Does watching an explainer video help learning with subsequent text? –
Only when prompt-questions are provided

Marie-Christin Krebs *, Katharina Braschoß , Alexander Eitel
Department of Educational Psychology, Justus-Liebig-Universität Giessen, Otto-Behaghel-Str. 10, Giessen, 35390, Germany

A R T I C L E I N F O

Keywords:
Explainer video
Cognitive prompts
Video-based learning
Illusion of understanding
Situational interest

A B S T R A C T

Background: Learning with explainer videos can foster learning. However, their effects on subsequent learning are
still unclear. On the one hand, they might increase situational interest and scaffold subsequent learning. On the
other hand, they might hinder subsequent learning by fostering an illusion of understanding. In case of the latter,
the question arises of whether providing prompt-questions after an explainer video would prevent an illusion of
understanding. Therefore, we investigated the effects of medium and prompt-questions on subsequent learning
with text.
Sample: One hundred thirty-three teacher students and psychology students from a German university.
Methods: In an online study with a 2x2 between-subjects design, we investigated the effects of medium (video vs.
video-script) in learning phase 1 and prompt-questions (yes vs. no) on subsequent learning with text.
Results: As expected, watching the video made the content seem more interesting and less difficult. Contrary to
the illusion-of-understanding-assumption, this did not result in learners overestimating but rather under-
estimating themselves. Moreover, while prompt-questions in the video condition fostered learning, they impaired
learning in the video-script condition. Exploratory mediation analyses revealed that in the prompt condition, the
superiority of the video was mainly driven by the quality of the prompt-answers rather than the time learners
invested in answering the prompt-questions.
Conclusions: Our findings suggest that explainer videos combined with prompt-questions can foster learning with
subsequent text. However, further research is necessary to replicate the findings under more controlled condi-
tions and to investigate the underlying processes in greater depth.

Learning with videos is popular. Millions of learners watch so-called
explainer videos (i.e., short instructional videos, which explain com-
plex/abstract issues in an entertaining way) on all kinds of topics. It is
therefore not surprising that there has been growing interest recently in
whether explainer videos effectively support learning. While there is
evidence that videos can be an effective instructional tool, they are not
always learning-effective, a fact we have known for almost 40 years (e.
g., Salomon, 1984).

To make learning with videos effective, the video needs to be 1) well
designed and 2) well implemented within an instruction. The first aspect
has been the focus of ample previous research, resulting in various
design principles (e.g., Brame, 2016; Kulgemeyer, 2020; Mayer, 2021).
However, also highly relevant are the questions of how videos should be
implemented, for example in learning environments also containing
other types of learning materials (e.g., texts), and what problems might

emerge in this context. For instance, learners preparing for an exam
should not just rely on a short explainer video but also review other
materials such as their notes or a textbook chapter. However, relatively
little is currently known about whether and how watching explainer
videos interacts with subsequently learning from text.

Based on previous research, there are – among others – two possi-
bilities how watching an explainer video first might influence subse-
quent learning with text. On the one hand, watching the explainer video
first might foster subsequent learning from text by raising situational
interest (Endres et al., 2020) and/or by providing a mental scaffold
(Eitel & Scheiter, 2015). On the other hand, watching the explainer
video first might hinder learning from text by fostering an illusion of
understanding (Kulgemeyer & Wittwer, 2023; Salomon, 1984), where
learners prematurely think that they have understood the topic well
enough (Kulgemeyer & Wittwer, 2023; Wittwer & Renkl, 2008). This

* Corresponding author.
E-mail addresses: marie-christin.krebs@psychol.uni-giessen.de (M.-C. Krebs), katharina.braschoss@psychol.uni-giessen.de (K. Braschoß), alexander.eitel@

psychol.uni-giessen.de (A. Eitel).

Contents lists available at ScienceDirect

Learning and Instruction

journal homepage: www.elsevier.com/locate/learninstruc

https://doi.org/10.1016/j.learninstruc.2024.101988
Received 31 March 2023; Received in revised form 21 July 2024; Accepted 25 July 2024

Learning and Instruction 94 (2024) 101988 

Available online 4 August 2024 
0959-4752/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC license ( http://creativecommons.org/licenses/by- 
nc/4.0/ ). 

mailto:marie-christin.krebs@psychol.uni-giessen.de
mailto:katharina.braschoss@psychol.uni-giessen.de
mailto:alexander.eitel@psychol.uni-giessen.de
mailto:alexander.eitel@psychol.uni-giessen.de
www.sciencedirect.com/science/journal/09594752
https://www.elsevier.com/locate/learninstruc
https://doi.org/10.1016/j.learninstruc.2024.101988
https://doi.org/10.1016/j.learninstruc.2024.101988
https://doi.org/10.1016/j.learninstruc.2024.101988
http://crossmark.crossref.org/dialog/?doi=10.1016/j.learninstruc.2024.101988&domain=pdf
http://creativecommons.org/licenses/by-nc/4.0/
http://creativecommons.org/licenses/by-nc/4.0/


would reduce time and effort to process subsequent text. In case of the
latter, the question arises of how to prevent such an illusion of under-
standing. One means may be cognitive and metacognitive prompts, that
is, content-related and monitoring-related questions (e.g., Berthold
et al., 2007) provided after watching the video. There is evidence that
such prompts benefit learning from cognitive and metacognitive per-
spectives because they serve as a retrieval practice task (e.g., Roediger&
Karpicke, 2006), and foster metacognitive monitoring (e.g., Fiorella
et al., 2020; Müller & Seufert, 2018; Szpunar et al., 2014). Accordingly,
they could also foster learning with explainer videos with subsequent
expository text by stimulating beneficial learning strategies.

Against this backdrop, the aim of the present study was to shed light
on the question of whether explainer videos hamper subsequent learning
by inducing an illusion of understanding, and whether providing
learners with prompts might counter possible negative effects of
explainer videos on subsequent learning.

1. Learning with explainer videos – is it as simple as it seems?

Explainer videos, also referred to as explanatory videos (e.g., Kul-
gemeyer, 2020) or educational videos (Brame, 2016), are short
audio-visual presentations (e.g., animated videos, screencasts, sketched
explanation videos) that cover all kinds of educational topics and can
thus be considered a subcategory of instructional videos (i.e.
audio-visual presentations intended to promote learning; Mayer, 2021).
Explainer videos explain relatively complex and abstract issues in rela-
tively short time, conversational language, and usually in an enter-
taining way (e.g., Krämer & Böhrs, 2017). Explainer videos can be
roughly divided into two categories: explainer videos that present more
abstract concepts and conceptual (declarative) knowledge, and
explainer videos that convey how-to (procedural) knowledge. The latter
category is usually referred to as video tutorials and is considered a
subcategory of explainer videos (Wolf, 2015). Importantly, explainer
videos do not exhaustively explain a topic, but rather highlight relevant
issues. In educational contexts, learners watch explainer videos, for
example, to review learning content they did not understand in class or
to prepare for exams (Wolf et al., 2021). But do explainer videos support
learning generally, or do learners merely succumb to this illusion?

On the one hand, there is empirical evidence that video tutorials are
more effective for learning than text tutorials (e.g., Lloyd & Robertson,
2012; van der Meij& van der Meij, 2014). For instance, van der Meij and
van der Meij (2014) found positive effects of tutorials for software
learning. They compared four types of tutorials that included either
paper-based or video-based instruction and were accompanied by either
a short paper-based or video-based preview, resulting in the following
conditions: Paper-based preview and paper-based instruction,
paper-based preview and video procedure, video preview and
paper-based procedure, video preview and video procedure. The
paper-based or video-based preview defined the problem and displayed
the start and end screen of the task. With regard to post-test perfor-
mance, results indicated that learners with a video preview out-
performed learners in the plain-text conditions.

On the other hand, there is also empirical evidence that for declar-
ative knowledge expository texts are more effective as a first information
source compared to videos. List and Ballenger (2019) investigated the
effect of two complementary media (text vs. video) on strategy use and
comprehension. All learners received a more expository source first (text
vs. video) before they received a second more narrative source (text vs.
video). Results indicated that learners differed regarding their strategy
use based on the medium they received. For instance, learners in the text
only condition reported more cross-textual elaboration as well as orga-
nizational strategies. Moreover, learners who received a text first scored
significantly higher on a knowledge post-test.

Results of a recent meta-analysis by Noetel et al. (2021) showed that
providing learners with supplemental videos in addition to existing
learning content facilitated student learning (g > 0.2). However, their

meta-analysis summarised effects from all kinds of videos (e.g., lecture
capture, educational multimedia, tutorials), and not just explainer
videos. Other research has indicated positive effects of animated
explainer videos on self-assessed outcome variables such as interest,
engagement, and understanding a topic, but not on objective learning
outcomes such as practical and written exam scores (Liu & Elms, 2019).
Hence, the evidence is mixed, potentially because the video design and
implementation must be considered.

According to Mayer (2021), learning with instructional videos is
effective for learning if they are designed in line with three principles,
namely the dual channels principle (Paivio, 1986), limited capacity
principle (Sweller et al., 2011), and generative activity principle (Mayer
& Fiorella, 2022). Brame (2016) considered cognitive load, learner
engagement, and active learning as the three most important factors for
designing and implementing explainer videos effectively. Kulgemeyer
(2020) postulated a framework for science explainer videos, which also
included several design recommendations (e.g., using minimal expla-
nation, follow-up learning tasks, focussing on complex scientific prin-
ciples, etc.). Kulgemeyer (2020) tested this framework by comparing a
science explainer video developed with a high vs. a low fit to the
postulated framework. It was found that the learners with an explainer
video fitting the framework outperformed those with the less-fitting
explainer video regarding their declarative knowledge, thus further
supporting the assumption that learners can benefit from well-designed
explainer videos. What “well-designed” and “well-implemented” means
in particular, requires an answer from three different perspectives that
are not mutually exclusive, but probably complement each other.

From amotivational perspective, explainer videos may foster learning
especially when they are visually appealing and comprise a personalised
frame story (e.g., Endres et al., 2020). They may thus make the learning
experience more joyful due to their “emotional” design, which some-
times leads to better learning outcomes (Brom et al., 2018). Specifically,
explainer videos comprising emotional design elements lead to
better-sustained attention during learning when they trigger and
maintain learners’ situational interest (e.g., Endres et al., 2020; Hidi &
Renninger, 2006). In contrast to individual interest, which is defined as
a relatively stable personal characteristic, situational interest is
described as a reaction to environmental characteristics (Hidi, 2001).
Hence, situational interest can be influenced by the design characteris-
tics of learning materials (Endres et al., 2020). While triggered situa-
tional interest refers to superficial features of the instruction that might
shortly catch learners’ attention (Endres et al., 2020; Hidi, 1990),
maintained situational interest is responsible for maintaining focused
attention and persistence over an extended learning phase (Hidi &
Renninger, 2006). According to Endres et al. (2020), maintained situa-
tional interest should yield desirable effects on learning, particularly
when learners are required to engage with the learning material in a
longer learning phase.

From a cognitive perspective, explainer videos may help learners to
gain initial conceptual and/or procedural understanding, which serves
as a kind of cognitive scaffold for further learning (e.g., Eitel & Scheiter,
2015). In the context of multimedia learning, for example, results of a
review by Eitel and Scheiter (2015) suggest that learners benefit from
receiving the mediumwith the less complex information first. According
to Eitel and Scheiter (2015), this could be because processing the less
complex information in the first medium might facilitate the processing
of the more complex information in the second medium. Consequently,
watching an explainer video that briefly summarises relevant informa-
tion of a topic in a structured overview might also facilitate subsequent
learning processes with other media containing more complex infor-
mation (e.g., a text). Moreover, explainer videos may be even more
effective as a cognitive scaffold compared to text.

According to the Cognitive Theory of Multimedia Learning (CTML;
Mayer, 2022), textual and visual information is processed in two
different channels. Therefore, watching an explainer video that contains
visual as well as auditory elements can result in a richer mental

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

2 



representation than solely reading a text with the same information (i.e.
multimedia effect; Mayer, 2022). On the other hand, there is also
empirical evidence that text might be the more facilitative medium for
information presentation, and thus should be presented first (List &
Ballenger, 2019). Overall, it is therefore an open question whether
learners would benefit more from watching an explainer video
compared to reading the same information as a text before learning with
an expository text.

From a meta-cognitive perspective, breaking down complex informa-
tion and presenting them in an entertaining way via video might lead to
an illusion of understanding (Kulgemeyer & Wittwer, 2023; List & Bal-
lenger, 2019; Salomon, 1984), meaning that watching the (explainer)
video is easy, so understanding the contents may appear easier than it
actually is. As a consequence, learners may overestimate their level of
understanding (Kulgemeyer &Wittwer, 2023; Wittwer & Renkl, 2008),
and thus invest (too) little mental effort in further processing the
learning content, which could negatively affect their learning outcomes
(e.g., Paik & Schraw, 2013). Moreover, Kornell et al. (2011) found that
the perceived ease of information processing could bias learners’ met-
acognitive judgements (i.e. ease-of-processing heuristic; Kornell et al.,
2011). If explainer videos are easy to process or at least appear to be so,
this might result in learners overestimating their actual knowledge level
(i.e. metacognitive calibration), because they base their judgement of
expected performance on the perceived easiness of processing/memo-
rizing the video content (i.e. data-driven self-regulation; Baars et al.,
2020). Additionally, learners tend to overestimate their learning when
learning with multimedia material (i.e. multimedia heuristic; Eitel,
2016; Hoch et al., 2023; Serra & Dunlosky, 2010). Since videos can be
considered multimedia learning material, the multimedia heuristic
might also apply to learning with videos. It is possible that learners
overestimate their knowledge level after watching the explainer video
because they expect that learning with explainer videos is easy.

In the case of explainer videos, learners may also find subsequent
learning material (e.g., book chapters) irrelevant and/or unnecessarily
complicated compared to the explainer video, and thus fail to engage in
further learning processes. Previous research suggests that this is
potentially especially problematic if explainer videos contain errors or
convey scientific misconceptions, because misconceptions are close to
everyday experiences, and thus potentially more attractive than the
scientifically correct explanations (Kulgemeyer & Wittwer, 2023).
Moreover, Senko et al. (2022) found that situational interest promotes
learners’ overconfidence. This too could be especially problematic for
explainer videos, as they are usually designed to promote situational
interest. Yet this is precisely what can lead to learners being less capable
of correctly assessing their actual knowledge level and of adapting their
further learning process accordingly. Moreover, as watching a video is a
rather passive activity, learners may not engage cognitively with the
content presented to them. Consequently, they might not even become
aware of their illusion of understanding (Kulgemeyer &Wittwer, 2023).

Metacognitive monitoring is a key factor for guiding learners’ control
of subsequent learning behaviour (e.g., Dunlosky & Rawson, 2012). For
instance, there is empirical evidence that accurate metacognitive
judgements improve learning performance (e.g., Ackerman & Leiser,
2014; Dunlosky & Rawson, 2012). However, explainer videos might
make it harder for learners to monitor their knowledge level accurately.
Consequently, this might have negative effects on subsequent learning
processes.

To the best of our knowledge, up to date there is only little empirical
research on the differences regarding metacognitive calibration between
text and video as single learning materials, let alone on possible effects
on subsequent learning phases with expository texts (e.g., Mason et al.,
2022; Tarchi et al., 2021; Tarchi & Mason, 2022). Tarchi et al. (2021)
compared learning from text with learning from instructional videos and
subtitled instructional videos. In this study, results did not indicate a
significant difference between conditions regarding metacognitive
calibration, although learners reported a higher perceived self-efficacy

when learning from video compared to text. Moreover, results did not
indicate an effect of medium on immediate post-test performance but an
advantage of the text condition compared to the subtitled video condi-
tion in a delayed post-test. In another study with second language
learners, Tarchi and Mason (2022) also compared the effect of text,
instructional video and subtitled instructional video on learning. Here,
learners in the video condition outperformed learners in a text condition
in a delayed post-test, but again there were no significant differences
between the different media in an immediate post-test. Moreover, there
were also no significant differences regarding the self-rated judgements
on their performance between the media.

In a recent study, Mason et al. (2022) also compared the effect of
medium and context on learning with multiple texts versus multiple
instructional videos. Results showed that learners in the video condition
invested longer learning time compared to those in the text condition.
However, there were no significant effects of medium or context on
learners’ post-test performance. Overall, empirical evidence on the ef-
fect of learning with multiple media (i.e. explainer video and text) on
learning processes and learning outcomes is still scarce and findings
from single medium studies are mixed.

However, if explainer videos might be especially prone to foster
inaccurate metacognitive judgements, how can we encourage learners
to engage more actively with the content and more accurately assess
their current level of understanding to counter potential negative effects
on subsequent learning?

1.1. Using prompts to counter the illusion of understanding

Previous research has shown that prompts can assume multiple
forms (e.g., hints, questions) and, as a subcategory of scaffolds, can
support both self-regulated learning and learning performance (for an
overview, see Zheng, 2016).

According to Berthold et al. (2007), prompts can serve as strategy
activators, i.e., they can elicit beneficial learning strategies that learners
are capable of but do not spontaneously and/or adequately demonstrate.
For instance, Berthold et al. (2007) used writing instructions with
cognitive prompts, metacognitive prompts, or a mixture of both to elicit
cognitive and metacognitive strategies in writing learning protocols.
Their results showed that the mixture of cognitive and metacognitive
prompts had positive effects on learning outcomes. Further, providing
learners with only cognitive prompts not only elicited cognitive learning
strategies, it also triggered significantly more metacognitive strategies.
Prompts are thus effective both as strategy activators and as monitoring
aids because they stimulate the comparison of one’s current learning
state with one’s actual learning goal, and thus reveal knowledge gaps or
illusions of understanding (Müller & Seufert, 2018).

Furthermore, content-related questions after watching an explainer
video might serve as retrieval practice task (for a review, see Agarwal
et al., 2021), and thus help learners to consolidate their knowledge - an
important part of the learning process (Roelle et al., 2022). Szpunar
et al. (2013), for example, demonstrated that administering tests after
each video segment of an online video-lecture significantly reduced
students’ mind wandering, and resulted in better learning outcomes
compared to giving a test only after the last video segment. In a similar
study, Szpunar et al. (2014) showed that students again benefitted from
repeated testing during a video-lecture: They tended to be significantly
less likely to underestimate or overestimate their knowledge and per-
formed better in a post-test. Overall, these findings suggest that
content-related questions can promote metacognitive monitoring and
learning outcomes.

Furthermore, in line with the generative activity principle (Mayer &
Fiorella, 2022) content-related questions might also serve as generative
learning task. Generative learning activities (e.g., self-explanations,
drawings, etc.) encourage learners to cognitively (re)engage with the
learning material, fostering deeper processing of the content, and thus
the construction of coherent mental representations, resulting in better

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

3 



comprehension and transfer (Fiorella et al., 2020; Mayer & Fiorella,
2022; Roelle et al., 2022). Accordingly, Wang et al. (2023) showed that
prompting learners to engage in a generative learning activity (i.e.
writing a summary) between video segments benefitted learning.

According to Roelle et al. (2022), both retrieval practice tasks and
generative learning tasks are beneficial for learning, because both tasks
can support self-monitoring, and thus affect subsequent learning.
Against this background, content-related and monitoring-related ques-
tions after watching the explainer video may 1) support learners in
constructing and consolidating coherent mental representations of the
new knowledge, 2) provide implicit feedback about the learners’ current
level of understanding, and thus 3) support learners in adapting their
later learning processes accordingly. Therefore, we expected that
content-related and monitoring-related questions would prevent
learners from developing an illusion of understanding after watching the
explainer video. Moreover, we expected that all learners in our study
would benefit from these prompts in terms of their learning performance
and outcome.

1.2. The present study and hypotheses

We investigate if prompts might support learning from explainer
videos before text. To do so, we compared learning processes and out-
comes of students who learned with explainer videos or the corre-
sponding video scripts, and received prompts or not, before they should
learn with a textbook. We formulated and preregistered
(AsPredicted#84960) the following (competing) hypotheses.

H1. Situational Interest Hypothesis
Explainer videos may trigger situational interest. Therefore, learners

in the video condition should report higher triggered situational interest
in the first learning phase compared to the video-script condition (H1a).
Moreover, the higher triggered situational interest for learners in the
video condition in the first learning phase should result in higher
maintained situational interest in the second learning phase and
accordingly in learners investing more learning time in the second
learning phase, and thus performing better at the post-test (H1b).

H2. Illusion of Understanding Hypothesis
Explainer videos without prompts may trigger an illusion of under-

standing whichmight be reflected inmore passive processing behaviour.
Accordingly, learners watching the explainer video should report lower
judgements of active as well as passive mental effort and lower judge-
ments of difficulty after the first learning phase compared to learners in
the video-script condition (H2a). Moreover, we expected that learners in
the video condition without prompts would report higher judgements of
learning in the first learning phase compared to actual learning out-
comes in the post-test (i.e. metacognitive calibration) compared to the
video-script condition without prompts (H2b). Additionally, we ex-
pected that an illusion of understanding in phase 1 would lead learners
in the video condition without prompts to invest less learning time in the
second learning phase, which in turn would result in poorer learning
outcomes (H2c).

H3. Prompts Support Learning Process Hypothesis
If prompts foster learning with explainer videos because they help

learners assess their knowledge more accurately, learners with prompts
(vs. no prompts) should have more accurate metacognitive judgements
(H3a) and invest higher mental effort (H3b) and more learning time
(H3c) in the second learning phase.

H4. Prompts Support Learning Outcomes Hypothesis
We expected that all learners, but especially those with explainer

videos, would benefit from prompts. Thus, learners with prompts should
outperform those without in the post-test (H4a). In addition, learners in
the video-prompt condition should perform best (H4b).

1.3. Exploratory analyses

Against the background that explainer videos may both arouse
situational interest and serve as a cognitive scaffold, we also investi-
gated whether learners who watched the video (vs. read the video-
script) would be better at answering the prompt questions after the
first learning phase, and whether higher-quality answers would mediate
the prompts’ effect on learning outcomes.

2. Methods

2.1. Participants and design

Initially, 141 teacher students and psychology students from a
German university participated in our online learning experiment. Four
participants reported not having faithfully followed the instruction
during the study and were thus excluded from data analyses. Another
person was excluded as they had a lot of prior knowledge about the
learning topic, and another three because they reported having been
distracted while learning and also having used external resources for the
post-test. After exclusion, 133 participants (108 female; 25 male; M =

21.54 years, SD = 4.42) remained. We conducted an a priori power
analysis using G*Power 3.1 (Faul et al., 2009) for an assumed medium
effect size (alpha = 0.05; power = 0.80, f = 0.25), which resulted in a
recommended sample size of N = 128 participants.

The design of the online experiment was a 2x2 between-subjects
design with medium in the first learning phase (video vs. video-script)
and prompts after the first learning phase (no prompts vs. prompts) as
independent variables. Participants were randomly assigned to one of
four conditions: video-script without prompts (n = 35), video-script
with prompts (n = 22), video without prompts (n = 46), and video
with prompts (n = 30). The median time to complete the study was 60
min. Participants could win one of 30 gift cards (10€) as compensation
and were recruited during lectures and courses for teacher students. The
study was approved by our local ethics board (LEK FB06 2021-0051).

2.2. Materials

2.2.1. Explainer video and video script
Learning material in phase 1 was either an explainer video or the

corresponding video-script about test anxiety. The content of the ma-
terial was based on the content of the chapter “test anxiety” from the
Handbook of Educational Psychology (Rost et al., 2018). It was struc-
tured in five sections: introduction, overview, definition of test anxiety,
causes for test anxiety, and interventions. The introduction section was
designed to trigger situational interest by starting with a relatable frame
story and directly addressing participants’ prior experiences (i.e.
“Perhaps you have experienced this before: you have an important exam
approaching, [ …]”) (e.g., Endres et al., 2020). Then, participants were
given a brief overview as a structure to help them gain an initial un-
derstanding of the content (i.e. cognitive scaffold; Eitel & Scheiter,
2015). The overview was followed by the presentation of the learning
content in everyday language, which also serves as an emotional design
element. The learning material ended with a brief summary. The content
for the first learning phase was selected to cover the main aspects of test
anxiety in order to give a short overview regarding the definition of test
anxiety, causes for test anxiety, and possible interventions.

The video-script comprised 814 words and was displayed on two
pages (page 1: introduction and overview; page 2: learning content and
summary). There was no time limit, but participants could not return to
the previous page. The explainer video used sketched explanations. It
was created based on the video-script using a video creator platform (htt
ps://simpleshow.com). The explainer video’s audio-text was the video-
script narrated by a female teacher student with minimal changes in the
wording to fit the medium (for an exemplary screenshot, see Fig. 1). For
the design of the explainer video, we adhered to the multimedia

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

4 

https://aspredicted.org/j7im6.pdf
https://simpleshow.com
https://simpleshow.com


principles and design recommendations for videos (e.g., coherence
principle, temporal contiguity; Mayer et al., 2020). Moreover, the video
included social cues (e.g., a human hand that moved images in the
video). Key words (e.g., cognitive interference) were presented in text
form (Adesope & Nesbit, 2012). Participants could pause the video but
not rewind it. Similar to the script, the video was divided into two parts
and presented on two pages. Again, participants could not return to the
previous page. The video duration was 5 min and 34 s.

The main difference between the script and the video was the
wording, when the spoken/written text explicitly referred to the video/
script. However, unfortunately, part of the summary section was also
missing from the video in the study due to converting. Hence, the
amount of information in the script and the video differed slightly in the
last section. Importantly, the technical issue did not affect any additional
information relevant to the prompts or the knowledge test.

2.2.2. Learning material in phase 2
For phase 2, participants received the chapter “test anxiety” by

Cortina (2008) from the Handbook of Educational Psychology
(Schneider & Hasselhorn, 2008). The ten content pages of the chapter
were displayed on the same page so that participants could scroll
through the text. Before the text was displayed, participants received a
brief explanation for difficult terms (e.g., phylogenetic). There was no
time limit for reading the chapter. The chapter was intended to com-
plement the learning material from phase 1 by containing both the in-
formation from the video/script, as well as more in-depth information
and additional aspects (e.g., psychological theories, empirical studies,
individual differences). To be able to answer all post-test items correctly,
learners had to learn the information from both learning materials.

2.2.3. Prompts
Based on Berthold et al. (2007), we created six prompts with five

thereof relating to cognitive processes (i.e., retrieval/generative pro-
cesses), and one to metacognitive processes. Within the three retrieval
prompts, participants were asked to write down the definition, compo-
nents, and symptoms of test anxiety (prompt-question 1: “How do you
define performance anxiety, what are its components and how do they affect
learners? Please give your answer in 3-4 sentences.”), to name different
causes for test anxiety (prompt-questions 2: “What are possible causes of
severe test anxiety? Please give your answer in 1-2 sentences.”), and to
describe possible interventions from the perspective of the affected
person, the teacher, and the parents (prompt-question 3: “How can
learners, parents and teachers counteract test anxiety? Please give your
answer in 3-4 sentences.”). Within the two generative prompts, partici-
pants were asked to write down possible connections within the content
(prompt-question 4: “What connections between individual contents did you
notice while working on the learning unit? Please give your answer in 1-2

sentences.”), and to create an example of the most important content
of the topic (prompt-question 5: “Think of your own example of the most
important content of the topic. Please give your answer in 2-3 sentences.”).
To foster monitoring, we also asked participants to write down issues
they found difficult to understand (prompt-question 6: “What was
difficult to understand? This will be followed by a second learning unit based
on a textbook text on the topic (corresponds to the basic literature of the
lecture). Please write down what you want to pay more attention to when
reading the text (e.g. definition/causes/interventions).”).

We displayed the prompts only after the first learning phase and not
in between the learning material. This decision was based on the idea
that the experimental design should resemble a real learning situation
with a relatively short explainer video.

2.3. Measures

2.3.1. Prior knowledge
We assessed prior knowledge using one self-rating item and an open-

ended question. First, we asked participants to indicate their current
knowledge on the topic of test anxiety on a 0–100 percent scale (i.e.
subjective prior knowledge). Then, we instructed them as follows:
“Please write down everything you know about the topic of test anxiety. Feel
free to take a moment to think about it. Write at least 2–3 sentences.” Fifty
percent of the participants’ responses were evaluated by two indepen-
dent raters in terms of the content’s quantity and the quality (for double
coded answers ICC = 0.873). For the double-coded responses, we then
calculated an average score across both raters. For a total of five aspects
(symptoms, definition, relevance, causes, and interventions), each
answer could receive two to three points for quality, and one extra point
per aspect for a detailed answer (i.e. quantity). Overall, participants
could receive 0–12 points for quality and 0–5 points for quantity,
resulting in a maximum total score for objective prior knowledge of 17
points.

2.3.2. Learning outcome measures
Overall, the post-test comprised 13 recall and transfer items in three

different formats (i.e. multiple choice, cloze, and open-ended text).
Recall performance in this study describes learners’ ability to answer

questions that required information that could be directly derived from
the learning material. We assessed recall performance with four items:
One open-ended question, and three multiple choice items with 1–5
correct answer options each (e.g., “Test anxiety is …”, example answer
options: “… the experienced fear in a certain exam situation”, incorrect; “…
a personality trait”, correct). For the multiple-choice items, participants
were given one point for each correct answer, resulting in a maximum
score of 15 points. The open-ended question was the same as in the prior
knowledge test and thus evaluated following the same criteria, resulting
in a maximum score of 17 points. Similar to prior knowledge, 50% of the
responses were double-coded by two independent raters (for double
coded answers ICC = 0.912), and then the score of the double-coded
responses was averaged across both raters. Overall, participants could
score 0–32 points for recall performance (McDonalds’ ω = 0.75).

Transfer performance in this study describes learners’ ability to
answer questions that required a deeper understanding of the learning
content. For instance, learners’ had to apply the knowledge they gained
in the learning phase to novel scenarios as well as making generaliza-
tions and/or inferences. We assessed transfer performance with five
multiple-choice items, two open-ended questions, and two cloze items.
Again, the multiple-choice items comprised 1–5 correct answer options
each (e.g., “Is it possible that the influence of performance anxiety is
underestimated?”, example answer options: “No, because performance
anxiety is compensated for.”, incorrect; “Yes, because people with high
performance anxiety might drop out of school earlier and thus systematically
distort the results of long-term studies.”, correct). For each correct answer,
participants received one point resulting in a maximum score of 25
points. In each of the two open-ended questions, an example situation

Fig. 1. Exemplary screenshot from the explainer video.

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

5 



was briefly described from everyday teaching that involved the test
anxiety topic. To include all three intervention approaches, one of the
questions focused on effective teacher behaviours, the other on effective
parent behaviours and recommendations for those affected by test
anxiety (for an example, see OSF). Participants could receive a
maximum score of 10 points. Fifty percent of the responses to the open-
ended questions were double coded by two independent raters (for
double coded answers ICCopen transfer 1 = 0.763 and ICCopen transfer 2 =

0.698). For the double-coded answers, an average score was calculated
for both raters. The first cloze question comprised four text gaps with
2–3 options and one correct answer each (“What are potential effects of
high test anxiety?”, e.g., “Learners with high test anxiety usually invest
[less, incorrect]/[more, correct]/[the same amount, incorrect] of time in
preparing for exams as other learners.”). The second cloze question
comprised two text gaps with five answer options and one correct
answer each. Because of a technical problem, only one of the two gaps
could be scored. For each correct answer, participants received one
point, resulting in a maximum score of 5 points. Overall, participants
could score 0–40 points for transfer performance (McDonalds’ ω =

0.52).

2.3.3. Prompts
We assessed participants’ responses on the six prompts in terms of

quantity and quality but used only the answers to the cognitive-related
prompts (i.e. retrieval prompts and generative prompts) for the
exploratory analyses. For quantity, we assessed the overall word count.
To assess the quality, the responses to each of the six prompt-questions
were coded, with 50% of responses double-coded (ICC = 0.947). For the
three retrieval prompts, participants could receive a maximum total
score of 25 points for correct responses. For the generative prompts,
participants could receive 0–2 points for their own example, and one
point for each mention of a relevant connection within the topic. We
constructed a scale based on the five retrieval and generative prompts to
measure possible effects of cognitive-related prompts (McDonalds’ ω =

0.63). Participants could earn 0–2 points for the metacognitive prompt.

2.3.4. (Meta-)Cognitive measures
After the introduction, we asked participants to rate on a 0–100

percent scale the difficulty of learning with the rest of the script/video (i.
e. EOL; ease of learning).

After each learning phase, we also assessed participants’ judgement
of learning (JOL), judgement of difficulty (JOD), mental effort, and
cognitive load.

To assess participants’ JOL, we asked them to rate on a 0–100
percent scale how confident they were that they would be able to answer
questions about the learning content on a test at the end of the study. To
calculate their metacognitive calibration (i.e., how well their self-
assessment matched their objective learning outcomes) after the first
learning phase, we first scaled their learning outcome on a 0–100
percent scale and then subtracted that value from their JOL-score from
the first learning phase. For metacognitive calibration, a value over zero
indicated overestimation and a value below zero underestimation. For
participants in the prompt condition, we also assessed their specific
judgements after the first learning phase on a 0–100 percent scale
regarding the three content sections covered in the script/video. To
assess participants’ JOD, we asked them to rate on a 7-point Likert scale
how difficult they found the learning unit to be (1= very easy to 7= very
difficult).

We also assessed participant’s active mental effort (e.g., “I exerted
myself to read the text”) and passive mental effort (e.g., “It has been
strenuous to read the text.”) with two items based on Klepsch and Seufert
(2021). We also assessed cognitive load via the Cognitive Load Ques-
tionnaire (CLQ; Klepsch et al., 2017). For this, participants rated eight
adapted statements regarding intrinsic (phase 1: Cronbach’s α = 0.73,
phase 2: Cronbach’s α = 0.83), extraneous (phase 1: Cronbach’s α =

0.80, phase 2: Cronbach’s α = 0.79), and germane cognitive load (phase

1: Cronbach’s α = 0.44; phase 2: Cronbach’s α = 0.66). All items were
assessed on a 7-point Likert scale (1 = absolutely wrong to 7 = absolutely
right).

2.3.5. Situational interest
We assessed situational interest based on Endres et al. (2020) along

two dimensions (i.e. affect and value; Schiefele, 2009). Participants’
affect was measured by asking participants to rate the fit of the adjec-
tives "exciting," "entertaining," and "boring" (reverse-coded) for the
materials and content. Participants’ value was measured by asking them
to rate the fit of the adjectives “useful,” “unnecessary” (reverse-coded),
and “unimportant” (reverse-coded) for the materials and content. Both
measures were rated on a 9-point Likert scale (1 = not at all to 9 = very
much). Based on Endres et al. (2020), for triggered situational interest,
we calculated a mean score for the items of the
affect-towards-the-material subscales (phase 1: Cronbach’s α = 0.80,
phase 2: Cronbach’s α = 0.65). For maintained situational interest, we
calculated a mean score for the items of the affect-towards-the-content
and value-towards-the-content subscales (phase 1: Cronbach’s α =

0.78, phase 2: Cronbach’s α = 0.80). Relying on Endres et al. (2020), we
also calculated an overall measure of positive affect by calculating a
mean score for the items of the affect-towards-the-material and
affect-towards-the-content subscales for the first learning phase (phase
1: Cronbach’s α = 0.80, phase 2: Cronbach’s α = 0.74).

2.3.6. Further measures
Because the present study was conducted online, we included a short

procedure check at the end of the experiment. First, all participants were
asked to answer truthfully, as this would not put them at any disad-
vantage. We then assessed their genuine participation with three items
(e.g., “Did you use external resources (e.g., Google, notes, etc.) to help you
complete the post-test?” “yes” vs. “no”), followed by a question regarding
which device they had used to participate in the study to check for
technical issues.

The measures below were also assessed for exploratory reasons but
not included in our analyses: Personal interest (Schiefele et al., 1993),
academic self-concept (Schöne et al., 2002), learning strategies (Kling-
sieck, 2018), intrinsic motivation (Wilde et al., 2009), and prior expe-
rience with explainer videos.

2.4. Procedure

After having provided informed consent to participate in the online
study, participants answered demographic questions, questions on their
personal interest and their academic self-concept regarding educational
psychology, as well as their general use of learning strategies and their
prior knowledge on the topic of test anxiety. They were then assigned to
either the video or the video-script condition for the first learning phase.

After reading or watching the short introduction, participants rated
their perceived difficulty of learning with the rest of the script or video
(i.e. ease of learning). Then, participants read or watched the rest of the
video-script or video. They then rated their cognitive load, perceived
difficulty, and situational interest regarding the content and design of
the learning material. Afterwards, half of the participants received six
prompts, and then rated how well they would be able to answer ques-
tions about the topic in a test at the end of the study. Participants in the
no-prompt condition rated their judgement of learning only globally,
while participants in the prompt condition also rated their specific
judgements regarding the three content sections covered in the script or
video.

Then, the second learning phase started. There was no time limit for
reading the book chapter, so that all participants were able to finish the
learning phase at their own pace. After the second learning phase, we
again assessed participants’ judgement of learning, cognitive load,
judgement of difficulty, and situational interest. This was followed by
the post-test. Because the study was conducted as an online study, we

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

6 

https://osf.io/tuq3h/?view_only=726bed246f8a4e8697525cce98c331c0


were not able to control their potential use of learning aids, thus we
explicitly instructed them to put those aside.

After participants completed the post-test at their own pace, we
assessed their interest and intrinsic motivation regarding their entire
learning experience. At the end of the study, we asked them to answer
three procedure-check items addressing their faithful participation, and
to indicate which device they used in the study. Finally, all participants
were debriefed and forwarded to a new questionnaire where they could
enter their e-mail addresses if they wanted to participate in the lottery
for the gift cards.

3. Results

For all statistical analyses, we applied an alpha level of α = 0.05, and
one-tailed test statistics for directional hypotheses. As indices for effect
size, for Analyses of Variances and Covariance we used ηp2 with values of
0.01, 0.06, 0.14, and for t-tests Cohens’ d with values of 0.2, 0.5, 0.8 as
small, medium, and large effect size, respectively. For mediation ana-
lyses, we used dummy coding for the experimental conditions medium
(coding [0] video-script and [1] video) and prompts (coding [0] no
prompts and [1] prompts).

3.1. Preliminary analyses

Analyses of Variance (ANOVAs) revealed no significant differences
between experimental conditions regarding subjective, F < 1, or
objective prior knowledge, F(1,129) = 1.62, p = .206, ηp2 = 0.01. Only
objective prior knowledge correlated significantly with post-test per-
formance (r = 0.20, p = .021). We thus included objective prior
knowledge as a covariate in Analyses of Covariance (ANCOVAs) and
mediation analyses with learning outcome as dependent variable.

3.2. Main analyses

3.2.1. Situational Interest Hypothesis (H1)
We conducted a one-tailed independent-samples t-test to compare

participants’ triggered situational interest after the first learning phase
between the video and video-script condition. In line with H1a, our
results revealed a significant difference: Learners in the video condition
reported significantly higher triggered situational interest (phase 1)
than learners in the video-script condition, t(131) = − 4.69, p < .001,
Cohens’ d = 0.82 (see Table 1 for the descriptive statistics).

To test whether triggered situational interest would result in higher
maintained situational interest and in learners investing more learning
time in the second learning phase, and thus performing better at the
post-test (H1b), we conducted two separate three-step-mediation ana-
lyses for recall and transfer performance, respectively (Model 6, Hayes,
2022). For each of the two dependent variables, we entered medium
(coding [0] video-script and [1] video) as independent variable, trig-
gered situational interest (phase 1), maintained situational interest
(phase 2) and invested learning time (phase 2) as mediators, and prior
knowledge as covariate. Against our expectations, our results indicated
no significant specific indirect effect either for recall, b = 0.02, SE =

0.03, 95% CI [− 0.02 | 0.10] (partially standardized indirect effect: b =

0.003, SE = 0.01, [-0.004 | 0.02]) or transfer performance, b = 0.01, SE
= 0.02, 95% CI [− 0.01 | 0.07] (partially standardized specific indirect
effect: b = 0.004, SE = 0.01, [− 0.004 | 0.02]).

3.2.2. Illusion of understanding hypothesis (H2)
Because lower mental effort and a lower judgement of difficulty for

the first learning phase can reflect an illusion of understanding, we
investigated whether there would be a difference between the video and
the video-script condition. For this, we conducted three one-tailed in-
dependent-samples t-tests with active mental effort (phase 1), passive
mental effort (phase 1) and perceived difficulty of the learning content
(phase 1) as dependent variables (see Table 1 for the descriptive

statistics). In line with our expectations, learners in the video condition
reported significantly less active mental effort, t(131) = 1.86, p = .032,
Cohens’ d = 0.33, less passive mental effort, t(103.7) = 1.77, p = .040,
Cohens’ d= 0.32, and assessed the content overall as being less difficult, t
(131) = 1.67, p = .049, Cohens’ d = 0.29, compared to learners in the
video-script condition, thus supporting the assumption that explainer
videos seem easier than text (H2a).

We also expected that without prompts learners watching the
explainer video would tend to overestimate their knowledge level
(H2b). To test this hypothesis, we conducted a one-tailed independent
simple t-test with medium (video vs. video-script) as independent vari-
able and metacognitive calibration as dependent variable for the no-
prompt condition. Against our expectations, for learners without
prompts, we identified no significant difference between the video and
video-script conditions, t(76) = 1.25, p = .108, Cohens’ d = 0.28.

Moreover, we expected that an illusion of understanding in the first
learning phase would lead to less learning time and effort in the second
learning phase, thus resulting in poorer learning outcomes (H2c). Thus,
for the no-prompt condition, we conducted two separate mediation
analyses with the dummy coded medium condition ([0] video-script and
[1] video) as independent variable, the two sequential mediators active
mental effort and learning time in phase 2, for recall and transfer per-
formance as dependent variables, respectively, and prior knowledge as
covariate (Model 6; Hayes, 2022). Against our hypotheses, our results
indicated no significant specific indirect effects either for recall, b =

− 0.21, SD= 0.18, 95% CI [− 0.66 | 0.05] (partially standardized indirect
effect: b = − 0.05, SE = 0.04, [− 0.15 | 0.01]) or transfer performance, b
= − 0.18, SE = 0.15, 95% CI [− 0.52| 0.05] (partially standardized in-
direct effect: b = − 0.06, SE = 0.05, [− 0.16 | 0.02]).

3.2.3. Prompts support learning process hypothesis (H3)
To test our hypothesis that prompts (vs. no prompts) after the first

Table 1
Descriptive statistics of important variables in phase 1.

Condition Range

Explainer
video

Video-
script

n = 76 n = 57

M (SD) M (SD)

Triggered situational interest (phase 1) 5.82 (1.65) 4.43 (1.77) 1–9
Maintained situational interest (phase
1)

7.27 (1.08) 6.98 (1.10) 1–9

Active mental effort (phase 1) 3.50 (1.53) 4.00 (1.54) 1–7
Passive mental effort (phase 1) 2.28 (1.18) 2.70 (1.50) 1–7
Judgement of difficulty (phase 1) 2.70 (0.86) 2.98 (1.11) 1–7

prompts no
prompts

prompts no
prompts

n = 30 n = 43 n = 22 n = 35

M (SD) M (SD) M (SD) M (SD)

Cognitive prompts
(quality)

11.43
(4.00)

– 9.07
(2.86)

– 0-27
(min.)*

Cognitive prompts
(amount of
words)

185.00
(84.98)

– 163.14
(77.02)

– –

Metacognitive
prompt (quality)

1.10
(0.89)

– 1.18
(0.66)

– 0–2

Metacognitive
prompt (amount
of words)

11.43
(12.55)

– 13.36
(13.45)

– –

Metacognitive
calibration
(phase 1)

− 11.41
(25.83)

− 8.66
(19.03)

− 4.70
(22.50)

− 3.35
(18.29)

− 100-
100

Note: For the second generative prompt, participants could receive one point for
each mention of a relevant connection. Thus, there was no upper point-limit for
this question.

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

7 



learning phase would result in more accurate metacognitive judgements
about the learning outcomes in a post-test (H3a), we calculated an
ANOVAwith metacognitive calibration (Phase 1) as dependent variable.
Due to missing values, three participants were excluded from the anal-
ysis. Against our expectation, there was neither a main effect of medium,
F(1,125) = 2.33, p = .129, ηp2 = 0.02, nor of prompts, F < 1, nor a sig-
nificant interaction effect between medium and prompts, F < 1.

Moreover, we expected that learners in the prompt condition (vs. in
the no-prompt condition) would invest more mental effort (H3b) and
more learning time (H3c) in phase 2. Therefore, we conducted two
separate ANOVAs with active mental effort (phase 2) and learning time
in the second learning phase as dependent variables.

Our results for active mental effort (phase 2) indicated a significant
main effect for the prompt condition, F(1,129) = 6.57, p = .012, ηp2 =

0.05, but not for the medium condition, F(1,129) = 2.02, p = .158, ηp2 =
0.02. Against H3a, our results indicated that with prompts learners re-
ported less active mental effort during the second learning phase.
However, based on the significant interaction between the experimental
conditions, F(1,129) = 10.27, p = .002, ηp2 = 0.07, this difference is
mainly attributable to the significant difference in the video-script
condition (see Table 2 for the descriptive statistics). Results from sim-
ple comparisons indicated that learners with prompts in the video-script
condition reported significantly less active mental effort than those
without prompts (Mdiff = − 1.71, SE = 0.45, p < .001). For the video
condition, this difference was not significant (Mdiff = 0.19, SE = 0.39, p
= .624). Moreover, with prompts, learners in the video condition re-
ported significantly more active mental effort compared to learners in
the video-script condition (Mdiff = 1.37, SE = 0.46, p = .004). For the
condition without prompts, the difference was not significant (Mdiff =

0.53, SE = 0.37, p = .156).
Regarding invested learning time in phase 2, the medium (video vs.

script) also revealed no main effect, F < 1. However, we observed a
significant main effect for the prompt condition, F(1,124) = 4.05, p =

.046, ηp2 = 0.03. But against H3b, learners with prompts invested
significantly less learning time than those without. Furthermore, the
interaction between experimental conditions was not significant, F

(1,124) = 3.56, p = .061, ηp2 = 0.03.
A possible explanation for why we failed to detect a positive effect of

prompts on mental effort and learning time in the second learning phase
might be that not all learners needed to use the indirect feedback from
answering the prompts for the second learning phase. Possibly, only
learners who accurately judged their performance in the retrieval
prompts as poor used this information to adjust their behaviour in the
second learning phase. Against this background, we performed two
exploratory quadratic regression analyses for those learners whose
performance in the retrieval prompts were overall below 50% percent.
In the first quadratic regression analysis, we entered metacognitive
calibration (phase 1) for the retrieval prompts as independent variable
and invested learning time (phase 2) as dependent variable. Against our
assumption, the regression model with the quadratic term was not sig-
nificant, F(2,36) = 2.04, p = .144, ηp2 = 0.05. In the second quadratic
regression analysis, we entered metacognitive calibration for the
retrieval prompts as independent variable, and active mental effort
(phase 2) as dependent variable. Again, however, the regression model
with the quadratic term was not significant, F < 1.

Another possible explanation is that learners in the video-script
condition with prompts were demotivated, and thus invested less
mental effort and learning time in the second learning phase. To test this
assumption, we calculated an exploratory ANOVA with participants’
intrinsic motivation as dependent variable, and medium (video vs.
video-script) as well as prompt condition (prompt vs. no-prompt) as
independent variables. Results revealed a significant main effect of
medium, F(1,129) = 8.52, p = .004, ηp2 = 0.06, indicating a higher
intrinsic motivation for learners in the video condition. For the prompt
condition, the main effect was not significant, F < 1. The interaction
between medium and prompt condition was also not significant, F
(1,129) = 3.87, p = .051, ηp2 = 0.03. Descriptively, however, learners in
the video-script condition with prompts reported the lowest intrinsic
motivation regarding their learning experience at the end of the study
(see Table 2 for mean scores).

It is possible that the demotivation for learners in the video-script
condition with prompts was related to more effort in answering the
prompts after the first learning phase which might have let to more
motivational depletion during the second learning phase. To test this
assumption, we calculated two exploratory independent t-tests. First, we
entered the time learners spent on the prompts as dependent variable
and medium (video vs. video-script) as independent variable. There was
no significant difference between the video and the video-script condi-
tion, t(50) = − 0.16, p = .870, Cohens’ d = 0.05, indicating that learners
in both conditions spent equally long on answering the prompt questions
after the first learning phase. Second, we entered the number of words
participants wrote to answer the prompt questions as dependent vari-
able and medium (video vs. video-script) as independent variable.
Again, the difference was not significant, t(50) = − 0.82, p = .417,
Cohens’ d= 0.23. On average, learners in both conditions wrote a similar
number of words to answer the prompt questions. We also tested
whether there would be a correlation between time spent on answering
the prompts and time spent in the second learning phase. In contrast to
the motivational depletion assumption, results indicated a positive
correlation, r= 0.405, p= .004, indicating that learners who spent more
time to answer the prompt questions also invested more learning time in
the second learning phase.

3.2.4. Prompts support learning outcomes hypothesis (H4)
To test whether the medium (video vs. video-script) and prompt

condition interaction (prompt vs. no prompt) would influence the
learning outcome, we conducted two ANCOVAs for recall and transfer
with prior knowledge as covariate (see Table 2 for the descriptive
statistics).

For recall, there was neither a main effect for medium, F(1,128) =
2.40, p = .124, ηp2 = 0.018, nor for the prompt condition, F < 1. How-
ever, our results indicated a significant interaction between

Table 2
Descriptive statistics of important variables in phase 2.

Condition Range

Explainer Video Video-script

prompts no
prompts

prompts no
prompts

n = 30 n = 43 n = 22 n = 35

M (SD) M (SD) M (SD) M (SD)

Triggered
situational
interest (phase 2)

2.82
(0.98)

2.96
(1.69)

3.59
(1.76)

3.50
(1.60)

1–9

Maintained
situational
interest (phase 2)

5.79
(1.38)

6.00
(1.26)

5.95
(1.48)

6.36
(1.28)

1–9

Active mental
effort (phase 2)

5.23
(1.72)

5.04
(1.61)

3.86
(1.83)

5.57
(1.52)

1–7

Passive mental
effort (phase 2)

1.97
(0.85)

2.48
(1.33)

2.77
(1.63)

2.66
(1.43)

1–7

Judgement of
difficulty (phase
2)

5.00
(0.89)

4.70
(1.07)

5.09
(0.92)

4.63
(1.09)

1–7

Recall performance 17.58
(5.12)

14.54
(4.27)

13.14
(3.54)

16.37
(4.71)

0–32

Transfer
performance

30.40
(3.09)

30.23
(3.03)

28.93
(2.35)

30.99
(3.25)

0–40

Metacognitive
calibration
(phase 2)

− 31.51
(20.00)

− 26.88
(23.40)

− 26.20
(18.20)

− 23.43
(14.46)

− 100-
100

Intrinsic
motivation

3.39
(0.73)

3.27
(0.68)

2.80
(0.65)

3.16
(0.62)

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

8 



experimental conditions, F(1,128) = 13.57, p < .001, ηp2 = 0.10. Simple
comparisons within the prompt condition revealed a significant differ-
ence between the video and video-script condition (Mdiff = 4.18, SE =

1.25, p = .001). In line with H4b, learners in the video condition per-
formed better than those in the script condition. With regard to the no-
prompt condition, simple comparisons revealed no significant difference
(Mdiff = − 1.72, SE= 1.00, p = 0.087). Furthermore, simple comparisons
within the video-script condition revealed a significant effect of prompts
(Mdiff = − 2.88, SE = 1.22, p = .019), with learners in the prompt con-
dition performing worse than those in the no-prompt condition. Simple
comparisons within the video condition also revealed a significant effect
of prompts (Mdiff = 3.01, SE = 1.04, p = .004). Here, learners in the
prompt condition performed significantly better than those in the no-
prompt condition (Fig. 2).

Against H4, for transfer, there was neither a main effect of medium, F
< 1, nor of the prompt condition, F(1,128) = 2.68 p = .104, ηp2 = 0.02,
nor a significant interaction, F(1,128) = 3.74, p = .055, ηp2 = 0.03.

3.3. Exploratory analyses

To test whether explainer videos serve as cognitive scaffold, we
tested whether learners who watched the video (vs. read the video-
script) would be better at answering the retrieval and generative
prompt questions after the first learning phase. For this, we calculated a
mean score of the quality scores for the answers to the retrieval and the
generative prompt questions. An ANCOVA for learners in the prompt
condition with medium (video vs. video-script) as independent variable
and prior knowledge as covariate, was not significant, F(1,49) = 4.00, p
= .051, ηp2 = 0.08. Descriptively, however, learners in the video condi-
tion wrote better answers to the retrieval and generative prompts than
those in the video-script condition. Moreover, we tested whether
learners in the video condition spent more time answering the prompt-
questions. Results of an independent-samples t-test did not show sig-
nificant differences between the video and video-script condition, t(50)
= − 0.16, p = .871, Cohens’ d = 0.46.

Furthermore, we conducted two exploratory mediation analyses for
learners in the prompt condition (Fig. 3). We entered the dummy coded
medium condition (coding [0] video-script and [1] video) as indepen-
dent variable, prior knowledge as covariate, time-on-prompts and
quality of prompt-answers for the retrieval and generative prompts as
potential mediators, and recall and transfer performance as dependent
variables, respectively (Model 4; Hayes, 2022).

Results of the mediation analysis with recall as dependent variable
indicated that only the quality of the prompt-answers partially mediates
the effect of medium (video vs. video-script) on recall, b = 1.91, SE =

0.90, 95% CI [0.24 | 3.80] (partially standardized indirect effect: b =

0.38, SE = 0.16, 95% CI [0.05 | 0.71]), but not time-on-prompts. For
transfer as dependent variable, we found similar results. Again, only the
quality of the prompt-answers mediates the effect of medium (video vs.
video-script) on recall, b = 0.79, SE = 0.48, 95% CI [0.03 | 0.87]
(partially standardized indirect effect: b= 0.27, SE= 0.16, 95% CI [0.12
| 0.62]).

4. Discussion

The aim of the present study was twofold. First, we investigated
whether watching explainer videos before text fosters an illusion of
understanding and is thus detrimental for subsequent learning with
textbook material. Second, we investigated whether prompts provided
as content-related and monitoring-related questions would prevent
learners from developing an illusion of understanding, and thus foster
learning with explainer videos and subsequent textbook material.

In summary, although we found no evidence that learners in our
study developed an illusion of understanding after watching the
explainer video, they did benefit from the prompt-questions for their
learning outcome. Exploratory analyses indicated that this beneficial
effect was mainly driven by the quality of the prompt-answers. Sur-
prisingly, learners in the video-script condition did not benefit from
prompt-questions, and actually performed worse in the post-test than
those without.

Regarding the situational interest hypothesis, in line with our ex-
pectations, we found that learners in the video condition reported more
triggered situational interest towards the learning material and the
content than learners in the video-script condition (H1a). This concurs
with previous research showing that explainer videos can trigger situ-
ational interest (Endres et al., 2020) - a precondition for further learning
processes. Against the situational interest hypothesis, however, learners
failed to invest more learning time after watching the explainer video.
Thus, the explainer video was helpful to catch learners’ attention, but
that had no beneficial effects on the subsequent learning with the text-
book chapter.

As a competing hypothesis, we formulated the illusion-of-
understanding hypothesis (H2). We expected that if explainer videos
fostered an illusion of understanding (Paik & Schraw, 2013), learners in
the video condition would exhibit signs of superficial processing, (e.g.,
perceiving the content to be easier, overestimating their level of un-
derstanding), and consequently invest less effort and time in the second
learning phase. In line with our expectations, we found that learners in
the video condition did report significantly less active and passive
mental effort, and assessed the content to be less difficult, which might
be initial indications that those learners invested less effort in processing
the content in the explainer video.

Nevertheless, that had no significant effect on their subsequent
learning. Our results do not reveal that learners overestimated their
knowledge after having watched the explainer video, or that this
resulted in more ineffective learning in the second learning phase and
thus in poorer learning outcomes. Against our expectations, we did not
find that learners in the video condition were significantly less well
metacognitively calibrated compared to learners in the video-script
condition. Similarly, Tarchi et al. (2021) found that even if learners in
a video condition perceived higher self-efficacy regarding their learning
compared to learners in a text condition, this was not mirrored by their
calibration judgements. Interestingly, in our study, learners in the video
condition tended to underestimate their performance - and not over-
estimate it, as we would have expected. Moreover, providing prompts
descriptively increased the underestimation. Consequently, our results
for metacognitive calibration did not support the illusion of under-
standing hypothesis (H2).

However, our results showed that learners in the video condition
reported more passive cognitive processing during the first learning
phase compared to learners in the video-script condition. It is possibleFig. 2. Recall performance as a function of condition.

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

9 



that they became aware that this might be problematic for learning,
which might have resulted in a more humble metacognitive mindset.
This might also explain the fact that learners in the video condition with
prompt-questions descriptively underestimated their knowledge level
even more than those in the video condition without prompt-questions.
This is because if learners became aware of their superficial processing
after watching the explainer video and were also prompted to explicitly
reflect on their learning process by answering the prompt-questions, this
could have led to a greater underestimation, as learners might have also
compensated for the perceived easiness of the video. In sum, it is
possible that learners experienced an illusion of understanding during
the first learning phase and reflecting on their learning behaviour might
have resulted in learners underestimating themselves more.

Overall, even though we found that learners watching the explainer
video were more prone to superficial processing which might indicated
an illusion of understanding during watching the video, this was not
reflected in their metacognitive calibration afterwards. Hence, our re-
sults do not support the illusion-of-understanding hypothesis (Salomon,
1984).

Concerning the use of prompt-questions after the first learning phase,
we found - in contrast to previous research (e.g., Berthold et al., 2007) -
that providing prompts was not necessarily helpful. In contrast to our
expectations that providing learners with prompt-questions would help
them monitor and consolidate their knowledge (e.g., Eitel et al., 2022;
Roelle et al., 2022), we failed to detect any general beneficial effect of

prompts on learners’ metacognitive calibration (H3) or learning out-
comes (H4).

A possible explanation for why we failed to detect a general positive
effect of prompt-questions on mental effort (H3b) and learning time in
the second learning phase (H3c) is that this might only apply for learners
who correctly rated their performance as poor after the first learning
phase. In contrast, learners who correctly rated their performance as
good might have required less effort and time because they could skim
information they already knew. Therefore, we performed an exploratory
non-linear regression analysis for learners with a recall performance
below 50% to test for a U-shaped relationship between learners’ meta-
cognitive calibration based on their recall performance and mental
effort or learning time in the second learning phase. Unfortunately, our
data did not support this alternative explanation.

For learning outcomes, our analyses’ results indicate that while
learners in the video condition (in line with our expectations and pre-
vious research, e.g., Szpunar et al., 2014) benefitted from prompts
regarding their learning outcome, learners in the video-script condition
even performed worse than those without prompts. It is possible that
learners in the video-script condition with prompts were more demoti-
vated compared to the learners in the other conditions. In line with this
assumption, we found that learners in the video-script condition with
prompts reported the lowest intrinsic motivation at the end of the study.
We tested in exploratory analyses whether an increased effort in form of
invested time and produced content in the prompt phase might explain a

Fig. 3. Results of the exploratory mediation analyses with recall and transfer as dependent variables.

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

10 



‘motivational depletion’ for learners in the script condition with
prompts. In contrast to this possible explanation, however, we did not
find significant differences for invested time and amount of content
regarding the prompts between learners in the video-script and the
video condition. Rather, we found a positive correlation between the
time learners spent on the prompts and the time learners invested in the
second learning phase, which also speaks against the assumption that
learners who invested more time to answer the prompt-questions were
less motivated to invest learning time in the second learning phase.
Overall, there appears to be a motivational problem in the video-script
condition with prompts but we were not able to determine the source
of the problem. It might be interesting to take a closer look at a possible
influence of epistemically-related emotions in this context (Pekrun et al.,
2017) in future studies.

For learners in the video condition with prompts, exploratory ana-
lyses revealed that the superiority of the video compared to the video-
script was mainly driven by the quality of the prompt-answers.
Learners in the explainer video condition delivered descriptively
higher quality answers for the retrieval and generative prompts which in
turn related positively to better recall performance.

Due to our study design, time-on-task in phase 1 (i.e. time spent on
video/video-script and time spent on answering the prompt-questions)
differed between the prompt conditions (prompt vs. no-prompt),
which might have influenced the effects of prompts on learning out-
comes in our study. However, answer quality rather than the mere time
learners took to answer prompt-questions led to better recall perfor-
mance in the video condition. In the conditions with prompts, learners
spent a similar amount of time answering the prompt-questions. Yet,
learners in the video condition were better at answering the prompt-
questions, which in turn lead to better recall performance. In contrast,
learners in the video-script condition invested a similar amount of time
in answering the prompt-questions. Consequently, we would argue that
they were similarly motivated to answer the prompt-questions. How-
ever, they performed worse in answering the prompt-questions after
phase 1 and in answering the recall questions at the end of the study.
Taken together, these results indicate that learners in the video condi-
tion were better able to utilise the prompt-questions to their advantage.

Against this backdrop, explainer videos in combination with prompt-
questions can have a positive influence on subsequent learning with
textbook material. On the one hand, explainer videos can trigger situ-
ational interest in the topic (Endres et al., 2020), and on the other hand,
they can – like pictures – scaffold subsequent learning from text (Eitel &
Scheiter, 2015). However, our results also show that the positive effects
of explainer videos only emerged when prompt-questions were pro-
vided. Although we detected no empirical evidence that learners in our
study benefitted from the prompt-questions in metacognitive moni-
toring and regulation terms, it still appears that learners in the video
condition were able to use the prompts to consolidate the knowledge
they had acquired in the explainer video, and thus benefitted from
prompts. Overall, in terms of the use of prompts, our results reveal that
the effective provision of instructional support is often not as simple as it
might first appear.

4.1. Limitations, implications and conclusions

We conducted the present study as an online experiment. On the one
hand, this can be problematic, as we had no control over what the
learners actually did during the experiment. To counter this problem, we
implemented control questions at the end of the study, checked for a
reasonable participation timeframe, and excluded participants who
stated that they were too distracted during learning and used external
resources for the post-test. Another challenge associated with the
reduced control is the limited insight into the reasons why participants
withdrew from the study. Overall, the dropout was in an acceptable
range for an online learning study. After the start of the first learning
phase, 40 participants terminated the study early. While the dropout of

these 40 participants was not equally distributed across the four con-
ditions (Х2(3) = 20.01, p < .001), a post-hoc analysis revealed that this
was mainly due to the fact that overall learners in the explainer video
condition without prompts dropped out significantly less than expected.
A closer look at the dropout revealed that in the first learning phase
significantly more learners from the video condition (n = 12) compared
to learners from the video-script condition (n = 3) dropped out (Х2(1) =
3.14, p = .041). It is possible that they dropped out due to smaller
technical issues (e.g., problems with video sound). Besides smaller
technical issues, it is also possible that some participants in the video
condition did not like the design of the video (e.g., the pictures, the
voice, the speed), and decided to withdraw because it was not possible to
skip the video or to increase the speed. However, it is also possible that
individual characteristics affected the dropout. Consequently, we took a
closer look at potential variables that might have affected a dropout. For
this, we conducted Mann-Whitney-U-Tests between participants who
withdrew from the study in the first learning phase and the other par-
ticipants. We neither found significant differences for personal interest
in educational psychology (U = 1537.000, Z = − 0.002, p = .998), nor
academic self-concept (U = 1711.000, Z = 0.912, p = .362), nor self-
rated prior knowledge (U = 1772.000, Z = 1.210, p = .226). During
answering the prompt-questions, an equal number of participants
dropped out from the video condition (n = 12) and the video-script
condition (n = 11). In the second learning phase, the dropout (n = 9)
was equally distributed between the experimental conditions (Х2(3) =
4.49, p = .213).

An advantage of the online setting, on the other hand, is that it en-
ables greater ecological validity, as learners – similar to a real learning
situation – self-regulated their learning with the learning material in a
familiar environment. Of course, it must be taken into account that the
learning material used in the study is to a certain extent artificial. For
example, the use of the video-script as learning material is not entirely
comparable with more complex text materials used by learners in nat-
ural learning contexts. Nonetheless, the video-script presents a narrative
summary of learning-relevant information which can also be found in a
similar form in textbooks. However, for further research, it would be
interesting to replicate the study in a lab and to collect more objective
process measures, such as eye-tracking data, to see what learners actu-
ally do while learning with different materials. Moreover, it could also
be interesting to examine factors that might influence the effectiveness
of explainer videos in authentic learning situations, such as prior
knowledge, mind wandering, distraction, and the use of supplemental
resources during test taking in more depth.

With regard to transfer performance as learning outcome measure, it
is possible that we failed to detect significant results because the
learning phases were too short. It is also possible that an effect of the
experimental conditions on transfer performance would only be visible
in a delayed post-test. Nonetheless, in typical studies on (multimedia)
learning, study sessions are even shorter than it was here, andMayer and
colleagues often find solid effects on transfer outcomes (see e.g., Mayer
& Fiorella, 2022; Noetel et al., 2022, for a meta-meta-analysis). It would
be interesting to investigate possible effects of medium and prompts on
transfer performance for longer learning phases as well as with delayed
testing.

Another limitation of our study has to do with the fact that we
employed a well-designed explainer video. Our results could have
differed, had we employed less well-developed explainer videos, espe-
cially in terms of metacognitive calibration, as they might prompt
learners to overestimate their knowledge more. Our study participants
rated the explainer video as being more interesting than the text only
version (i.e. the video-script), but the mean value was still rather close to
the middle area of the scale. Hence, it is possible that the explainer video
was not interesting enough to foster overestimation. Furthermore, we
only used text material (i.e. textbook chapter) in the second learning
phase. It is possible that effects would change for other types of materials
such as video-lectures because the explainer video’s positive effect on

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

11 



learning outcomes in the prompt-condition might be partially due to a
multimedia effect (e.g., Mayer, 2022). In contrast to previous research
(e.g., Bai et al., 2022; Moos & Bonde, 2016; Zheng et al., 2023), we
decided to not include the prompts in the video but displayed them only
after the learners watched the whole explainer video. There is empirical
evidence that for longer videos it is beneficial to include prompts in the
video (e.g., Moos & Bonde, 2016; Zheng et al., 2023). Thus, it might be
interesting to investigate whether there would also be beneficial effects
for prompts on subsequent learning processes if they were included in
the explainer video.

Another limitation concerns the measure we used to assess meta-
cognitive calibration. We calculated the measure using the post-test
performance at the end of the study and the judgement of learning
(JOL) after the first learning phase. For one, the measure we used in the
study reflects a convincement regarding participants’ later performance
on a global level but was not based on the percentage of correctly
answered questions. However, the merged measure of post-test perfor-
mance and judgement of learning in the form we used reflects the de-
viation from the learners’ conviction about their performance after the
first learning phase and their actual performance (global prediction;
Baars et al., 2020). It might be interesting to use a more specific measure
for learners’ JOL in future studies. For another, we administered the JOL
measure in the prompt-condition only after the prompt questions, and
not before and after answering the questions. Therefore, we cannot
investigate a possible change in the assessment due to answering the
prompt questions within a learner. Based on our study design, we can
only compare the respective JOL assessments between conditions
(prompt vs. no-prompt). It might be interesting to take a closer look at a
potential change of the JOL assessment in relation to the performance in
prompt questions in future studies.

Another problem regarding the monitoring calibration measure re-
lates to the test expectancy of the participants. Participants did not
receive specific information on what kind of questions they had to
expect in the post-test when answering the judgement of learning items
after the first learning phase. It is therefore not entirely clear what the
participants based their judgement on. Possibly, participants in the
prompt condition expected open-ended questions based on the format of

the prompt questions. However, participants in the no-prompt condition
had no prior anchor of test items for their judgement. Accordingly, they
might have had more difficulties in judging their performance accu-
rately because they had less test information compared to participants in
the prompt condition. Previous research, however, indicates that the
knowledge about test demands alone does not automatically increase
monitoring accuracy (Eitel, 2016). Further research should consider
this, however, and give participants more information about the
post-test when asking them to judge their later performance.

Against this background, further research is needed to be able to
generalise our findings to other types of explainer videos and learning
materials.

Overall, our findings contribute to research evidence on using
explainer videos for formal learning purposes by showing that prompt-
questions can foster learning with explainer videos. In line with previ-
ous research, we also found that explainer videos can induce situational
interest, but in contrast to other research, we found no evidence for the
assumption that explainer videos foster an illusion of understanding.
Interestingly, we failed to observe that prompts exert a generally
beneficial effect –we even found that the video-script condition prompts
hampered learning. Concerning practical recommendations, it is
important to note that more time is needed when prompts are added to
the instruction, which is an issue when there are close time restrictions
(e.g., school hours). In a nutshell, our results imply that the appropriate
provision of instructional support such as prompts is often not as simple
as it might first appear.

CRediT authorship contribution statement

Marie-Christin Krebs:Writing – review& editing, Writing – original
draft, Formal analysis, Conceptualization. Katharina Braschoß: Inves-
tigation, Conceptualization. Alexander Eitel: Writing – review & edit-
ing, Writing – original draft, Formal analysis, Conceptualization.

Acknowledgements

We thank Charlotte Hild for her support in data collection.

Appendix A

Table 3
Time on Learning Task in Phase 1 and Phase 2

Condition

Explainer video Video-script

prompts no prompts prompts no prompts

n = 30 n = 46 n = 22 n = 35

M (SE) M (SE) M (SE) M (SE)

Time spent on material in phase 1 (sec)a 448.47 (29.88) 466.72 (24.95) 196.14 (35.71) 276.23 (27.66)
Time spent on prompts in phase 1 (sec)b 1042.27 (176.80) – 990.95 (211.31) –
Time spent on material in phase 2 (sec) 678.97(138.07) 694.02(96.78) 443.90(169.10) 910.58(131.65)

Note.
a Two participants were excluded based on an exploratory data analysis due to extreme processing times for learning material 1.
b Two participants were excluded based on an exploratory data analysis due to extreme times for answering the prompt-questions.

References

Ackerman, R., & Leiser, D. (2014). The effect of concrete supplements on metacognitive
regulation during learning and open-book test taking. British Journal of Educational
Psychology, 84, 329–348. https://doi.org/10.1111/bjep.12021

Adesope, O. O., & Nesbit, J. C. (2012). Verbal redundancy in multimedia learning
environments: A meta-analysis. Journal of Educational Psychology, 104, 250–263.
https://doi.org/10.1037/a0026147

Agarwal, P. K., Nunes, L. D., & Blunt, J. R. (2021). Retrieval practice Consistently
benefits student learning: A systematic review of applied research in schools and
Classrooms. Educational Psychology Review, 33, 1409–1453. https://doi.org/
10.1007/s10648-021-09595-9

M.-C. Krebs et al. Learning and Instruction 94 (2024) 101988 

12 

https://doi.org/10.1111/bjep.12021
https://doi.org/10.1037/a0026147
https://doi.org/10.1007/s10648-021-09595-9
https://doi.org/10.1007/s10648-021-09595-9


Baars, M., Wijnia, L., de Bruin, A., & Paas, F. (2020). The relation between students’
effort and monitoring judgments during learning: A meta-analysis. Educational
Psychology Review, 32, 979–1002. https://doi.org/10.1007/s10648-020-09569-3

Bai, C., Yang, J., & Tang, Y. (2022). Embedding self-explanation prompts to support
learning via instructional video. Instructional Science, 50, 681–701. https://doi.org/
10.1007/s11251-022-09587-4

Berthold, K., Nückles, M., & Renkl, A. (2007). Do learning protocols support learning
strategies and outcomes? The role of cognitive and metacognitive prompts. Learning
and Instruction, 17, 564–577. https://doi.org/10.1016/j.learninstruc.2007.09.007

Brame, C. J. (2016). Effective educational videos: Principles and Guidelines for
Maximizing student learning from video content. CBE-Life Sciences Education, 15.
https://doi.org/10.1187/cbe.16-03-0125

Brom, C., Stárková, T., & D’Mello, S. K. (2018). How effective is emotional design? A
meta-analysis on facial anthropomorphisms and pleasant colors during multimedia
learning. Educational Research Review, 25, 100–119. https://doi.org/10.1016/j.
edurev.2018.09.004

Cortina, K. S. (2008). Leistungsängstlichkeit. In W. Schneider, & M. Hasselhorn (Eds.),
Handbuch der Psychologie. Handbuch der Pädagogischen Psychologie (pp. 50–61).
Hogrefe.

Dunlosky, J., & Rawson, K. A. (2012). Overconfidence produces underachievement:
Inaccurate self evaluations undermine students’ learning and retention. Learning and
Instruction, 22, 271–280. https://doi.org/10.1016/j.learninstruc.2011.08.003

Eitel, A. (2016). How repeated studying and testing affects multimedia learning:
Evidence for adaptation to task demands. Learning and Instruction, 41, 70–84.
https://doi.org/10.1016/j.learninstruc.2015.10.003

Eitel, A., Endres, T., & Renkl, A. (2022). Specific questions during retrieval practice are
better for texts containing seductive details. Applied Cognitive Psychology, 36,
996–1008. https://doi.org/10.1002/acp.3984

Eitel, A., & Scheiter, K. (2015). Picture or text first? Explaining sequence effects when
learning with pictures and text. Educational Psychology Review, 27, 153–180. https://
doi.org/10.1007/s10648-014-9264-4

Endres, T., Weyreter, S., Renkl, A., & Eitel, A. (2020). When and why does emotional
design foster learning? Evidence for situational interest as a mediator of increased
persistence. Journal of Computer Assisted Learning, 36, 514–525. https://doi.org/
10.1111/jcal.12418

Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using
G*Power 3.1: Tests for correlation and regression analyses. Behavior Research
Methods, 41, 1149–1160. https://doi.org/10.3758/BRM.41.4.1149

Fiorella, L., Stull, A. T., Kuhlmann, S., & Mayer, R. E. (2020). Fostering generative
learning from video lessons: Benefits of instructor-generated drawings and learner-
generated explanations. Journal of Educational Psychology, 112, 895–906. https://
doi.org/10.1037/edu0000408

Hayes, A. F. (2022). Introduction to Mediation, Moderation, and Conditional Process
Analysis: A Regression-Based Approach (Third Edition). The Guilford Press.

Hidi, S. (1990). Interest and its contribution as a mental resource for learning. Review of
Educational Research, 60, 549–571. https://doi.org/10.3102/00346543060004549

Hidi, S. (2001). Interest, reading, and learning: Theoretical and practical considerations.
Educational Psychology Review, 13, 191–209. https://doi.org/10.1023/A:
1016667621114

Hidi, S., & Renninger, K. A. (2006). The four-phase model of interest Development.
Educational Psychologist, 41, 111–127. https://doi.org/10.1207/s15326985ep4102_4

Hoch, E., Fleig, K., & Scheiter, K. (2023). Can monitoring prompts help to reduce a
confidence bias when learning with multimedia? Zeitschrift für
Entwicklungspsychologie und Pädagogische Psychologie. https://doi.org/10.1026/0049-
8637/a000279

Klepsch, M., Schmitz, F., & Seufert, T. (2017). Development and Validation of two
Instruments measuring intrinsic, extraneous, and germane cognitive load. Frontiers in
Psychology, 8, 1997. https://doi.org/10.3389/fpsyg.2017.01997

Klepsch, M., & Seufert, T. (2021). Making an effort versus experiencing load. Frontiers in
Education, 6, Article 645284. https://doi.org/10.3389/feduc.2021.645284, 56.

Klingsieck, K. B. (2018). Kurz und knapp – die Kurzskala des Fragebogens
„Lernstrategien im Studium“ (LIST). Zeitschrift für Padagogische Psychologie, 32,
249–259. https://doi.org/10.1024/1010-0652/a000230

Kornell, N., Rhodes, M. G., Castel, A. D., & Tauber, S. K. (2011). The ease-of-processing
heuristic and the stability bias: Dissociating memory, memory beliefs, and memory
judgments. Psychological Science, 22, 787–794. https://doi.org/10.1177/
0956797611407929

Krämer, A., & Böhrs, S. (2017). How do Consumers evaluate explainer videos? An
empirical study on the effectiveness and Efficiency of different explainer video
formats. Journal of Education and Learning, 6, 254–266.

Kulgemeyer, C. (2020). A framework of effective science explanation videos informed by
criteria for instructional explanations. Research in Science Education, 50, 2441–2462.
https://doi.org/10.1007/s11165-018-9787-7

Kulgemeyer, C., & Wittwer, J. (2023). Misconceptions in Physics explainer videos and
the illusion of understanding: An experimental study. International Journal of Science
and Mathematics Education, 21, 417–437. https://doi.org/10.1007/s10763-022-
10265-7

List, A., & Ballenger, E. E. (2019). Comprehension across mediums: The case of text and
video. Journal of Computing in Higher Education, 31, 514–535. https://doi.org/
10.1007/s12528-018-09204-9

Liu, C., & Elms, P. (2019). Animating student engagement: The impacts of cartoon
instructional videos on learning experience. Research in Learning Technology, 27.
https://doi.org/10.25304/rlt.v27.2124

Lloyd, S. A., & Robertson, C. L. (2012). Screencast tutorials Enhance student learning of
statistics. Teaching of Psychology, 39, 67–71. https://doi.org/10.1177/
0098628311430640

Mason, L., Tarchi, C., Ronconi, A., Manzione, L., Latini, N., & Bråten, I. (2022). Do
medium and Context Matter when learning from multiple complementary Digital
texts and videos? Instructional Science, 50, 653–679. https://doi.org/10.1007/
s11251-022-09591-8

Mayer, R. E. (2021). Evidence-based principles for how to design effective instructional
videos. Journal of Applied Research in Memory and Cognition, 10, 229–240. https://
doi.org/10.1016/j.jarmac.2021.03.007

Mayer, R. E. (2022). Cognitive Theory of Multimedia Learning. In R. E. Mayer, &
L. Fiorella (Eds.), Cambridge handbooks in psychology. The Cambridge handbook of
multimedia learning (Third edition, pp. 57–72). Cambridge University Press. htt
ps://doi.org/10.1017/9781108894333.008.

Mayer, R. E., & Fiorella, L. (Eds.). (2022). Cambridge handbooks in psychology. The
Cambridge handbook of multimedia learning (3rd ed.). Cambridge University Press.
https://doi.org/10.1017/9781108894333.

Mayer, R. E., Fiorella, L., & Stull, A. (2020). Five ways to increase the effectiveness of
instructional video. Educational Technology Research & Development, 68, 837–852.
https://doi.org/10.1007/s11423-020-09749-6

Moos, D. C., & Bonde, C. (2016). Flipping the classroom: Embedding self-regulated
learning prompts in videos. Technology, Knowledge and Learning, 21, 225–242.
https://doi.org/10.1007/s10758-015-9269-1

Müller, N. M., & Seufert, T. (2018). Effects of self-regulation prompts in hypermedia
learning on learning performance and self-efficacy. Learning and Instruction, 58,
1–11. https://doi.org/10.1016/j.learninstruc.2018.04.011

Noetel, M., Griffith, S., Delaney, O., Harris, N. R., Sanders, T., Parker, P., del Pozo
Cruz, B., & Lonsdale, C. (2022). Multimedia design for learning: An overview of
reviews with meta-meta-analysis. Review of Educational Research, 92, 413–454.
https://doi.org/10.3102/00346543211052329

Noetel, M., Griffith, S., Delaney, O., Sanders, T., Parker, P., Del Pozo Cruz, B., &
Lonsdale, C. (2021). Video improves learning in higher education: A systematic
review. Review of Educational Research, 91, 204–236. https://doi.org/10.3102/
0034654321990713

Paik, E. S., & Schraw, G. (2013). Learning with animation and illusions of understanding.
Journal of Educational Psychology, 105, 278–289. https://doi.org/10.1037/a0030281

Paivio, A. (1986).Mental representations: A dual coding approach. Oxford University Press.
Pekrun, R., Vogl, E., Muis, K. R., & Sinatra, G. M. (2017). Measuring emotions during

epistemic activities: The epistemically-related emotion scales. Cognition & Emotion,
31, 1268–1276. https://doi.org/10.1080/02699931.2016.1204989

Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests
improves long-term retention. Psychological Science, 17, 249–255. https://doi.org/
10.1111/j.1467-9280.2006.01693.x

Roelle, J., Schweppe, J., Endres, T., Lachner, A., Aufschnaiter, C. von, Renkl, A., Eitel, A.,
Leutner, D., Rummer, R., Scheiter, K., & Vorholzer, A. (2022). Combining retrieval
practice and generative learning in educational contexts. Zeitschrift für
Entwicklungspsychologie und Pädagogische Psychologie, 54, 142–150. https://doi.org/
10.1026/0049-8637/a000261

Rost, D. H., Sparfeldt, J. R., & Buch, S. (Eds.). (2018). Handwörterbuch pädagogische
Psychologie (5., überarbeitete und erweiterte Auflage). Beltz.

Salomon, G. (1984). Television is “easy” and print is “tough”: The differential investment
of mental effort in learning as a function of perceptions and attributions. Journal of
Educational Psychology, 76, 647–658. https://doi.org/10.1037/0022-0663.76.4.647

Schiefele, U. (2009). Situational and individual interest. In K. R. Wentzel, & D. B. Miele
(Eds.), Handbook of motivation at school (1st ed., pp. 197–222). Routledge.

Schiefele, U., Krapp, A., Wild, K.-P., & Winteler, A. (1993). Der“ Fragebogen zum
Studieninteresse“(FSI). Diagnostica, 39, 335–351.

Schneider, W., & Hasselhorn, M. (Eds.). (2008). Handbuch der Psychologie. Handbuch der
Pädagogischen Psychologie. Hogrefe.

Schöne, C., Dickhäuser, O., Spinath, B., & Stiensmeier-Pelster, J. (2002). Skalen zur
Erfassung des schulischen Selbstkonzepts : SESSKO. Hogrefe. https://madoc.bib.un
i-mannheim.de/42724.

Senko, C., Perry, A. H., & Greiser, M. (2022). Does triggering learners’ interest make
them overconfident? Journal of Educational Psychology, 114, 482–497. https://doi.
org/10.1037/edu0000649

Serra, M. J., & Dunlosky, J. (2010). Metacomprehension judgements reflect the belief
that diagrams improve learning from text. Memory, 18, 698–711. https://doi.org/
10.1080/09658211.2010.506441

Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive load theory. Springer.
Szpunar, K. K., Jing, H. G., & Schacter, D. L. (2014). Overcoming overconfidence in

learning from video-recorded lectures: Implications of interpolated testing for online
education. Journal of Applied Research in Memory and Cognition, 3, 161–164. https://
doi.org/10.1016/j.jarmac.2014.02.001

Szpunar, K. K., Khan, N. Y., & Schacter, D. L. (2013). Interpolated memory tests reduce
mind wandering and improve learning of online lectures. Proceedings of the National
Academy of Sciences of the United States of America, 110, 6313–6317. https://doi.org/
10.1073/pnas.1221764110

Tarchi, C., & Mason, L. (2022). Learning across media in a second language. European
Journal of Psychology of Education, 38, 1593–1618. https://doi.org/10.1007/s10212-
022-00652-7

Tarchi, C., Zaccoletti, S., & Mason, L. (2021). Learning from text, video, or subtitles: A
comparative analysis. Computers & Education