Immersive Learning Predicted: Presence, Prior
Knowledge, and School Performance Inﬂuence
Learning Outcomes in Immersive Educational
Chair of School Pedagogy
University of W¨
Chair of Educational Science
University of Passau
Abstract—Media Learning is an internal process nested within
a complex combination of learner-speciﬁc and external factors.
The Educational Framework for Immersive Learning hypoth-
esizes learners’ presence, motivational traits, emotional states,
cognitive capabilities, and previous knowledge as predictors of
learning outcomes in immersive educational virtual environ-
ments. This article proposes a research model for investigating
relations between these variables that seem to be crucial for
explaining Immersive Learning processes. Three Virtual Realities
for learning three topics of Computer Science Education (com-
ponents of a computer, asymmetric cryptography, and ﬁnite state
machines), each provided on three distinct levels of technological
immersion, were used to carry out a study with 78 participants.
Path analysis was used to test the hypotheses deriving from the
research model, showing that presence, prior knowledge about
the content, and school performance inﬂuence learning outcomes.
Presence was predicted by the users’ academic emotional states
prior to the study and the provided level of immersion. The emo-
tional states were inﬂuenced by the students’ school performance.
Prior knowledge and school performance of the students were
affected by the motivational variables. This study contributes to
existing research as it adds factors that are crucial for learning
processes to the discussion on Immersive Learning.
Index Terms—Immersive Learning, Learning Outcomes, Pres-
ence, Path Analysis, Educational Framework
Although investigations of isolated variables in the process
of teaching and learning are necessary for explaining fun-
damental relations, educational characteristics show complex
dependencies. Shuell describes scholastic learning as ”[...] a
rich psychological soup of a classroom, a soup comprised of
cognitive, social, cultural, affective, emotional, motivational,
and curricular factors” [1, p. 726]. Media Learning adds even
more complexity to this interlinked structure: Characteristics
of the technology, its effects on the learner, and its embedding
in a broader teaching sequence sum up just a few of the factors
that require additional consideration. On the basis of Helmke’s
supply-use-framework for the explanation of scholastic learn-
ing , Dengel and M¨
agdefrau propose a localization of the
factors immersion as part of the teacher’s instructional supply
and presence inside the learner’s perception process, which
is inﬂuenced by motivational, cognitive, and emotional char-
acteristics of the learner. Discussions on Immersive Learning
can beneﬁt from ﬁndings in the Educational Sciences to ex-
plain how learning works in educational virtual environments
(EVEs) by investigating multiple rather than isolated factors
. This paper investigates the relation between motivational
traits, cognitive capabilities, emotional states, experience with
the learning content, presence and learning outcomes. First, we
formulate a research model on the basis of the Educational
Framework for Immersive Learning (EFiL) and formulate
hypotheses. We report a study with 78 middle school students
and describe the setting with three EVEs for Computer Science
(CS) Education topics. The ﬁndings are presented using a
path analysis approach. The discussion then connects the
hypotheses to the ﬁndings and argues the limitations of the
study and its threats to validity. The articles closes with
remarks on the contribution of the study to existing research
and implications for the design of immersive EVEs.
II. FACTORS INFLUENCING LEARNING OUTCOMES IN
EDU CATI ONAL VIRTUA L ENVIRONMENTS
Dengel and M¨
agdefrau propose the Educational Framework
for Immersive Learning (EFiL) as a localization of two main
characteristics of EVEs, immersion and presence , in the
popular supply-use framework by Helmke . They instru-
mentalize Slater’s view of immersion as a quantiﬁable descrip-
tion of technology  and presence as the resulting perception
of non-mediation . Using this distinction into an objec-
tive (immersion) and a subjective (presence) characteristic of
EVEs, immersion becomes part of the instructional supply that
is provided by a teacher, while presence is connected to the
learner’s internal perception and interpretation processes. The
instructional supply, as well as the teacher, can inﬂuence the
perceptual processes and the learning potential, consisting of978-1-7348995-0-4/20$31.00 ©2020 Immersive Learning Research Network
Fig. 1. Factors Inﬂuencing Learning Outcomes in Educational Virtual Environments, based on the Educational Framework for Immersive Learning 
several motivational, cognitive, and emotional characteristics
and the learner’s previous experience. The perceptual pro-
cesses are also inﬂuenced by the learning potential. Together,
these two factors determine the learning activities, which lead
to learning outcomes. While the EFiL includes more factors
into its perspective on Immersive Learning, this study focuses
on the main characteristics of EVEs, immersion and presence,
the variables of the learning potential, and learning outcomes.
Fig. 1 shows an overview of the variables in the EFiL (with
the user’s family characteristics included in the context factor).
Learning outcomes can relate to cognitive, affective, or
psychomotor objectives. The majority of scholastic learning
focuses the cognitive area. According to Bloom’s taxon-
omy of cognitive learning objectives, this comprises different
cognitive levels, such as remembering contents (knowledge),
understanding concepts (comprehension), or the application of
certain skills. More complex learning objectives can consist
of analyzing, synthesizing, or evaluating contents . Modern
approaches foster actional objectives regarding competencies
as ”context-speciﬁc cognitive dispositions that are acquired
and needed to successfully cope with certain situations or tasks
in speciﬁc domains” . This reorientation has led schools to
focus on evidence-based policy and practice .
The EFiL understands the factors inﬂuencing learning out-
comes as collections of various theoretical theories. The next
section helps straightening out the factors and their relations
for a research model with predictors of Immersive Learning.
A. Learning Activities
Learning activities are the direct antecedents of learning
outcomes. Modern understandings of how learning works deny
that a direct transfer of knowledge is possible. Acquiring
knowledge can be understood as a process of construction 
as well as a process of assimilation and accommodation of
existing knowledge structures . In Immersive Learning,
teaching concepts based on the idea of constructivism are ac-
tive learning , situated learning , and experience-based
learning  to facilitate such learning activities. In Helmke’s
original supply-use framework, the learning activities describe
the leaner’s active use of the supplied learning material .
While intentional learning with a focus on the active use of
learning material comprises the major part of school education,
research shows that VR can contribute to implicit learning
processes as well, e.g. .
B. Instructional Media
Regarding the instructional supply, which describes the
quality of the teaching and learning material , the in-
structional medium transports represented information for
educational purposes. Its content is designed with a certain
didactical quality. For digital learning media, especially EVEs,
the content represents the software as a whole.
The technology describes the provided level of technological
immersion. In our understanding, we follow Slater’s perspec-
tive of immersion being a quantiﬁable description of technol-
ogy rather than integrating the psychological effect on the user
. Using this distinction between technology (immersion)
and effect (presence), it is possible to distinguish between
the learning supply and the perception on the learner’s side
. Dalgarno and Lee propose the dimensions representational
ﬁdelity and learner interaction as technological characteristics
of 3D virtual learning environments (VLEs) . Following
their understanding, the representational ﬁdelity subsumes fac-
tors for the realistic, smooth, and consistent display of visual,
auditory and kinaestetic/tactile display of the environment and
its (moving) objects. The learner interaction includes embod-
ied actions and communication as well as the control of the
environment’s attributes and behavior and the construction of
objects and their behavior. There are certain characteristics that
intersect the technological and the content aspect, such as the
implementation of embodiment, locomotion and interaction
with other player/non-player avatars, as these design decisions
have technological as well as didactical dimensions.
An instructional medium is embedded in an educational
setting, such as the integration in a teaching sequence. An
instructional medium can be used as an introduction to a
new topic/problem (phase of task deﬁnition), for acquiring
knowledge that is related to the problem (phase of working
out fundamentals for solving the task) or in the phase of
task solution . In addition to embedding the medium
in the larger teaching sequence, factors such as procedure,
composition of the classroom group (individual experience,
collaborative experience), and technical support are part of the
considerations related to the setting. It has to be noted that the
setting does not necessarily frame an intentional learning pro-
cess, but can also derive from unintentional learning situations
C. Perception and Interpretation
In order to use the supplied learning material, it has to be
perceived and interpreted by the learner. In Helmke’s original
supply-use-framework, perception and interpretation of the
instructional supply are mediating processes, determining if
and which learning activities result on the learner’s side .
Presence is a subjective, psychological response to a given
VR  that can be seen as the ”perception of non-mediation”
. In an elaborated model of learning in 3-D VLEs, Dalgarno
and Lee argue that the sense of presence (as the feeling of
’being there’), co-presence (as ’being there together’), and
the construction of identity (linking the visual representation
of the user to him-/herself) are a result of the environment
characteristics representational ﬁdelity and learner interaction
presented in section B . The hypothesis that the level of
immersion is linked to the sense of presence is backed by
many research efforts in the past and present, e.g. –.
Dalgarno and Lee assume that ”the greater ﬁdelity of a 3-D
VLE leads to a greater sense of presence, and consequently,
greater transfer” [15, p. 20]. We hypothesize that
H1: The participant’s level of presence will be signiﬁ-
cantly inﬂuenced by the level of technological immersion.
The learning beneﬁts spatial knowledge representation,
experiential learning,engagement,contextual learning, and
collaborative learning, which are supported through a higher
sense of presence, can be found within the learning activities
reported in section A. The idea that presence is related to
learning has been investigated in several studies, e.g. ,
, which reported positive effects of presence on learning
outcomes. The model for this study assumes that presence
contributes to the learning activities  by offering certain
learning beneﬁts to the learner . Our second hypothesis is
H2: The learning outcomes will be signiﬁcantly inﬂu-
enced by the learner’s level of presence.
Perception and interpretation are connected to several other
factors that are part of the learning potential . The assump-
tion of presence being a part of these mediation processes
 implies an interesting network of factors inﬂuencing Im-
mersive Learning, as presence and learning activities are both
inﬂuenced by factors of the learning potential, while presence
mediates the effect of the medium on the learning activities. In
an extension of Dalgarno and Lee’s work, Fowler emphasizes
the pedagogical immersion as ”a complex interaction of dif-
ferent pedagogical variables” [21, p. 416] that inﬂuences the
perception and interpretation of a VLE.
D. Learning Potential
Factors subsumed under the learning potential moderate
the effect of how the instructional supply is perceived and
interpreted. Variables afﬁliated with the learning potential
are the learner’s motivational traits, cognitive capabilities,
emotional states, and previous experiences .
The emotional state of the learner is a crucial inﬂuence on
learning that is connected to the learner’s motives, activities,
and outcomes. Emotions that have an effect on learning
(achievement emotions) can be distinguished according to
their degree of activation (activating or deactivating) and their
valence (being pleasant/positive or unpleasant/negative) .
A virtual simulation training study from Fraser et al. revealed
correlations between the learners’ reported emotions and their
perceived cognitive load, which inﬂuences the learning out-
comes . Research shows that emotions are connected to
the sense of presence as well, see , . Following these
ideas, it can be assumed that emotions affect the cognitive
resources related to the learner’s perception and interpretation
(in our terms: presence), which, in turn, inﬂuence learning
H3: The learner’s presence will be signiﬁcantly inﬂu-
enced by (a) positive emotions and (b) negative emotions.
Cognitive factors are some of the most important predictors
for learning outcomes: In his meta study, Hattie names various
cognitive variables that predict learning like the cognitive
development (d=1.28), prior achievement (d=.67), and con-
centration/persistence/engagement (d=.48) . There is evi-
dence that cognitive capabilities shown as school performance
inﬂuence success in adult life as well , . Regarding
the effect of cognition on presence, the feeling of ’being
there’ is induced by (1) the mental construction of one’s own
bodily actions as interaction possibilities within the virtual
world while (2) suppressing incompatible sensory input from
the physical reality, resulting in a suspension of disbelief
. Thus, the feeling of ’being there’ is accompanied by
cognitive processes leading from stimuli to presence . To
measure cognitive capabilities, this study subsumes previous
school performance in the subjects German (the learners’
native language) and Maths, because a correlation between
intelligence and scholastic performance can be assumed .
together with previous knowledge about the EVEs’ topics
as indicators for cognitive beneﬁts. While this simpliﬁcation
cannot represent the effect of cognitive factors on presence (as
school grades do not measure the cognitive processes relevant
for inducing presence), hypotheses 3, 4, and 5 model the
effect of cognition on learning outcomes (through learning
H4: The post-test performance will be signiﬁcantly inﬂu-
enced by the scholastic performance in (a) German and
(b) Maths, and by (c) the pre-test performance.
The control-value theory proposed for the emotional factors
depicts a relation between emotional states and cognitive
resources . As the emotional state of the user will be
assessed with a focus on the prospective outcome emotions,
Fig. 2. Proposed Research Model With the Factors Intrinsic Motivation (MOTin), External Regulation (MOTex), Pre-Test Performance (PERpre), Scholastic
Performance in German (SPGer), Scholastic Performance in Maths (SPMat ), Post-Test Performance (PERpost), Positive Emotions (EMOpo), Negative Emotions
(EMOne), Presence (PRS), and Immersion (IMM)
consisting of emotions that are related to expected positively
valued success or expected negatively valued failure in the
future , it is assumed that these emotions are predicted by
the cognitive factors related to scholastic performance.
H5: Positive emotions will be signiﬁcantly inﬂuenced by
the scholastic performance in (a) German and (b) Maths.
H6: Negative emotions will be signiﬁcantly inﬂuenced by
the scholastic performance in (a) German and (b) Maths.
Motivational constructs symbolize the underlying ”why”
of learning (Vallerand 2010). With an average effect size of
d=1.28, they are an important predictor of learning outcomes
. Deci and Ryan’s self-determination theory distinguishes
intrinsic motivation (experiencing pleasure and satisfaction
from an activity itself) and extrinsic motivation (performance
of an activity because of some sort of separable outcome)
. Extrinsic motivation can be either externally regulated
(where rewards and constraints motivate the person), intro-
jected (internalization of past contingencies), or identiﬁed
(perception of extrinsic motives as chosen by oneself). In a
two-dimensional taxonomy, Bartle identiﬁes different reasons
of why people play games . According to their tendencies
to either be motivated by the game’s players or its world
and to focus their behavior on rather acting or interacting,
Bartle distinguishes the player types achievers (accomplish
game-related goals), explorers (explore the game), socializers
(interact with fellow players), and killers (impose themselves
on others). A player’s motivation is a central criterion for the
design of immersive EVEs. The engagement induced through a
motivating virtual learning environment can enhance learning
outcomes and presence . The intrinsic motivation to know
refers to pleasure and satisfaction gained through learning,
exploring, and understanding something new , such as the
academic outcomes described by Hattie. Bartle’s player types
and Pirker’s Player Type Design  derive from an under-
standing of the intrinsic motivation towards accomplishments
and the intrinsic motivation to experience stimulation . We
focus on the motivation related to learning and formulate three
hypothesis concerning the effects of intrinsic motivation as
the most and external regulation as the least self-determinated
dimensions of motivation on the cognitive factors:
H7: The previous scholastic performance in German will
be signiﬁcantly inﬂuenced by (a) intrinsic motivation and
(b) external regulation.
H8: The previous scholastic performance in Maths will
be signiﬁcantly inﬂuenced by (a) intrinsic motivation and
(b) external regulation.
H9: Pre-test performance will be signiﬁcantly inﬂuenced
by (a) intrinsic motivation and (b) external regulation.
E. Teacher and Context
The teacher(s) have a strong impact on the learning poten-
tial. The characteristics of the teacher subsume all personal,
professional, pedagogical factors contributing to or inhibiting
learning. Primary factors are professional, didactical, and
diagnostic expertise, classroom management competencies,
personal values, goals, expectancies, and self efﬁcacy .
Context variables integrate all factors deriving from the
learner’s family, his/her cultural, regional, educational and
classroom environment, the didactical context as well as
the school and classroom atmosphere. Family characteristics
include structural factors such as social class, language, and
the parents’ education. In the original supply-use framework
 and the EFiL , the family is proposed as a separate
factor, independently inﬂuencing the learning potential.
The teacher(s) and the context are crucial factors when
it comes to evaluating learning. To focus the effect of the
medium on learning outcomes and its correlates and moderat-
ing/mediating effects, those factors are simpliﬁed as external
factors and not depicted in the research model (Fig. 2).
Fig. 3. (a) Bill’s Computer Workshop (b) Fluxi’s Cryptic Potions (c) Pengu’s Treasure Hunt (d) A Student During the Study
III. MET HO D
A. Sample and Procedure
78 Austrian middle school students (36 female, 4 missing
values) from eight to ninth grade took part in the one-time
study. The students’ mean age was 13.95 ±0.74 years.
Most students had no prior experience with VR technology
(44.90%), some of them tried it once (29.50%), while only few
have used VR multiple times (29.50%) and none of them used
VR frequently. The students achieved moderate performances
in the subjects German (M= 2.42, SD = 0.96) and Maths
(M= 2.51, SD = 0.91).
Three days before the study, the students ﬁlled out a
questionnaire assessing demographic data, their motivation
towards learning CS, and the pre-tests. At the beginning of
the study, they ﬁlled out an emotion questionnaire. There were
three environments on three different devices, each student
experienced every EVE on one device (covering all levels of
immersion: one on the laptop, one on the Mobile VR, and
one on the HTC Vive), see Fig. 3d. After each experience
(about 15 minutes, respectively, at the students own pace),
they ﬁlled out a presence questionnaire and a post-test for the
EVE (same as the pre-test). The students collected stamps for
each VR experience and questionnaire to track their progress.
In the end, the investigator explained the connection of the
games’ analogies to Computer Science topics.
The EVEs used for the study focused on different topics
from Computer Science (components of a computer, asymmet-
ric cryptography, and ﬁnite state machine) and were provided
each on three different levels of immersion (laptop: 16-inch
monitor with Intel HD Graphics 5500, Mobile VR: Daydream
View with Moto Z smartphone, HTC Vive generation 1).
Bill’s Computer Workshop (components of a computer):
The user shrinks him-/herself and enters a broke-down com-
puter to repair it. Bill, a mechanic inside the computer, assigns
certain tasks to the player (ﬁnding/reconnecting components),
see Fig. 3a. Every time, the student repairs a component,
he/she gets more information about its function and how
it is connected to the other parts of the computer. Fluxi’s
Cryptic Potions (asymmetric cryptography): The user ﬁnds
him-/herself in a medieval love story and encounters Fluxi,
a personal carrier dragon. By writing letters and encrypting
or decrypting them with magic public or private potions
(analogies for public-private key encryption, see Fig. 3b), the
player can decide to win either the heart of Prince Charming,
Princess Isolde, or the dragon’s aunt Gertrud, with the help
of his/her friend Nikolai. The aim to ask the beloved for a
dance at a ball gets interfered by a Man-in-the-Middle, Sir
Dance-A-Lot. Pengu’s Treasure Hunt (ﬁnite state machines):
On the quest to ﬁnd the treasure of the fearless penguin
Fig. 4. Path Analysis of the Research Model Showing Effects and Relations Between Intrinsic Motivation (MOTin), External Regulation (MOTex), Pre-
Test Performance (PERpre), Scholastic Performance in German (SPGer), Scholastic Performance in Maths (SPMat ), Post-Test Performance (PERpost), Positive
Emotions (EMOpo), Negative Emotions (EMOne), Presence (PRS), and Immersion (IMM)
pirate Pengu, the player travels various islands (the machines
states) by using boats (representing the transition functions)
and gradually builds up a map of which boat leads to which
island (Fig. 3c). The student gets a new part of the key needed
to open up the next treasure chest (the ﬁnite state) for each
newly discovered route on the map. In the end, the treasure
map depicts a ﬁnite state machine of the island world (with
three different worlds to be discovered). The game is based
on the CS Unplugged activity Treasure Hunt .
Multiple questionnaires were used to assess the different
constructs; the students self-reported their grades in German
and Maths if their parents allowed it and if they were willing
to impart them to the investigators (anonymized).
The students’ motivation towards learning Computer Sci-
ence (the topic of the EVEs) was assessed using the Hanfstingl
et al. questionnaire for testing students’ motivation towards
learning in a certain subject , adapted for learning in
Computer Science (the questionnaire was evaluated in blinded
for review). 12 rating items on a ﬁve-point Likert scale were
used to assess intrinsic, identiﬁed, and introjected motivation,
as well as external regulation. The scales for intrinsic motiva-
tion (M= 3.10, SD = 1.02, α= .85) and external regulation
(M= 2.70, SD = 1.02, α= .65) were used for this study. The
external regulation scale’s reliability was slightly low.
The emotional state of the participants was assessed via
an emotion questionnaire from Titz . The students self-
reported ten basic emotions concerning positive emotions (en-
joyment, hope, pride, relief, security) and negative emotions
(shame, anger, hopelessness, anxiety, boredom) on a six-point
Likert scale. The emotions security and fear were excluded for
the analysis due to better scale reliabilities for both positive
emotions (M= 2.91, SD = 0.98, α= .73) and negative
emotions (M= 0.69, SD = 0.68, α= .68). Still, the negative
emotions’ scale reliability was slightly low.
Presence was measured using the translated version of the
Slater-Usoh-Steed (SUS) questionnaire  assessing phys-
ical presence with six questions on a seven-point Likert
scale. The students ﬁlled out one SUS questionnaire for
each of the three environments: Bill’s Computer Workshop
(M= 4.41, SD = 1.49, α= .87), Fluxi’s Cryptic Potions
(M= 4.14, SD = 1.14, α= .91), and Pengu’s Treasure Hunt
(M= 3.96, SD = 1.56, α= .92).
The students’ performance was measured before and af-
ter each game on different cognitive levels with a separate
performance test for each topic: components of a computer
(knowledge, e.g. Name the function of the working memory.),
asymmetric cryptography (understanding, e.g. Describe why
only Bob can read Alice’s message when she encrypts her
message with Bob’s public key.), and ﬁnite state machines
(application, e.g. Complete the map so that you can only get
to Treasure Island if you take three ’apple boats’ in a row.).
For asymmetric cryptography and ﬁnite state machines, the
analogies of the potions (public-/private-keys) and the islands
and boats (states and transition functions) were maintained to
make the topics more accessible to the age group. The results
of the pre- and the post-test were z-standardised to make the
learning gains on the different topics comparable.
1) Test of the Proposed Model: A path analysis was
performed on MPlus with the 234 datasets using maximum
likelihood estimation to test the ﬁt between the research model
displayed in Fig. 4 and the data obtained. The resulting path
model is displayed in Fig. 4. The model ﬁt indices for the x2
test (p= .45), the comparative ﬁt index (CFI = .99), the root
mean square error of approximation (RMSEA = .005) index,
and the standardised root mean residual (SRMR = .043) value
satisﬁed the recommended level of acceptable ﬁt according to
, indicating a good ﬁt of the proposed research model.
2) Hypotheses Testing: Fig. 4 shows the resulting path
coefﬁcients. Not all hypotheses were supported by the data.
There were correlations within the learning potential con-
structs between the motivational variables (r= -.40, p <.001),
the previous scholastic performances (r= .51, p <.001), and
the emotional variables (r= -.33, p <.001). The ﬁndings
show that presence was signiﬁcantly inﬂuenced by the level
of technological immersion (β= .49, p <.001), support-
ing H1. Presence was found to be a signﬁcant inﬂuence
on post-test performance (β= .16, p <.01), supporting
H2. Positive emotions were signiﬁcantly inﬂuencing presence
(β= .15, p <.05), while negative emotions were not (β= -
.05, p >.05), supporting H3a, but not H3b. The post-test
performance was signiﬁcantly inﬂuenced by cognitive factors.
Pre-test performance (β= .34, p <.001) and the previous
scholastic performance in Maths (β= -.29, p <.001) had a
positive effect on post-test performance, the performance in
German (β= .26, p <.001) had a negative effect on post-test
performance, supporting H4a, H4b, and H4c. Positive emotions
were found to be predicted by both scholastic performances,
German (β= .37, p <.001) and Maths (β= -.26, p <.001),
supporting H4aand H4b. Negative emotions were found to
be predicted by the scholastic performance in German (β= -
.27, p <.001), but not Maths (β= .07, p >.05). This supports
H6a, but not H6b. Scholastic performance in German was sig-
niﬁcantly inﬂuenced by external regulation (β= .23, p <.01),
but not by intrinsic motivation (β= .12, p >.05), supporting
H7b, but not H7a. There were no signiﬁcant effects on the
scholastic performance in Maths from intrinsic motivation
(β= -.08, p >.05) and extrinsic motivation (β= .12, p >.05).
Thus, H8aand H8bwere not supported by the data. Pre-test
performance was signiﬁcantly affected by intrinsic motivation
(β= .16, p <.05) and external regulation (β= -.23, p <.01),
supporting H9aand H9b.
Seven endogenous variables were investigated in the study.
Variance explanation was low for the cognitive factors pre-
test performance (R2= .09) and scholastic performance in
German (R2= .05) and Maths (R2= .03) as well as for the
emotional factors positive emotions (R2= .10) and negative
emotions (R2= .06). Moderate variance explanation was found
for presence (R2= .28) and post-test performance (R2= .23).
D. Discussion and Conclusion
1) Contribution to Theory: This study investigated the pre-
dictors of learning outcomes using path analysis on the basis
of the EFiL. The ﬁndings show that the learning outcomes
can be predicted by the learners’ previous knowledge on
the topic, their previous scholastic performance, and their
sense of presence within the EVE. Some cognitive factors are
inﬂuenced by the learners’ motivational traits and affect their
emotional states. The positive emotions of the learner and the
provided level of technological immersion inﬂuence his/her
sense of presence. These results conform to the theoretical
assumptions deriving from the EFiL and support other ideas
of learning in 3D VLEs, such as Dalgarno and Lee  and
Fowler  that are based on the premise that learning with
technology is mediated through the perception of the learn-
ers and moderated by their individual characteristics. While
there are existing approaches that include person-speciﬁc and
technological factors to explain presence, , and learning
outcomes, e.g. , this is the ﬁrst study known to the
authors that investigates multiple person-speciﬁc factors that
are relevant to learning and connects them with technological
characteristics of virtual environments. Results that the authors
want to highlight concern the effects of the learners’ scholastic
performance in their native language, German, on the post-
test performance and on their emotional states. These effects
were signiﬁcant, but oppose to the hypothesized directions. It
is assumed that this could result from gender differences, but
this relation has to be investigated further.
2) Contribution to Practice: Nesting technological charac-
teristics and their perceptual correlates within existing ideas
on how learning works has implications for using VEs (in
especial immersive VEs) in practice as well. Teachers who
want to use VEs as instructional media need to consider
•how a suitable VE can be embedded in the selected
teaching sequence and to which phase of the learning
progress it can contribute,
•which hardware and software can be used to support the
intended learning beneﬁts, and
•how the VE interacts with the learners’ person-speciﬁc
characteristics and how this affects their learning.
Regarding the design of EVEs, the same considerations
inﬂuence the development process:
•presence has an effect on learning outcomes, which is
why an EVE should be designed in a way so that it
induces a sense of presence;
•the level of immersion has a strong impact on the user’s
sense of presence, which is why technological consider-
ations play a crucial role in the design of EVEs;
•emotional characteristics can have an impact on the user’s
sense of presence, which is why EVEs’ emotional cues
can contribute to the corresponding learning processes.
3) Limitations of the Study: The EFiL proposes a broad
overview of theories. To carry out studies investigating the
relationships between the factors, several constraints and as-
sumptions have to be made. We assessed motivation as a situ-
ational motivation towards learning the topics of the EVEs in
general and the emotional state directly before the experiences.
There are more nuances in motivational factors that are related
to presence and those in the emotional factors that are related
to learning. Another simpliﬁcation made for this study was the
decision to subsume the learner’s cognitive factors under the
pre-test performance and the previous scholastic performances
in the core subjects German and Maths only.
Some scale reliability values were questionable. When car-
rying out more studies with the developed EVEs, the pre-
and post-tests have to be revised according to the problems
that occured during this study. The level of immersion was
measured as an ordinale variable which allows a ranking,
but is not sufﬁcient for an adequate scaling. Future studies
could integrate a more detailed approach to quantify the
technological quality of different devices. Also, the outcomes
of one-time experiences might differ in various ways from a
long-term use of immersive media in the classroom.
As a concluding remark: Explaining and exploring Im-
mersive Learning can beneﬁt from multiple perspectives on
the technology, the content, the teacher(s), the learner’s per-
ception, and his/her characteristics as well as the setting
and external factors. An interdisciplinary approach including
the Educational Sciences, Game Design, Human-Computer-
Interaction, and Psychology can contribute to further research.
 T. J. Shuell, “Teaching and learning in the classroom,” in International
Encyclopedia of the Social & Behavioral Sciences. Elsevier, 2001, pp.
15 468–15 472.
 A. Helmke, Unterrichtsqualit¨
at und Lehrerprofessionalit¨
Evaluation und Verbesserung des Unterrichts ; [ber¨
Hattie-Studien], 5th ed., ser. Schule weiterentwickeln, Unterricht
verbessern : Orientierungsband. Seelze-Velber: Klett Kallmeyer, 2014.
 A. Dengel and J. M¨
agdefrau, “Immersive learning explored: Subjective
and objective factors inﬂuencing learning outcomes in immersive educa-
tional virtual environments,” in Proceedings of 2018 IEEE International
Conference on Teaching, Assessment, and Learning for Engineering
(TALE), M. J. W. Lee, Ed. [Piscataway, New Jersey]: [IEEE], 2018,
 M. Slater, “A note on presence terminology,” Presence Connect, vol. 3,
 M. Lombard and T. Ditton, “At the heart of it all: The concept of
presence,” Journal of Computer-Mediated Communication, vol. 3, no. 2,
p. 0, 1997.
 B. S. Bloom, Taxonomy of educational objectives: The classiﬁcation
of educational goals, [ﬁrst edition] ed. New York, New York: David
McKay Company, 1968.
 K. Koeppen, J. Hartig, E. Klieme, and D. Leutner, “Current issues
in competence modeling and assessment,” Zeitschrift f¨
/ Journal of Psychology, vol. 216, no. 2, pp. 61–73, 2008. [Online].
 D. Leutner, J. Fleischer, J. Gr¨
unkorn, and E. Klieme, Competence
Assessment in Education: Research, Models and Instruments, ser.
Methodology of Educational Measurement and Assessment. Springer
International Publishing, 2017.
 M. P. Huang and N. E. Alessi, “Presence as an emotional experience,”
Studies in health technology and informatics, vol. 62, pp. 148–153, 1999.
 J. Piaget and M. Cook, The origins of intelligence in children. New
York: W W Norton & Co, 1952.
 C. Dede, “Introduction to virtual reality in education,” Themes in Science
and Technology Education, vol. 2, no. 1-2, pp. 7–9, 2010.
 H. Mandl, H. Gruber, and A. Renkl, “Situiertes lernen in multimedialen
lernumgebungen,” in Information und Lernen mit Multimedia und
Internet : Lehrbuch f¨
ur Studium und Praxis. 3. Auﬂ., L. J. Issing
and P. Klimsa, Eds. Weinheim: Beltz, Psychologie-Verlags-Union,
2002, pp. 138–148, 1.u.2. Auﬂ. des Buches u.d. Buchtitel: Information
und Lernen mit Multimedia. [Online]. Available: https://epub.uni-
 C. Schott and S. Marshall, “Virtual reality and situated experiential
education: A conceptualization and exploratory trial,” Journal of
Computer Assisted Learning, vol. 34, no. 6, pp. 843–852, 2018. [Online].
 M. Slater, “Implicit learning through embodiment in immersive vir-
tual reality,” in Virtual, augmented, and mixed realities in education,
ser. Education, D. Liu, Ed. Singapore: Springer, 2017, pp. 19–33.
 B. Dalgarno and M. J. W. Lee, “What are the learning affordances of
3-d virtual environments?” British Journal of Educational Technology,
vol. 41, no. 1, pp. 10–32, 2010.
 G. Tulodziecki, S. Grafe, and B. Herzig, Medienbildung in Schule und
Unterricht: Grundlagen und Beispiele. UTB GmbH, 2019. [Online].
 J.-W. Lin, H. Duh, D. E. Parker, H. Abi-Rached, and T. A. Furness,
“Effects of ﬁeld of view on presence, enjoyment, memory, and simulator
sickness in a virtual environment,” in Proceedings IEEE Virtual Reality
2002. IEEE Comput. Soc, 2002, pp. 164–171.
 T. A. Mikropoulos, “Presence: A unique characteristic in educational
virtual environments,” Virtual Reality, vol. 10, no. 3-4, pp. 197–206,
 M. Chow, “Factors inﬂuencing presence in virtual worlds,” NI 2012 :
11th International Congress on Nursing Informatics, June 23-27, 2012,
Montreal, Canada. International Congress in Nursing Informatics (11th
: 2012 : Montreal, Quebec), vol. 2012, p. 083, 2012.
 E. A.-L. Lee, K. W. Wong, and C. C. Fung, “How does desktop virtual
reality enhance learning outcomes? a structural equation modeling
approach,” Computers & Education, vol. 55, no. 4, pp. 1424–1442, 2010.
 C. Fowler, “Virtual reality and learning: Where is the pedagogy?” British
Journal of Educational Technology, vol. 46, no. 2, pp. 412–422, 2015.
 R. Pekrun, T. Goetz, W. Titz, and R. P. Perry, “Academic emotions
in students’ self-regulated learning and achievement: A program of
qualitative and quantitative research,” Educational Psychologist, vol. 37,
no. 2, pp. 91–105, 2002.
 K. Fraser, I. Ma, E. Teteris, H. Baxter, B. Wright, and K. McLaughlin,
“Emotion, cognitive load and learning outcomes during simulation
training,” Medical education, vol. 46, no. 11, pp. 1055–1062, 2012.
 J. Diemer, G. W. Alpers, H. M. Peperkorn, Y. Shiban, and
uhlberger, “The impact of perception and presence on emotional re-
actions: A review of research in virtual reality,” Frontiers in psychology,
vol. 6, p. 26, 2015.
 J. A. Hattie, Visible learning: A synthesis of over 800 meta-analyses
relating to achievement. London: New York : Routledge, 2008.
 G. E. Samson, M. E. Graue, T. Weinstein, and H. J. Walberg, “Academic
and occupational performance: A quantitative synthesis,” American
Educational Research Journal, vol. 21, no. 2, pp. 311–321, 1984.
 P. L. Roth, C. A. BeVier, F. Switzer, and J. S. Schippmann, “Meta-
analyzing the relationship between grades and job performance,” Journal
of Applied Psychology, vol. 81, pp. 548–556, 1996.
 M. Slater and M. Usoh, “Representations systems, perceptual position,
and presence in immersive virtual environments,” Presence: Teleopera-
tors & Virtual Environments, vol. 2, pp. 221–233, 1993.
 T. Schubert, F. Friedmann, and H. Regenbrecht, “The experience of
presence: Factor analytic insights,” Presence: Teleoperators and Virtual
Environments, vol. 10, no. 3, pp. 266–281, 2001.
 J. A. Hattie and B. C. Hansford, “Self measures and achievement:
Comparing a traditional review of literature with a meta-analysis,”
Australian Journal of Education, vol. 26, no. 1, pp. 71–75, 1982.
 R. Pekrun, A. Frenzel, T. G¨
otz, and R. Perry, “The control-value
theory of achievement emotions: An integrative approach to emotions
in education,” Publ. in: Emotion in education / ed. by Paul A. Schutz
and Reinhard Pekrun. Amsterdam : Academic Press, 2007, pp. 13-36,
 R. Ryan and E. Deci, “Self-determination theory and the facilitation of
intrinsic motivation, social development, and well-being,” The American
psychologist, vol. 55, pp. 68–78, 2000.
 R. Bartle, “Hearts, clubs, diamonds, spades: players who suit muds,”
Journal of MUD research, vol. 1, no. 1, 1996.
 J. Pirker, “Immersive and engaging forms of virtual learning,” Graz,
 R. J. Vallerand, L. G. Pelletier, M. R. Blais, N. M. Briere, C. Senecal, and
E. F. Vallieres, “The academic motivation scale: A measure of intrinsic,
extrinsic, and amotivation in education,” Educational and Psychological
Measurement, vol. 52, no. 4, pp. 1003–1017, 2016.
 csunplugged.org, “The book,” 2019. [Online]. Available:
 B. Hanfstingl, T. Almut, I. Andreitz, and F. H. M¨
uler- und lehrerbefragung 2008/09. interner arbeitsbericht.”
 W. Titz, Emotionen von Studierenden in Lernsituationen: Explorative
Analysen und Entwicklung von Selbstberichtskalen: Zugl.: Regensburg,
Univ., Diss., 2000, ser. Internationale Hochschulschriften. M¨
unchen: Waxmann, 2001, vol. 367.
 M. Slater, M. Usoh, and A. Steed, “Depth of presence in virtual en-
vironments,” Presence: Teleoperators and Virtual Environments, vol. 3,
no. 2, pp. 130–144, 1994.
 R. Kline, Principles And Practice Of Structural Equation Modeling, 01