to complete the information processing. Therefore, transition counts can be regarded as a indicator to confirm CL. Korbach et al. (2018) also proposed that the number of transitions could provide information on the causes of CL. Another eye-based measurement of CL is an increase in pupil size (Van Gog et al., 2009), but it is necessary to acknowledge that pupil size is too sensitive to other factors such as changes in light and brightness, and so researchers should be careful about interpreting this measure (Van Gog and Jarodzka, 2013). Therefore, in the present study, we chose the average fixation duration and transition counts as objective indicators to measure participants’ CL.
In the present study, eye-tracking technology was used to measure students’ VA to specific parts (learning-content materials or the teacher) of a certain screen and their cognitive load in the cognitive process. By analyzing the data, we discuss herein how TE influences students’ VA and CL in VLs.
H1. Compared to the students in the group that experienced a low level of TE, those in the group that experienced a high level of TE would show higher SE.
H2. Compared to the students in the group that experienced a low level of TE, those in the group that experienced a high level of TE would pay more attention to the teacher.
H3. Compared to the students in the group that experienced a low level of TE, those in the group that experienced a high level of TE would have a lower CL.
H4. Compared to the students in the group that experienced a low level of TE, those in the group that experienced a high level of TE would show higher LP.
Variable name | Description |
---|---|
Dwell time | Total time for which the eye gaze of a learner remained on a certain AOI |
Fixation count | Total number of fixations on a certain AOI |
Average fixation duration | Average duration of a fixation point |
Transition count | Number of direct-gaze switches between two AOIs |
The two experimental groups were taken to separate classrooms and followed the same procedures. Participants were asked to sit in a chair in front of an experimental monitor equipped with an eye-tracking device. Before starting to play the video instruction, the research assistant verbally explained the steps that the participants were expected to complete, and the same research assistant supervised the participants on every occasion. While viewing the uninterrupted video, the participants’ eye movements were recorded. After the video instruction, the participants completed three post questionnaires (SE questionnaire, CL questionnaire, and a test of post-learning comprehension) for 20 min at once. A flowchart of our method is shown in Fig. 1.
While viewing the VLs, participants’ eye movements were recorded. To address the second research hypothesis, we analyzed dwell time and fixation counts. For this purpose, AOIs were defined to separate the teacher and the learning content of the video in visually (Fig. 2). The AOIs were created manually, and in the video, we identified two areas: (A1) learning-content materials; (A2) the teacher. The dwell time measures how long the learner's eye gazed at an AOI (including fixations and quick scans) and was relativized by the duration of the video. Areas with higher dwell times can be interpreted as having been prioritized more. Such measures have recently been used in teacher-gaze studies (Stahnke and Blömeke, 2021). Similarly, we analyzed the numbers of fixations on the two AOIs as a second indicator of participant’ VA; a high fixation count indicates that the learner repeatedly allocated their attention to the corresponding area (Holmqvist et al., 2011). Concerning the third research hypothesis, we focused on average fixation duration and transition count, which reflect students’ mental activities. The data can provide indirect evidence for judging the size and change of students’ CL.
Before running the analyses described below (t-test), we checked the normality by a Shapiro–Wilk test (p > 0.05) and the homogeneity of variance by Levene's test (p > 0.05). When these conditions were not met, we ran nonparametric statistics on independent samples; the nonparametric test was the Mann–Whitney U test. In the present study, the measures for hypotheses H1 (SE) and H4 (LP) were analyzed using a t-test, those for hypothesis H2 (VA) were analyzed by nonparametric tests, and those for hypothesis H3 were analyzed using a t-test (CL scales) and nonparametric tests (eye-tracking outcomes).
In addition, the statistical hypotheses were tested at an alpha error rate of 5%. Effect sizes estimate the magnitude of effect or association between two or more variables (Snyder and Lawson, 1993); effect sizes are resistant to sample-size influence and thus provide a truer measure of the magnitude of effect between variables (Ferguson, 2009). In present study, to describe whether the effects had a relevant magnitude, the effect-size measures of Cohen's d and η2 were used to describe the strength of the phenomenon. Specifically, Cohen's d was used for the t-test, and Cohen (1992) defined the size of Cohen's d value as follows: 0.2, indicating that the experimental effect was small, 0.5, indicating that the experimental effect was medium, and 0.8, indicating that the experimental effect was large. η2 was used for the Mann–Whitney U test, with baseline definitions of small (0.01), medium (0.06), and large (0.14) effects provided by Cohen (1988).
Fig. 3 shows the outcome for the SE of all the students in the two groups. The t-test indicated that the SE scores in group B were significantly higher than those in group A (MgroupA = 22.77, MgroupB = 24.69, p = 0.002, Cohen's d = 1.13). Cohen's d was 1.13, indicating a large effect size. This result supports hypothesis H1 and is consistent with emotional response theory and the results of some other scholars (Rosenshine, 1970; Horan et al., 2012). The mere act of watching a video is primarily a passive learning experience, likely leading to a lack of student engagement and hindered learning (Mirriahi et al., 2021). For learning to be successful, students must hold adequate SE beliefs and keep investing sufficient effort in learning. There is evidence that students’ SE increases significantly when they receive positive emotional expressions and when motivating messages are present (Van der Meij, 2013). Consistent with our research results, positive emotions exhibited by enthusiastic teachers may enhance students’ SE, and students with more SE may transform the passive experience of watching videos into a more active experience and promote learning (Mirriahi et al., 2021).
H2. Compared to the students in the group that experienced a low level of TE, those in the group that experienced a high level of TE would pay more attention to the teacher.
Hypothesis H2 focuses on how the different levels of TE affect the distribution of the students’ attention on both the learning-content materials (area A1) and the teacher (area A2) during the VL learning. The eye-tracker data showed that the different VLs resulted in a significant difference in the students’ VA. We analyzed the students’ total dwell time, fixation counts, and heat maps, and the specific data for each AOI are given in Table 3.
Indicator | Group | N | Mdn | IQR | Mann–Whitney U test | ||
---|---|---|---|---|---|---|---|
U | p | η 2 | |||||
Dwell time [ms] | A | 35 | 573124.18 | 137812.85 | 498.00 | 0.179 | 0.056 |
B | 35 | 538932.09 | 232035.18 | ||||
Dwell time: A1 [ms] | A | 35 | 459754.10 | 80163.80 | 422.00 | 0.025 | 0.045 |
B | 35 | 428060.49 | 153114.89 | ||||
Dwell time: A2 [ms] | A | 35 | 78684.23 | 44959.21 | 316.00 | 0.000 | 0.131 |
B | 35 | 99893.79 | 59642.09 | ||||
Fixation counts: A1 | A | 35 | 1825.61 | 403.11 | 485.00 | 0.134 | 0.011 |
B | 35 | 1534.13 | 488.74 | ||||
Fixation counts: A2 | A | 35 | 300.69 | 213.02 | 474.00 | 0.104 | 0.046 |
B | 35 | 328.12 | 107.38 |
According to the Shapiro–Wilk tests, the eye-tracking data were not normally distributed. All measures were compared between groups A and B by the Mann–Whitney U test for independent samples. Basic descriptive statistics (median Mdn and interquartile range IQR) of the numerical variables were determined. The Mdn of the dwell time and the fixation counts in the teaching-content display area (area A1) for group A were higher than those for group B, and the situation was just the opposite for the teacher display area (area A2). The comparison in Table 3 shows that the dwell time in area A1 for group A was significantly higher than that for group B with small effect sizes (p = 0.025, η2 = 0.045), but the Mdn of the dwell time in area A2 for group A was significantly lower than that in group B with large effect sizes (p = 0.000, η2 = 0.131). That is to say, TE had a significant effect on students' dwell time in area A2, significantly increasing students' attention on the teacher, but the effect on the dwell time in area A1 is not obvious. This indicates that TE did not have a significant effect on students' attention to the learning-content materials (area A1), i.e., the “distraction effect” was not significant, although the high-TE teacher caused students' attention to shift more toward her. Also, heat maps are often used to show observation areas. This representation includes the so-called “hot spots” (marked in red) that show where students paid most attention. The heat maps in Fig. 4 show that the course content for group A received considerable VA.
The analysis of the eye-movement data showed that the teacher with the higher level of TE added to the students’ VA, which supports hypothesis H2. Analyzing the watching behavior of individual learners combined with detailed learning assessments offers rich insights into how different levels of TE might aid or hinder learning. This study found that characterizing learners’ watching behavior emphasized the great impact of TE on VA. When watching the VL with the higher level of TE, the students paid more attention to the teacher (area A2) and less attention to other areas, this being because the teacher exhibited richer expressions (smiles, surprise, questions), gestures, lively intonation, etc., which led to the students being drawn more to the teacher. As shown in Fig. 4, the teacher was asking the question “Is chemistry a subject that studies explosions?” The teacher with the lower level of TE asked this question in a neutral mood, while the teacher with the higher level of TE asked this question with a brisk tone and hand gestures describing explosions. The heat maps show that students in the low-TE group paid more attention to area A1, while students in the high-TE group paid more attention to area A2. This is consistent with other research that showed that teachers’ gestures (Rueckert et al., 2017) and facial expressions (Gullberg and Holmqvist, 2006) can draw people's attention to the speaker.
H3. Compared to the students in the group that experienced a low level of TE, those in the group that experienced a high level of TE would have a lower CL.
The CL scales were analyzed using a t-test, and the results showed that the CL of group A (M = 65.06) was higher than that of group B with medium effect sizes (M = 62.66, p = 0.015, Cohen's d = 0.598). This shows that the group that experienced the higher level of TE experienced the lower CL. Furthermore, we also used the transition counts and average fixation duration as indicators for measuring the students’ CL, and the CL scores of the two groups are shown in Fig. 5. The nonparametric Mann–Whitney U test indicated that there was a significant difference with medium effect sizes (p = 0.015, η2 = 0.076) in the transition counts between AOIs in the groups, and the average fixation duration for group B was much lower than that for group A with medium effect sizes (p = 0.023, η2 = 0.065).
In VLs, attentional cues such as gestures and facial expressions can guide learners’ VA in a timely manner to the information mentioned by the instructor, which can be a very effective way to reduce visual searches in a multimedia explanation. The teacher with the lower level of TE kept her body still, did not move her arms, did not have a directional gaze (i.e., did not shift her eyes from the camera to the learning content being discussed), and kept facing the students, whereas the teacher with the higher level of TE used pointing gestures while turning her upper body to the knowledge area; for example, the latter teacher pointed to the periodic table of elements in the demonstration area and turned her eyes to it when she said that it was discovered by the Russian chemist Mendeleev. Visual search efficiency describes the speed at which observers locate relevant visual information when they hear or read the corresponding word. The gestures and facial expressions of the high-TE teacher helped those participants shift their attention from her to the demonstration area in a timely and effective manner, and she guided students to pay attention to important informative words or concepts, thereby enabling the students to pay attention to relevant information more quickly and increasing their search efficiency (de Koning et al., 2009).
In addition, high fixation counts are associated with low search efficiency (Goldberg and Kotval, 1999), so the lower fixation counts of group B compared to group A (area A1) support the assertion that TE provides attention cues. Previous work has shown that attention cues can guide learners’ VA in a timely manner and reduce the competition for working memory resources between relevant information and irrelevant information in VLs, thereby reducing unnecessary CL (Van Gog et al., 2014). The present results are consistent with those of Pi et al. (2017) in that when the students were watching the VLs, their eyes and attention consciously followed the teacher's guiding behaviors to lead to the teaching content, thereby better allocating attention to the process of memorizing the teaching content and reducing CL. Therefore, TE can be effective in helping students to capture those parts of the learning materials that need to be processed deeply in rapid learning, thereby avoiding increased consumption of ineffective cognitive resources. Our research results support hypothesis H3, and the specific data for each AOI are given in Table 4.
Indicator | Group | N | Mdn | IQR | Mann–Whitney U test | ||
---|---|---|---|---|---|---|---|
U | p | η 2 | |||||
Average fixation duration [ms] | A | 35 | 281.18 | 82.79 | 419.50 | 0.023 | 0.065 |
B | 35 | 230.77 | 128.38 | ||||
Transition counts | A | 35 | 132.32 | 63.86 | 404.50 | 0.015 | 0.076 |
B | 35 | 107.23 | 65.98 |
H4. Compared to the students in the group that experienced a low level of TE, those in the group that experienced a high level of TE would show higher LP.
We used a t-test to identify how the different levels of TE affected the students’ post-test scores. As was shown earlier, there was no significant difference in the students’ prior knowledge between the two groups (MgroupA = 2.60, MgroupB = 2.57, p = 0.921). Also, all the students in this study had blank knowledge about this field and needed to learn more about the content of “Changes and Properties of Matter” in the VLs. The LPs of the two groups are shown in Fig. 6, and clearly the LP scores of group B (M = 9.08, SD = 1.221) were higher than those group A (M = 6.46, SD = 0.817). Moreover, there was a significant difference between the two groups with large effect sizes (p = 0.000, Cohen's d = 1.817). The results show that the students in group B outperformed those in group A, which supports hypothesis H4. This is consistent with previous work that shows that TE has a positive impact on teaching effectiveness and that student performance is positively correlated with TE (Streeter, 1986; Kunter, 2013; Mahler et al., 2018). Based on hypothesis H4, we carefully conclude that the enthusiasm of chemistry teachers in VLs may be very important for students’ LP and should therefore be cultivated in pre-service and in-service teacher education. For example, teachers could develop awareness and skills to remain enthusiastic when facing the camera by participating in professional development courses or self-study.
According to the “split-attention effect” (Ayres and Sweller, 2014), more nonverbal information delivered by teachers with high TE will distract students’ attention and lead to them paying more attention to their teachers. We also showed the existence of this effect in hypothesis H2. However, the high-TE teacher did not reflect the harmful effect of split attention on learning. In the present study, the LP of the students in the high-TE group was significantly higher than that of those in the low-TE group. This result is similar to previous studies. For example, Wang and Antonenko (2017) studied the effects of instructor presence or absence on students’ VA distribution and learning effect in video teaching, and the results showed that the presence of the teacher divided the students’ attention to the teacher (26% of the students’ dwell time), and the students’ learning of simple questions was significantly better. Van Gog et al. (2014) conducted an eye-tracking study in which students learned problem-solving tasks from video modeling examples in which the model's face was either visible or invisible; their results showed that the percentage of fixation duration on the demonstration area decreased by 22.91% when the model's face was visible. This suggests that the model's face distracted learners from the presentation area, but the students’ performance improved significantly after they had seen the video modeling example twice. As discussed in the theoretical framework of the present study, the nonverbal information of the high-TE teacher played an important signaling function, directing the learners’ attention to the most relevant and important aspects of the teaching content, enabling them to participate in cognitive processing of the learning materials, and promoting the integration of visual demonstration and verbal explanation. It is reasonable to speculate that high TE has a positive effect on students’ LP because the signaling advantage brought by teachers’ nonverbal cues is greater than the negative effect of distraction.
In addition, facial expressions (eye, eyebrow, and mouth movements) are particularly important for online learning (Johnson et al., 2000; Atkinson, 2002). The face is the primary source for expressing emotions nonverbally (Mehrabian, 1971). In the present VLs, the high-TE teacher used different types of facial expressions: happy, surprised, confused, etc. The emotion expressed by the instructor's facial expressions in VLs is critical to shortening this distance between students and online teachers, i.e., it could overcome students’ feelings of isolation and improve their arousal level and participation in learning activities (Borup et al., 2014). Baylor and Kim (2009) noted that persuasion can be facilitated though the physical social cues of an agent's eyes, face, hands, and body; because the agent's facial expressions contain emotions, this enhances the persuasiveness of the lesson to the students, thus benefiting their learning. Mayer and DaPra (2012) reported a similar result that virtual instructors’ appropriate use of facial expressions could promote students’ deep cognitive processing and encourage their learning. Therefore, the nonverbal information conveyed by the high-TE teacher prompted the processing of verbal information.
However, nonverbal cues should be designed carefully to ensure that they do not increase CL. From a practical viewpoint, this study showed that TE can be formulated via the operation of teachers’ expressions and body gestures. Note that cues via immediate controls of teachers’ voice tones, facial expressions, and animation gestures open a myriad of exciting potential benefits regarding the development of learner interaction in VLs. In multimedia learning environments, TE helps students to maintain a higher level of SE. Moreover, TE maintains a higher attention level from student to teacher and improves students’ scores. This study also showed that CL was positively influenced by TE. Therefore, whether in a conventional classroom or a multimedia learning environment, teachers should pay attention to the proactive expression of their enthusiasm.
When interpreting our findings, some limitations should be kept in mind. First, the limited sample size did not involve a nationally representative sample of students, which may restrain us from generalizing the results of our study to all students in China; thus, future studies should expand the sample size by including students from different subjects, grade levels, and regions. Second, because some students were curious about the eye tracker, they might have paid more attention to it than to the VL, and this restricted the scientific nature of the research results to some extent. Third, this study involved only chemistry DK, and PK could be studied in the future. Finally, an important direction of future study would be to understand well the effects and mechanism of teachers’ other nonverbal cues regarding students’ learning.
This journal is © The Royal Society of Chemistry 2022 |