You are on page 1of 8

Computers & Education 144 (2020) 103713

Contents lists available at ScienceDirect

Computers & Education


journal homepage: http://www.elsevier.com/locate/compedu

Instructor presence in video lectures: Eye gaze matters, but not


body orientation
Zhongling Pi a, Ke Xu a, Caixia Liu b, Jiumin Yang a, c, *
a
School of Educational Information Technology, Central China Normal University, Wuhan 430079, China
b
Collaborative Innovation Center for Information Technology and Balanced Development of K-12 Education, Central China Normal University,
Wuhan 430079, China
c
National Engineering Research Center for E-learning, Central China Normal University, Wuhan 430079, China

A R T I C L E I N F O A B S T R A C T

Keywords: The instructor’s on-screen presence, rather than just voice, has rapidly become a popular feature
Teaching/learning strategies in video lectures. Eye gaze and body orientation are core indicators of an instructor’s attentional
Multimedia/hypermedia systems focus in classroom settings, but it is not well known how these factors influence learners’
Pedagogical issues
attention allocation and learning performance in video lectures with the on-screen presence of
Distance education and
instructors presenting slides. This study tested the effects of the instructor’s eye gaze and body
telelearning
orientation on attention and learning from a video lecture in a sample of 174 undergraduates.
Specifically, the instructor’s presence was manipulated in terms of eye gaze (direct, guided, or
averted) and body orientation (frontal or lateral). Eye tracking data revealed that regardless of an
instructor’s body orientation, learners who viewed the video lectures with the instructor’s guided
gaze paid greater attention to the slides, and those who viewed the video lectures with the in­
structor’s direct gaze paid greater attention to her face; paper-and-pencil assessments showed that
learners who viewed the video lectures with guided gaze showed better retention and transfer.
These results held regardless of body orientation, suggesting that an instructor’s eye gaze has a
stronger influence than body orientation on attention and learning from video lectures. The
findings suggest that an instructor should not look directly at the camera continuously throughout
the lecture, and should instead use guided gaze to draw learners’ attention to the learning
materials.

1. Introduction

The on-screen presence of instructors in video lectures has become a topic of interest to researchers with the growth of online
learning (Crook & Schofield, 2017). In this type of video lecture, learners can not only listen to the instructor’s speech but also see her
nonverbal cues, such as her eye gaze and body orientation, which indicate her attentional focus across the visual field (Beege, Nebel,
Schneider, & Rey, 2019; Stull, Fiorella, & Mayer, 2018). According to social agency theory (Mayer, 2014), as typical social cues, an
instructor’s eye gaze and body orientation promote a sense of interaction between the instructor and learners. That is, learners believe
that the instructor is personally teaching them; this sense will socially motivate them to engage in deeper cognitive processing and
hence they show better learning performance.

* Corresponding author. No. 152 Luoyu Road, Hongshan District, Central China Normal University, Wuhan 430079, Hubei Province, China.
E-mail address: yjm@mail.ccnu.edu.cn (J. Yang).

https://doi.org/10.1016/j.compedu.2019.103713
Received 20 March 2019; Received in revised form 24 September 2019; Accepted 29 September 2019
Available online 30 September 2019
0360-1315/© 2019 Elsevier Ltd. All rights reserved.
Z. Pi et al. Computers & Education 144 (2020) 103713

The instructor’s eye gaze and body orientation in video lectures have been evidenced to have a variety of consequences for learners’
attention allocation and learning performance (Beege, Schneider, Nebel, & Rey, 2017; Leong et al., 2017). Thus there is an open
question: When multiple social cues are presented by the instructor in video lectures, do body orientation and eye gaze equally in­
fluence learning, or is one factor more influential than the other? This question is especially relevant for video lectures with slides.

1.1. The role of eye gaze

In video lectures, there are three types of instructor’s eye gaze: direct gaze, guided gaze, and averted gaze (Fiorella, Stull, Kuhl­
mann, & Mayer, 2018; Holler et al., 2014). The instructor’s direct gaze in video lectures refers to gazing at the camera as if making eye
contact with learners. The instructor’s guided gaze refers to the instructor gazing at learning materials (e.g., slides, whiteboard) in the
video lecture. The instructor’s averted gaze refers to the instructor being recorded from an angle and not gazing at learning materials in
the video lecture, as if she is teaching other students in the class. Recent research has suggested that different types of instructor eye
gaze play different roles in learning from video lectures (Stull, Logan, & Mayer, 2018).

1.1.1. An instructor’s direct gaze


Previous studies found consistent effects of the instructor’s direct gaze on students’ attention in video lectures (Pi & Hong, 2016;
van Wermeskerken, Ravensbergen, & van Gog, 2018; van Wermeskerken & van Gog, 2017). That is, the instructor’s direct gaze draws
learners’ attention to herself, especially her face. Eye tracking research has shown that learners looked at the instructor’s face over 30%
of the time (indexed by percentage dwell time) during learning video lectures in which the instructor directly gazed at the camera, and
there was no decline in attention to the instructor over time (Pi & Hong, 2016; van Wermeskerken et al., 2018).
However, previous studies have found inconsistent effects of the instructor’s direct gaze on students’ learning performance in video
lectures. On one hand, in line with social agency theory (Mayer, 2014), video lectures with an instructor’s direct gaze can potentially
benefit learners’ social partnership with the instructor, and as a consequence, facilitate learners’ attention and learning performance.
Recent studies indeed provided evidence that the instructor’s direct eye contact, relative to averted gaze, had positive learning effects
among learners of a wide range of ages (Beege et al., 2019; Fiorella et al., 2018; Leong et al., 2017). The results of these studies suggest
that when an instructor provides direct gaze, learners show better learning performance.
On the other hand, the instructor’s direct gaze could be a potential cause of split attention, as direct gaze attracts learners’ attention
to the instructor while learners are also attending to the learning materials (Ayres & Sweller, 2014; Wilson, Martinez, Mills, D’Melllo,
& Smilek, 2018). Cognitive Load Theory (CLT) suggests that learners’ cognitive resources are limited, so that learners can only pay
attention to a part of incoming information at any given time (Sweller, Ayres, & Kalyuga, 2011). Therefore, in video lectures, when
learners are looking at the instructor’s direct gaze, they may not show sufficient attention to the learning materials in the slides, and
vice versa. However, contrary to this assumption, previous studies found that although direct gaze attracts learners’ attention to the
instructor (and away from the learning materials), it does not appear to hinder learners’ retention and transfer relative to guided gaze
(Ouwehand, van Gog, & Paas, 2015; Pi et al., 2019).

1.1.2. An instructor’s guided gaze


Previous studies have consistently shown that like other attentional cues (e.g., gestures and pointing), guided gaze draws learners’
attention to learning materials in the visual screen of video lectures (Pi et al., 2019; Pouw, van Gog, Zwaan, Agostinho, & Paas, 2018).
According to social cognitive theory, learners even at a young age automatically follow the instructor’s gaze in order to look at what
the instructor is looking at (Clifford & Palmer, 2018; Hoehl, Wiese, & Striano, 2008). Surprisingly, guided gaze has not been shown to
be more effective than direct gaze in improving performance on learning tests (Ouwehand et al., 2015; van Wermeskerken & van Gog,
2017). In fact, Fiorella et al. (2018) found that learners viewing a video lecture with direct gaze by an instructor who was writing on a
transparent board showed better retention and transfer than those viewing a video lecture with guided gaze by an instructor writing on
a conventional whiteboard. However, the comparison of guided and direct gaze might have been confounded by the use of different
types of writing boards in the two conditions.

1.2. The role of body orientation

Instructors’ body orientations can be categorized as frontal and lateral (Beege et al., 2017; Nagels, Kircher, Steines, & Straube,
2015). Frontal orientation in video lectures refers to the front of the instructor’s body being presented to the camera. Lateral orien­
tation refers to the side of the instructor’s body being presented to the left or right front of the camera in the video lecture. Psycho­
logical studies have provided evidence that learners can use the instructor’s body orientation to understand her focus of attention
(Cooney, Brady, & Ryan, 2017; Paulus, Murillo, & Sodian, 2016). In this respect eye gaze and body orientation are similar. An
instructor uses frontal body orientation to convey that she is talking directly to the learners (similar to direct gaze); the instructor uses
lateral body orientation to convey that she is gazing at the slides (similar to guided gaze) or other places (similar to averted gaze).
Recent studies showed that the instructor’s frontal body orientation in a video lecture enhanced learning performance (Beege et al.,
2017, 2019). However, these studies did not differentiate the effect of an instructor’s body orientation from that of eye gaze, so that no
conclusion can be drawn about whether the benefits can be attributed to direct gaze, frontal body orientation, or both.
Then the question arises: If an instructor uses both body orientation and eye gaze to express her attentional focus in a video lecture
(as in Beege et al., 2017, 2019), is learning influenced equally by body orientation and eye gaze, or has one factor a stronger effect than
the other? Some studies on social attention have asked a similar question. These studies suggested that when the instructor was visible

2
Z. Pi et al. Computers & Education 144 (2020) 103713

(showing both eye gaze and body orientation) in the video lecture, learners’ response to eye gaze overrode their response to body
orientation in simple attention guiding tasks (Cooney et al., 2017; Lawson & Calder, 2016; Perrett, Hietanen, Oram, Benson, & Rolls,
1992). Another study showed that in video lectures learners’ attention to the instructor mainly focused on the face rather than the body
(van Wermeskerken et al., 2018).

1.3. The present study

This study tested the effects of the instructor’s eye gaze (direct, guided, or averted) and body orientation (frontal or lateral) on
learners’ attention allocation (i.e., attention to instructor’s face, attention to instructor’s body, and attention to slides) and learning
performance (i.e., retention and transfer). Participants viewed a video lecture showing an instructor and slides presented to the right
side of her, in one of six conditions: (1) direct gaze þ frontal body (DG þ FB), (2) guided gaze þ frontal body (GG þ FB), (3) averted
gaze þ frontal body (AG þ FB), (4) direct gaze þ lateral body (DG þ LB), (5) guided gaze þ lateral body (GG þ LB), or (6) averted
gaze þ lateral body (AG þ LB).
Based on theoretical considerations and the results of previous studies, we assumed that an instructor’s eye gaze would override
body orientation in influencing attention allocation and learning from video lectures. Specifically, we hypothesized that learners
viewing video lectures with an instructor’s direct gaze would show a greater level of attention to the instructor’s face; learners viewing
video lectures with an instructor’s guided gaze would show a greater level of attention to the slides; and learners viewing video lectures
with an instructor’s averted gaze would show worse learning performance (retention and transfer).

2. Method

2.1. Participants and design

Undergraduate students from a Chinese university participated in the study (N ¼ 174; 102 females; Mage ¼ 19.83, SDage ¼ 1.80, age
range: 17–27). All participants had normal or corrected-to-normal vision and hearing. All of them gave written informed consent. The
participants received a small present (i.e., facial tissue, candy) at the end of the study to thank them for their time and effort.
Participants were randomly assigned to one of six conditions in a three-by-two factorial between-subjects design (eye gaze: direct,
guided, or averted; body orientation: frontal or lateral). There were 29 participants in each condition.

2.2. Apparatus and eye movement data analysis

Eye movements were registered by a Tobii T120 eye tracker with a sampling rate of 120 Hz. Participants sat at a distance of 60 cm
away from the screen. There were three areas of interest (AoI) in the video lecture: face area, body area, and slides area. To measure
how the instructor’s eye gaze and body orientation affected learners’ attention, the percentage dwell time was computed as the total
dwell time on a particular AoI divided by the total fixation duration of the video lecture.

Fig. 1. Screen shots of the six conditions.

3
Z. Pi et al. Computers & Education 144 (2020) 103713

2.3. Materials

2.3.1. Video lectures


The topic of the video lecture was “wind.” In each condition, the same female instructor gave the same information about wind
(described wind comprehensively and provided the definition, causes, categories, and characteristics) and used the same slides. She
spoke about 230 Chinese words per minute and her oral explanation did not explicitly signal learners to pay attention to particular
information on the slides. We used an informal interview before the study to ensure that none of the participants had been enrolled in a
class with the instructor, to avoid any expectations about the lecture. Each lecture lasted 7:20 min.
There were six conditions in a 3 � 2 factorial design (see Fig. 1). (1) DG þ FB: the instructor gazed into the camera and stood in a
frontal body position. (2) GG þ FB: the instructor gazed at the slides and stood in a frontal body position. (3) AG þ FB: the instructor
gazed at her left side at an angle of 40� from the camera and stood in a frontal body position, as if she were talking to other learners
(Beege et al., 2017). (4) DG þ LB: the instructor gazed into the camera and stood in a lateral body position. (5) GG þ LB: the instructor
gazed at the slides and stood in a lateral body position. (6) AG þ LB: the instructor gazed at her left side at an angle of 40� from the
camera and stood in a lateral body position.

2.3.2. Measurements
The prior knowledge test was created by the instructor who appeared in the lecture, to evaluate participants’ general knowledge of
wind. The test included four multiple-choice items (total 8 points), two fill-in-the-blank items (total 4 points), and two open-answer
items (total 10 points); the total score was the sum of all items, with a maximum possible score of 22. Answers to the two open-answer
items were rated separately by two trained raters, with high inter-rater reliability (respectively, α ¼ 0.98 and 0.99). Following are
examples of each type of question. Multiple-choice: “Wind is the horizontal movement of air, from: A. high pressure to low pressure; B.
low pressure to high pressure; C. mutual convection; D. a stationary front.” Fill-in-the-blank: “The two basic elements of measuring
wind are ______ and ______.”Open-answer: “So far, why isn’t it particularly effective to try to predict tornadoes?” The prior knowledge test
showed moderate internal consistency (α ¼ 0.63).
The first learning test was a retention test created by the instructor in the lecture. It included three multiple-choice items (two
scoring methods, total 8 points), eight fill-in-the-blank items (ten blanks, total 10 points), and two open-answer items (each scored
differently, total 11 points); the total score was the sum of all items, with a maximum possible score of 29. On the two open-answer
questions, the two raters showed high inter-rater agreement (respectively, α ¼ 0.97 and 0.99). Following are examples of each type of
question. Multiple-choice: “What are the main causes of wind: A. differences in elevation; B. differences in air humidity; C. differences
in atmospheric temperature; D. differences in air pressure.” Fill-in-the-blank: “Wind is a physical quantity that represents the motion of
air flow. It has both magnitude and direction. Therefore, wind is a.” Open-answer: “Please draw a diagram of wind formation.” The
retention test showed moderate internal consistency (α ¼ 0.53).
The second learning test was a transfer test created by the instructor in the lecture. The test required learners to use their knowledge
of wind (the definition, causes, categories, and characteristics) to analyze wind direction and air pressure in new situations. For
example, in the video lecture, participants learned the principle of wind blowing from a high pressure area to a low pressure area, and
then they had to judge the distribution of sea and land pressure in an area east of Zhejiang province in China according to the direction
of the wind. The test included seven multiple-choice items (total 14 points), a line matching item (5 points total), and two open-answer
items (each scored differently, total 13 points). The total score was the sum of all items, with a maximum possible score of 32. The two
open-answer items were rated separately by the two trained raters, with high inter-rater agreement (respectively, α ¼ 0.99 and 0.98).
Following are examples of each type of question. Multiple-choice: “East of Zhejiang, in summer, the wind blows from the sea to the
land. In winter, the wind blows from the land to the sea. It can be inferred that: A. Winter and summer winds will change constantly in
Zhejiang; B. In summer, the air pressure on the sea is higher than that on land; C. Wind direction is unchanged in Zhejiang; D. In
summer, the air pressure on land is higher than that on the sea.”Matching items required participants to match wind scale, the type of
wind and the wind’s characteristics. There were 5 wind scales, 5 types of wind, and 5 sentences describing the characteristics of the
wind, constructing 5 pairs. Short-answer item:“Please explain why a ‘waterspout’ appears.” The transfer test showed moderate internal
consistency (α ¼ 0.53).

Table 1
Means of all dependent variables, with standard deviations in parentheses, for each condition.
Dependent variable DG þ FB n ¼ 29 GG þ FB n ¼ 29 AG þ FB n ¼ 29 DG þ LB n ¼ 29 GG þ LB n ¼ 29 AG þ LB n ¼ 29

Prior knowledge 6.84 (3.35) 8.03 (3.85) 6.65 (2.36) 7.93 (4.51) 9.37 (4.14) 7.86 (2.07)
Dwell time on face 6.43 (6.86) 3.00 (2.42) 5.35 (5.57) 6.99 (8.10) 3.66 (3.56) 2.93 (3.12)
Dwell time on body 0.26 (0.39) 0.19 (0.19) 0.26 (0.39) 0.32 (0.51) 0.28 (0.83) 0.20 (0.29)
Dwell time on slides 60.84 (14.79) 65.26 (10.27) 61.32 (8.28) 57.67 (13.81) 68.49 (10.90) 64.58 (9.25)
Retention 12.58 (2.61) 13.90 (4.10) 11.81 (2.87) 12.38 (3.83) 15.15 (3.57) 10.31 (3.86)
Transfer 16.74 (3.45) 20.33 (3.77) 15.78 (3.48) 17.17 (4.60) 19.72 (4.47) 15.40 (3.30)

Note: DG þ FB, GG þ FB, AG þ FB, DG þ LB, GG þ LB, and AG þ LB respectively represents the direct gaze þ frontal body condition, guided
gaze þ frontal body condition, averted gaze þ frontal body condition, direct gaze þ lateral body condition, guided gaze þ lateral body condition, and
averted gaze þ lateral body condition.

4
Z. Pi et al. Computers & Education 144 (2020) 103713

2.4. Procedure

The study was conducted in individual sessions of 30–40 min. Firstly, participants took the prior knowledge test. Then they entered
an eye-tracking room. Before watching the lecture, the eye-tracking system was calibrated using 5-point calibration plus a 5-point
validation procedure. Participants viewed the lecture by themselves without a break, and then they took the learning tests (reten­
tion and transfer).

3. Results

To test differences across the six experimental groups, five 3 � 2 analyses of covariance (ANCOVAs) were conducted. For all an­
alyses, eye gaze (direct, guided, or averted) and body orientation (frontal or lateral) were used as independent variables and the score
on the prior knowledge test was used as the covariate. The dependent measures were the three dwell times (on face, body, and slides),
retention score and transfer score. Descriptive statistics (means and standard deviations) and the results of the ANCOVAs are
respectively presented in Table 1 and Table 2.

3.1. Attention

3.1.1. Dwell time on face


As predicted, there was a significant main effect of eye gaze on dwell time on face. A post hoc test (LSD) found that learners who
watched the video lectures with the instructor’s direct gaze paid greater attention to the instructor’s face than those who watched the
video lectures with guided gaze and averted gaze (respectively, MD ¼ 3.29, p ¼ .001; MD ¼ 2.58, p ¼ .010). No significant main effect
of body orientation and no significant interaction were observed. The results suggested that the instructor’s direct gaze attracted
learners’ greater attention to her face.

3.1.2. Dwell time on body


No significant main effect of eye gaze, no significant main effect of body orientation and no significant interaction were found in
relation to dwell time on body. The results suggested that the instructor’s eye gaze and body orientation did not affect learners’
attention towards her body.

3.1.3. Dwell time on slides


Only the main effect of eye gaze was significant. The post hoc test found that learners who watched the video lectures with the
instructor’s guided gaze paid greater attention to the slides than those who watched the video lectures with direct gaze and averted
gaze (respectively, MD ¼ 7.76, p < .001; MD ¼ 4.08, p ¼ .061). No significant main effect of body orientation and no significant
interaction were present. The above results suggest that the instructor’s guided gaze drew greater attention to the slides.

3.2. Learning performance

3.2.1. Retention
ANCOVA showed that the main effect of eye gaze was significant. A post hoc test found that learners who watched the video lecture
with the instructor’s guided gaze showed the highest retention, followed by those who watched the video lecture with the instructor’s
direct gaze, and finally by those who watched the video lecture with the instructor’s averted gaze (Fig. 2). No significant main effect of
body orientation and no significant interaction were observed. The results suggest that the instructor’s guided gaze promoted learners’
retention and her averted gaze hindered their retention.

3.2.2. Transfer
There was a main effect of eye gaze on transfer. Post hoc tests found that learners who watched the video lecture with the in­
structor’s guided gaze showed the best transfer, followed by those who watched the video lecture with the instructor’s direct gaze, and

Table 2
Results of ANCOVAs.
Dependent variable Main effect Interaction effect

Eye gaze Body orientation

Dwell time on face F (2, 168) ¼ 6.30, p ¼ .002, η2p ¼ 0.07 F (1, 168) ¼ 0.15, p ¼ .702, η2p ¼ 0.001 F (1, 168) ¼ 1.55, p ¼ .215, η2p ¼ 0.02
Dwell time on body F (2, 168) ¼ 0.30, p ¼ .740, η2p ¼ 0.004 F (1, 168) ¼ 0.60, p ¼ .809, η2p< 0.001 F (2, 168) ¼ 0.36, p ¼ .696, η2p ¼ 0.004
Dwell time on slides F (2, 168) ¼ 6.46, p ¼ .002, η2p ¼ 0.07 F (1, 168) ¼ 0.49, p ¼ .485, η2p ¼ 0.003 F (2, 168) ¼ 1.52, p ¼ .222, η2p ¼ 0.02
Retention F (2, 168) ¼ 12.26, p < .001, η2p ¼ 0.13 F (1, 168) ¼ 0.37, p ¼ .542, η2p ¼ 0.002 F (2, 168) ¼ 2.22, p ¼ .112, η2p ¼ 0.03
Transfer F (2, 168) ¼ 16.99, p < .001, η2p ¼ 0.17 F (1, 168) ¼ 0.62, p ¼ .434, η2p ¼ 0.004 F (2, 168) ¼ 0.33, p ¼ .721, η2p ¼ 0.004

Note: We used a significance level of 0.05 for all analyses. Partial eta square (η2p) was reported as a measure of effect size for ANCOVAs, with η2 ¼ 0.01,
η2 ¼ 0.06, and η2 ¼ 0.14 corresponding to small, medium, and large effects, respectively (Cohen, 1988). Score on the Prior Knowledge Test was the
covariate.

5
Z. Pi et al. Computers & Education 144 (2020) 103713

Fig. 2. There was a main effect of eye gaze on retention and transfer. The figure depicts post hoc comparisons among the three types of eye gaze.

finally those who watched the video lecture with the instructor’s averted gaze (Fig. 2). No significant main effect of body orientation
and no significant interaction were observed. These results suggest that the instructor’s guided gaze enhanced learners’ transfer and
her averted gaze hindered their transfer.

3.3. Summary

Taken together, the ANCOVAs showed a pattern of several main effects of eye gaze on learners’ attention and behavior, but no main
effects for body orientation and no interactions. The results support our hypothesis that there would be a stronger effect of an in­
structor’s eye gaze compared to body orientation on learners’ attention, retention and transfer.

4. Discussion

The current study focused on the effects of an instructor’s eye gaze and body orientation on learners’ attention and learning
performance in video lectures with slides. We found that only eye gaze had a significant positive effect on learners’ attention allocation
and learning performance. This is the first study to provide evidence that an instructor’s eye gaze has more influence than her body
orientation on students’ attention and learning performance.
As hypothesized, the instructor’s eye gaze, relative to her body orientation, played the dominant role in guiding learners’ attention
allocation. When the instructor used direct gaze, learners paid greater attention to her face regardless of her body orientation; once the
instructor used guided gaze, learners paid greater attention to the slides regardless of her body orientation. The results are consistent
with social cognitive studies showing that even though eye gaze and body orientation could both be attentional cues, when eye gaze
and body orientation are present in the same social context, eye gaze was superior in guiding attention (Cooney et al., 2017; Lawson &
Calder, 2016). Furthermore, the eye tracking data in the current study showed that learners took significantly more time to process the
instructor’s face than body. Therefore, in the context of video lectures, the instructor’s eye gaze is more likely to attract and thus guide
learners’ attention than body orientation is.
Regarding learning performance, we found that learners viewing the video lectures with the instructor’s guided gaze showed better
performance regardless of her body orientation, followed by those viewing the video lectures with the instructor’s direct gaze, and
finally those viewing the video lectures with the instructor’s averted gaze. The positive effect of guided gaze is not consistent with
previous studies (Fiorella et al., 2018; Ouwehand et al., 2015; Pi et al., 2019; van Wermeskerken & van Gog, 2017). There are two
possible explanations for the inconsistent results. First, there were differences in how the instructor’s guided gaze was used in previous
studies and in the present study. In previous studies, the instructor used guided gaze to direct attention to the slides only at several
crucial moments in the video lectures (Ouwehand et al., 2015; Pi et al., 2019; van Wermeskerken & van Gog, 2017); in the current
study, the instructor used guided gaze throughout the video to draw learners’ attention to the slides. The new type of gaze examined in
the study, namely a continuous guided gaze, might be helpful to study in an experimental context, but it is very atypical of how an
instructor would behave in an actual video lecture. Future studies should consider the effects of guided gaze in video lectures in terms
of frequency and timing.
A second reason that might account for the inconsistent findings is that the video lecture conditions were defined differently across
studies. For example, in the direct gaze condition in Fiorella et al. (2018), the instructor made eye contact with the learners while

6
Z. Pi et al. Computers & Education 144 (2020) 103713

standing behind and writing on a transparent whiteboard, and the dynamic drawing guided learners’ attention to the learning ma­
terials. In the guided gaze condition the instructor wrote on a non-transparent whiteboard, meaning that both the guided gaze and
dynamic drawings drew attention towards the learning materials. In our study, the instructors did not write on a board and the only
learning materials were the slides. In the direct gaze condition the instructor only provided eye contact, with no dynamic drawings to
guide learners’ attention; in the guided gaze condition the instructor used their gaze to guide learners’ attention towards the learning
materials. Thus the benefit of the guided gaze, relative to the direct gaze condition (which involved no guidance at all in the absence of
dynamic drawings), might have been more evident in our study than in Fiorella et al. (2018), where both the direct gaze and guided
gaze conditions provided some amount of guidance toward the learning materials.
It should be noted that previous studies that evidenced a positive effect of an instructor’s direct gaze mostly compared direct gaze
with averted gaze (Beege et al., 2019; Leong et al., 2017), whereas studies that obtained null results mostly compared direct gaze with
guided gaze (Ouwehand et al., 2015; Pi et al., 2019). Furthermore, in the studies evidencing a positive effect of direct gaze, the visual
screen of the video lectures showed an instructor in a lecture hall or an instructor writing behind a transparent whiteboard (Beege
et al., 2019, 2017; Stull et al., 2018), whereas in the studies with null results, the visual screen showed an instructor with visual
learning information presented on slides or on a non-transparent whiteboard (Ouwehand et al., 2015; Pi et al., 2019). That is, in studies
with significant effects, there was no visual-attentional competition because the visual screen only included the instructor or an
instructor seen writing behind a transparent whiteboard. Based on the current study and previous studies (Beege et al., 2017; Fiorella
et al., 2018), we conclude that whether the instructor should use direct gaze or guided gaze might depend on the characteristics of the
learning materials.
Three limitations should be considered. Firstly, we did not measure learners’ perceptions of the instructor’s experience level. These
perceptions might influence learning in video lectures, and learners might perform differently when learning from a perceived expert
than a perceived novice (Beege et al., 2019). When learners perceive the instructor as an expert, they would trust the instructor, which
is considered as fundamental for effective learning (Johnston, 2015). Secondly, the pace of the lecture should be considered in future
research. In the present study, the instructor could go at any pace they wanted during recording the lectures. There might be dif­
ferences in how well the instructor’s guided gaze helps guide attention to the slides when the instructor’s oral explanations are slow
versus fast. Thirdly, the study had high internal validity but might have had low ecological validity, as the results might not generalize
to all video lecture formats. Although the instructor/slides format is becoming increasingly popular in online learning, this study draws
attention to the need to study the value of direct and guided gaze in other types of video lectures as well. For example, a video lecture
with demonstration format includes an instructor modeling the problem solving steps or manipulating an object. Comparisons across
formats would allow us to conjecture about the type of format in which direct gaze or guided gaze would be most effective. Finally, the
reliabilities of the prior knowledge test and learning tests were somewhat low, perhaps for two reasons. The first reason was that there
were a small number of items on each test. The second reason was that the tests assessed a broad range of knowledge on the topic of
wind; therefore, the items were heterogeneous and had different levels of difficulty, which might have created variation within each
participant’s set of answers (Beege et al., 2019).
The results of this study make unique contributions to the research on social cues in video lectures. To our knowledge, this is the
first study to differentiate the effect of the instructor’s eye gaze from that of body orientation on attention and learning in this teaching
format. In addition, in earlier research both the instructor’s eye gaze and body orientation were believed to guide learners’ attention,
but no studies to date have directly tested whether one factor was more influential than the other when they are presented simul­
taneously in video lectures. In the present study, we used learners’ eye movements to test how an instructor’s eye gaze and body
orientation captured learners’ attention to the instructor and the slides. Paper-and-pencil tests also provided information about
retention and transfer.
In sum, it was striking that the video lecture instructor’s eye gaze was superior to body orientation in influencing learners’ attention
allocation and learning performance. An instructor’s guided gaze appears to have more benefits than direct gaze for learning in video
lectures, drawing learners’ attention to the important information in the learning materials. Based on the findings, the first practical
implication for designing video lectures is that an instructor does not need to consider her body orientation. Furthermore, our specific
findings about types of eye gaze have implications for designing video lectures, including that an instructor should not look directly at
the camera throughout the lecture, and should instead use guided gaze to draw learners’ attention to the learning materials.

Declaration of competing interest

Before we conducted this study, we reported it to the Ethics Committee of the School of Educational Information Technology at
Central China Normal University and received permission from the committee to conduct the research. In our study, all participants
were volunteers who provided written informed consent. Furthermore, they knew that they had the right to withdraw from the study at
any time during the experiment. We used numbers to refer to the participants instead of their names. Their data were only used for the
purpose of research.
There is no conflict of interest, as we conducted this study only because of our research interests.

Acknowledgments

This work was supported by the Research Projects of the Humanities and Social Sciences Foundation of the Ministry of Education of
China (19XJC880006; 17YJAZH104), and the National Natural Science Foundation of China (61877024).

7
Z. Pi et al. Computers & Education 144 (2020) 103713

References

Ayres, P., & Sweller, J. (2014). The split-attention principle in multimedia learning. In R. E. Mayer (Ed.), The Cambridge handbook of multimedia learning (pp. 206–226).
New York, NY: Cambridge University Press.
Beege, M., Nebel, S., Schneider, S., & Rey, G. D. (2019). Social entities in educational videos: Combining the effects of addressing and professionalism. Computers in
Human Behavior, 93, 40–52.
Beege, M., Schneider, S., Nebel, S., & Rey, G. D. (2017). Look into my eyes! Exploring the effect of addressing in educational videos. Learning and Instruction, 49,
113–120.
Clifford, C. W. G., & Palmer, C. J. (2018). Adaptation to the direction of others’ gaze: A review. Frontiers in Psychology, 9, 2165.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Cooney, S. M., Brady, N., & Ryan, K. (2017). Spatial orienting of attention to social cues is modulated by cue type and gender of viewer. Experimental Brain Research,
235(5), 1–10.
Crook, C., & Schofield, L. (2017). The video lecture. The Internet and Higher Education, 34, 56–64.
Fiorella, L., Stull, A. T., Kuhlmann, S., & Mayer, R. E. (2018). Instructor presence in video lectures: The role of dynamic drawings, eye contact, and instructor visibility.
Journal of Educational Psychology. Advance online publication https://doi.org/10.1037/edu0000325.
Hoehl, S., Wiese, L., & Striano, T. (2008). Young infants’ neural processing of objects is affected by eye gaze direction and emotional expression. PLoS One, 3(6),
e2389.
Holler, J., Schubotz, L., Kelly, S., Hagoort, P., Schuetze, M., & Ozyurek, A. (2014). Social eye gaze modulates processing of speech and co-speech gesture. Cognition,
133(3), 692–697.
Johnston, J. (2015). Issues of professionalism and teachers: Critical observations from research and the literature. Australian Educational Researcher, 42(3), 299–317.
Lawson, R. P., & Calder, A. J. (2016). The “where” of social attention: Face and body direction aftereffects arise from representations specific to cue type and not
direction alone. Cognitive Neuroscience, (7), 103–113.
Leong, V., Byrne, E., Clackson, K., Georgieva, S., Lam, S., & Wass, S. (2017). Speaker gaze increases information coupling between infant and adult brains. Proceedings
of the National Academy of Sciences, 114(50), 13290–13295.
Mayer, R. E. (2014). The Cambridge handbook of multimedia learning (2nd ed.). New York: Cambridge University Press.
Nagels, A., Kircher, T., Steines, M., & Straube, B. (2015). Feeling addressed! the role of body orientation and co-speech gesture in social communication. Human Brain
Mapping, 36(5), 1925–1936.
Ouwehand, K., van Gog, T., & Paas, F. (2015). Designing effective video-based modeling examples using gaze and gesture cues. Educational Technology & Society, 18
(4), 78–88.
Paulus, M., Murillo, E., & Sodian, B. (2016). When the body reveals the mind: Children’s use of others’ body orientation to understand their focus of attention. Journal
of Experimental Child Psychology, 148, 101–118.
Perrett, D. I., Hietanen, J. K., Oram, M. W., Benson, P. J., & Rolls, E. T. (1992). Organization and functions of cells responsive to faces in the temporal cortex [and
discussion]. Philosophical Transactions of the Royal Society of London B Biological Sciences, 335(1273), 23–30.
Pi, Z., & Hong, J. (2016). Learning process and learning outcomes of video podcasts including the instructor and PPT slides: A Chinese case. Innovations in Education &
Teaching International, 53, 135–144.
Pi, Z., Zhang, Y., Zhu, F., Xu, K., Yang, J., & Hu, W. (2019). Instructors’ pointing gestures improve learning regardless of their use of directed gaze in video lectures.
Computers & Education, 128, 345–352.
Pouw, W., van Gog, T., Zwaan, R. A., Agostinho, S., & Paas, F. (2018). Co-thought gestures in children’s mental problem solving: Prevalence and effects on subsequent
performance. Applied Cognitive Psychology, 32, 66–80.
Stull, A. T., Logan, F., & Mayer, R. E. (2018). An eye-tracking analysis of instructor presence in video lectures. Computers in Human Behavior, 88, 263–272.
Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive load theory. New York, NY: Springer.
van Wermeskerken, M., Ravensbergen, S., & van Gog, T. (2018). Effects of instructor presence in video modeling examples on attention and learning. Computers in
Human Behavior, 89, 430–438.
van Wermeskerken, M., & van Gog, T. (2017). Seeing the instructor’s face and gaze in demonstration video examples affects attention allocation but not learning.
Computers & Education, 113, 98–107.
Wilson, K. E., Martinez, M., Mills, C., D’Mello, S., Smilek, D., & Risko, E. F. (2018). Instructor presence effect: Liking does not always lead to learning. Computers &
Education, 122, 205–220.

You might also like