ORIGINAL RESEARCH

published: 31 August 2021

doi: 10.3389/fpsyg.2021.703857

Validation of Cognitive Load During

Inquiry-Based Learning With
Multimedia Scaffolds Using
Subjective Measurement and Eye
Movements
Marit Kastaun1*, Monique Meier1, Stefan Küchemann2 and Jochen Kuhn2
1
Department of Biology Education, Institute for Biology, Universität Kassel, Kassel, Germany, 2 Physics Education Research
Group, Department of Physics, Technische Universität Kaiserslautern, Kaiserslautern, Germany

Edited by: Subject-method barriers and cognitive load (CL) of students have a particular importance
Moritz Krell, in the complex learning process of scientific inquiry. In this work, we investigate the valid
Freie Universität Berlin, Germany
measurement of CL as well as different scaffolds to reduce it during experimentation.
Reviewed by:
Andreas Korbach,
Specifically, we examine the validity of a subjective measurement instrument to assess
Saarland University, Germany CL [in extraneous cognitive load (ECL), intrinsic cognitive load, and germane cognitive
Paul Ayres,
load (GCL)] during the use of multimedia scaffolds in the planning phase of the scientific
University of New South Wales,
Australia inquiry process based on a theoretical framework of the CL theory. The validity is analyzed
Kun Yu, by investigating possible relationships between causal (e.g., cognitive abilities) and
University of Technology Sydney,
Australia
assessment (e.g., eye-tracking metrics) factors in relation to the obtained test scores of
*Correspondence:
the adapted subjective measurement instrument. The study aims to elucidate possible
Marit Kastaun relationships of causal factors that have not yet been adequately investigated in relation
[email protected]
to CL. Furthermore, a possible, still inconclusive convergence between subjective test
Specialty section:
scores on CL and objectively measured indicators will be tested using different eye-tracking
This article was submitted to metrics. In two studies (n = 250), 9th and 11th grade students experimentally investigated
Educational Psychology,
a biological phenomenon. At the beginning of the planning phase, students selected one
a section of the journal
Frontiers in Psychology of four multimedia scaffolds using a tablet (Study I: n = 181) or a computer with a stationary
Received: 30 April 2021 eye-tracking device (Study II: n = 69). The subjective cognitive load was measured via
Accepted: 31 July 2021 self-reports using a standardized questionnaire. Additionally, we recorded students’ gaze
Published: 31 August 2021
data during learning with the scaffolds as objective measurements. Besides the causal
Citation:
Kastaun M, Meier M,
factors of cognitive-visual and verbal abilities, reading skills and spatial abilities were
Küchemann S and Kuhn J (2021) quantified using established test instruments and the learners indicated their representation
Validation of Cognitive Load During
preference by selecting the scaffolds. The results show that CL decreases substantially
Inquiry-Based Learning With
Multimedia Scaffolds Using with higher grade level. Regarding the causal factors, we observed that cognitive-visual
Subjective Measurement and Eye and verbal abilities have a significant influence on the ECL and GCL in contrast to reading
Movements.
Front. Psychol. 12:703857.
skills. Additionally, there is a correlation between the representation preference and different
doi: 10.3389/fpsyg.2021.703857 types of CL. Concerning the objective measurement data, we found that the absolute

Frontiers in Psychology | www.frontiersin.org 1 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

fixation number is predictive for the ECL. The results are discussed in the context of the
overall methodological research goal and the theoretical framework of CL.
Keywords: cognitive load, cognitive abilities, representation preference, scaffolding, eye tracking, scientific
inquiry

INTRODUCTION objective measures as indicators of CL, such as heart rate

(Paas and van Merriënboer, 1994; Minkley et al., 2021), pupil
Cognitive load (CL) is a theoretical, psychological construct dilation (Chen and Epps, 2014; Huh et al., 2019), blink rate
that describes the individual loads of learners during the (Chen and Epps, 2014), or gaze behavior (Korbach et al., 2018;
processing, construction, and memorizing of (new) information Zu et al., 2020), different approaches to measuring CL have
(Sweller et al., 2019). Within the Cognitive Load Theory (CLT), been investigated. Possible relationships and convergences
one of many basic assumptions is that only a limited number between subjective and objective measurement tools for
of information elements can be processed simultaneously due identifying CL are also coming more into the focus of research
to the limited working memory capacity (Paas et al., 2010). (among others Minkley et al., 2021). However, the use of
Therefore, the processing of information that is relevant to subjective measurement instruments by self-assessment is still
learning should be optimized and loads irrelevant to learning an established way to assess CL based on a learning task or
are to be minimized (among others van Merriënboer et al., instructional learning approach (Cierniak et al., 2009b; Leppink
2006; Paas and van Merriënboer, 2020). Another assumption et al., 2013; Klepsch et al., 2017; Krell, 2017; Zu et al., 2020;
claims the processing of information in two different channels Thees et al., 2021). Although the subjective measurement
– the verbal and the visual channel (dual channel assumption; approach is controversial in research regarding valid measurement
Skuballa et al., 2019; Sweller et al., 2019; Scheiter et al., 2020). of CL (de Jong, 2010; Kirschner et al., 2011; Sweller et al.,
Furthermore, it is assumed that CL is composed of three 2011), studies, such as the one by Klepsch et al. (2017),
distinct categories – intrinsic (ICL), extraneous (ECL), and demonstrate a proven measurement accuracy. Other studies
germane cognitive load (GCL; among others Sweller, 1994; Zu also reveal that measuring the three types of CL can pose
et al., 2020). ICL is the cognitive load associated with high different difficulties (DeLeeuw and Mayer, 2008; Ayres, 2018).
or low element interactivity during the processing of information/ First, it is unclear how, for example, the different CLs (ICL,
learning material (Sweller et al., 2011, 2019). Accordingly, the GCL, and ECL) can be affected by the wording of items, since
degree of interactivity between essential information elements variations of the items lead to different results in performance
required for learning determines the ICL. If the element and CL (Sweller et al., 2011; Klepsch and Seufert, 2021). Second,
interactivity is low, only a few elements are processed by the besides the wording, it is uncertain whether learners are able
learner in working memory at the same time. If the element to assess their competences in relation to the differentiated
interactivity increases, the load on working memory also rises items (ICL; GCL; and ECL; Sweller et al., 2011). Nevertheless,
(Sweller et al., 2011). In contrast, ECL includes the cognitive approaches to measure the three types of CL via self-assessments,
load generated by suboptimal instructional design. Consequently, e.g., via the design of controlled conditions (among others
the learner needs to spend more cognitive resources on irrelevant Sweller et al., 2011; Klepsch and Seufert, 2021), suggest that
or poorly constructed information during task performance a categorical differentiation and measurement of the individual
or information processing than on the actual task solution CLs are possible. However, Thees et al. (2021) also showed
(Sweller et al., 2019). The ECL and ICL together determine that relying on established subjective instruments to measure
the total CL in the working memory by the learner. They are CL may lead to various difficulties. Composing a subjective
additively related because the resources for processing come measurement instrument from established items could lead to
from the same working memory pool (Sweller et al., 2011). a more valid measurement of the three types of cognitive load
The third category of cognitive load is GCL. This is expended (ECL, GCL, and ICL). However, these established self-assessments
by the learner to construct schemas and form mental models have mostly been used with university students, whereas research
(Paas et al., 2003; Sweller et al., 2019), and it is therefore on subjective instruments measuring the different types of CL
often described as the load that generates learning. In recent for middle-school students in instructional teaching-learning
years, the study of CL has become increasingly important. To settings has been scarce (e.g., van de Weijer-Bergsma and van
test and consequently optimize instructional designs or as a der Ven, 2021). Therefore, the focus of this research is to
control variable for, e.g., self-regulated learning processes, such design and test a self-assessment questionnaire to measure the
as inquiry-based learning (see “Environment (E) & Task (T) three types of CL for 9th and 11th grade students. To examine
as causal factor – scaffolding & inquiry learning”; Kaiser et al., a possible valid measurement in combination with a convergence
2018), the valid measurement of CL represents a central goal between subjective and objective measurement instruments of
of educational and psychological research (among others CL, various methods from validity research can be used.
Kirschner et al., 2006; Minkley et al., 2018; Sweller et al., Consistent with the Standards for Educational and Psychological
2019). Using subjective self-assessments (Cierniak et al., 2009b; Testing (AERA et al., 2014), more recent validity approaches
Leppink et al., 2013; Klepsch et al., 2017; Krell, 2017) or open possible opportunities for proving evidence and convergence

Frontiers in Psychology | www.frontiersin.org 2 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

via argumentative, theory-based test score interpretations prior knowledge and task complexity or task success (expertise-
(Messick, 1995; Kane, 2006, 2013). According to Kane (2013, reversal effect: Cierniak et al., 2009a; Kalyuga, 2013).
p. 13), validity is “an integrated evaluation judgment of the Mental load, mental effort, and task performance are the
degree to which empirical evidence and theoretical rationales assessment factors by which individual CL can be judged in
support the adequacy and appropriateness of inferences and relation to, for example, mastering a task or instruction. Both
actions based on test scores or other modes of assessment”. mental load and mental effort generally describe the CL that
Based on the validation framework of Cronbach and Meehl must be exerted by the learner to complete a given task. Mental
(1955), the Argument-Based Approach to Validation follows load is the term used to summarize the subject-independent
the examination of test score interpretation. This consists of CL that relates solely to the characteristics of the task (Paas
making different validity assumptions from an interpretation- and van Merriënboer, 1994; Choi et al., 2014). Therefore, mental
usage argument and testing them using different methods as load is equivalent to the construct of ICL (e.g., Minkley et al.,
well as relating them back to the existing theories (Kane, 2013). 2021). In contrast, mental effort describes the loads “which
Hence, validity investigations do not represent a mere testing refers to the amount of capacity or resources that is actually
of formalized theories. They rather show the establishment allocated by the learner to accommodate the task demands”
and testing of coherent and plausible chains of reasoning that (Choi et al., 2014, p. 228). Mental effort therefore refers to
examine and explain the results of a measurement instrument the cognitive resources that are used during the specific problem
in a theory- and evidence-based manner. Within these chains solving/task and can accordingly be equated with GCL and
of reasoning, other validity criteria should be integrated and ECL (e.g., Paas and van Merriënboer, 1994; Choi et al., 2014;
the test score should be examined from different perspectives Krell, 2017). Regardless of the classification of the individual
(Messick, 1995; Kane, 2013). Regarding the validity testing of types of CL, these are related to the causal factors that can
a subjective instrument to measure the three types of CL in influence both task performance and the CL of the learner.
this study (ICL, ECL, and GCL), the theory-based analysis of The level of prior knowledge (L), for example, has an influence
possible influences on CL and interactions within the learning on the expression of the ICL (see “Prior Knowledge”), whereas
environment needs to be investigated (causal factors: the ECL is considered to be mainly dependent on the instructional
environment, task, and learner). The identification of possible design (T). Both types of loads are perceived by the learner
relationships between these causal factors and the occurring as rather passive (Klepsch and Seufert, 2021). The GCL, on
ICL, ECL, and GCL (assessment factors) when using multimedia the other hand, represents the perceived by the learner as
scaffolds is therefore required first. Then, a possible convergence rather active load (Seufert, 2018). The assessment factors can
between the results of the subjective instrument and objective be measured by subjective (self-assessment) and objective (for
indicators of CL will be examined in more detail. Therefore, example, heart rate) instruments as an indicator for CL (see
the theoretical assumptions between the causal and the assessment “Introduction” and Figure 1). Subjective instruments, such as
factors will be presented first in order to deduct assumptions the ones of Zu et al. (2020); Klepsch et al. (2017); Krell (2017);
with regard to the validation of the measurement instrument or Leppink et al. (2013), have the benefit that they can be easily
that will be tested within the performed studies. used via paper-pencil questionnaires within the intervention.
However, contrary to objective measures, such as Heart Rate
or Gaze Behavior, they depend on the self-assessment
competences of learners/participants who assess and report
CONSTRUCT OF COGNITIVE LOAD their own load based on the item formulations (see
WITH CAUSAL AND ASSESSMENT “Introduction”). The objective instruments provide another way
FACTORS to measure CL. Although they are technical and costly, objective
measurement methods are increasingly used. Although a valid
The distinction between causal and assessment factors in relation measurement of CL cannot yet be granted, various studies
to CL was described by Paas and van Merriënboer (1994) show close correlations between subjective and objective
based on study results in the context of problem solving. Choi measurement of CL categories (Korbach et al., 2018; Solhjoo
et al. (2014) extended this theory-based construct of CL et al., 2019; Minkley et al., 2021). Eye-movement measurements
(Figure 1). In addition to the influencing factors of the task represent one research approach to investigate CL during
(task: e.g., complexity and format) and learner-specific learning with visual (and auditory) material or problem solving.
characteristics (learner: e.g., cognitive abilities or prior Thus, different metrics, such as fixations duration (Zu et al.,
knowledge), the model was extended by the factor of learning 2020) or transitions (Korbach et al., 2018), are used to derive
environment. Like the task as well as the learner characteristics, learners’ CL. Previous work on the relationship of students’
the environment can influence the load that occurs, e.g., physical gaze data and their subjective CL ratings demonstrated that
conditions. The causal factors are in close interaction with certain eye-tracking metrics discriminated between the three
each other which can significantly influence the expression of types of CL (Zu et al., 2020). Zu et al. (2020) found that
individual CL (Choi et al., 2014; Paas and van Merriënboer, there is a significant relation of the mean fixation duration
2020). The interaction between the causal factors can be seen, of students with low prior knowledge in a physics context
for example, between the task and the characteristics of the and the ECL. Whereas the ICL significantly relates to the
learner. Thus, different studies show the close interaction among transitions between an animation and a text, the GCL is linked

Frontiers in Psychology | www.frontiersin.org 3 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

FIGURE 1 | Construct of cognitive load with for the study-specific causal and assessment factors (adapted from Choi et al., 2014).

to the dwell time on an animation. These results in the context inquiry learning: Bell et al., 2005), which is usually self-regulated
of electric motors were observed for students with low prior and cooperative, has higher learning effects and can contribute
knowledge and they were independent on the working better to long-term learning of scientific skills than direct forms
memory capacity. of instruction (Alfieri et al., 2011; Furtak et al., 2012). However,
other studies also suggest that learning along the scientific
inquiry process can cause student-specific problems, as the
Environment (E) & Task (T) as Causal complexity and openness of the learning process exceed the
Factor – Scaffolding & Inquiry Learning individual capacity of working memory (Leutner et al., 2005;
As shown in Figure 1, the learning environment and task are Kirschner et al., 2006). As a result, supporting methods, such
causal factors affecting CL. As Choi et al. (2014) noted, these as protocol sheets provided with tasks, or the use of scaffolds,
two factors usually cannot be clearly separated in (terms of) such as prompts and feedback, can minimize these occurring
practical implementation which also becomes clear by looking loads (Hmelo-Silver et al., 2007) and support the learner in
at the learning environment and the task in our study. In his or her inquiry-based learning process (guided inquiry
science education, inquiry-based learning is an approach to learning: Bell et al., 2005). Studies show, among other things,
promote scientific reasoning as an integral part of science that short, direct prompts do not hinder the constructivist
education (e.g., OECD, 2007). Within inquiry-based learning, learning method of inquiry-based learning. However, depending
learners are asked to actively apply different facets of knowledge on the design and the use, they can minimize individual CLs
to investigate a scientific phenomenon along the different phases (e.g., Kirschner et al., 2006). Based on this, different approaches
of the scientific knowledge process (Kind and Osborne, 2017; to scaffolding have been explored empirically (e.g., Arnold
de Jong, 2019). Through the generation of a research question et al., 2014; Kaiser and Mayer, 2019; Meier and Kastaun, 2021).
and testable hypotheses, the planning of a scientific investigation Using digital media, scaffolds can enable individual learning
to the execution and interpretation of the experimentally in a variety of forms.
obtained results, learners actively use their content-related In addition to prior knowledge and motivational factors,
knowledge as well as their methodological skills (inquiry skills/ learners bring in further characteristics that offer starting points
scientific reasoning skills) to investigate the phenomenon/ for differentiation. Especially when dealing with multimedia-
problem. Studies show that this problem-solving process (open based representation combinations (e.g., video or animation),

Frontiers in Psychology | www.frontiersin.org 4 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

cognitive characteristics, such as learning preferences for visual very detailed worked-out examples can minimize the learning
or verbal learning material (representation preference; Mayer outcome for students with a high level of prior knowledge,
and Massa, 2003), spatial ability (Höffler and Leutner, 2011), as their own individual knowledge construction is inhibited
or cognitive style (Höffler and Schwartz, 2011; Koć-Januchta by the content specifications, whereas learners with a low level
et al., 2019), can have an impact on the learning process. of prior knowledge can particularly benefit from these scaffolds
Regarding the construction of multimedia scaffolds, these may (expertise-reversal effect: Kalyuga, 2013; Chen et al., 2016). In
not only be composed of descriptive and depictive representations the context of inquiry-based learning, studies show that novices
(Ainsworth, 2006), but also of verbal audio tracks or symbolic benefit less from individual knowledge construction than experts
texts in combination with static and dynamic images (among because they are missing prior experience or concrete information
others Nitz et al., 2014). For a targeted, effective as well as about the individual knowledge facets (e.g., Shrager and Siegler,
preferably individualized use of technology-enhanced scaffolds, 1998; Siegler, 2005; Kendeou and van den Broek, 2007). Reasons
multimedia design principles based on CLT and the theory for this could be associated with high levels of CL and a
of multimedia learning (Mayer, 2014) should be resorted to consequent reduction of GCL, because the resources of the
minimize ECL and create free capacities in the working memory. working memory for processing the learning content are exceeded
Therefore, the use of multimedia scaffolds in the complex (element interactivity, see “Introduction”). If the complexity of
problem-solving process of inquiry-based learning (Figure 1: the learning task increases, explicit instruction is beneficial
E + T) along with learner characteristics as causal factors of for long-term retention (Chen et al., 2016, 2018).
CL should be elicited.
Cognitive Abilities
Along with the investigation of instructional forms and its
Learners’ (L) Characteristics as Causal pros and cons for specific learners, different correlations
Factor regarding CL can also be identified (e.g., Mayer and Massa,
Learner-specific characteristics are elementary in the context
2003). In addition to prior knowledge, spatial ability is
of CL (Figure 1) and thus also in the long-term memorization
increasingly a focus of investigation (e.g., Quaiser-Pohl et al.,
of new knowledge elements (among others Choi et al., 2014).
2001). Basically, a positive correlation can be found between
Besides motivational and affective factors, interest (Baars et al.,
the expression of spatial ability and learning performance
2017) or fun (van de Weijer-Bergsma and van der Ven, 2021),
(Höffler, 2010). Regarding dynamic and static representations,
the influence of cognitive characteristics, such as prior knowledge
such as animations or static images, spatial ability can act
(Kalyuga et al., 2003; Cierniak et al., 2009a; Kalyuga, 2013),
like a compensator according to the ability-as-compensator
in relation to CL has been widely investigated. Interactions
hypothesis (e.g., Höffler, 2010; Münzer, 2012; Sanchez and
with the theory of multimedia learning as well as distinction
Wiley, 2014), but only if this ability is also actively needed
of further cognitive characteristics aiming at cognitive abilities
for processing the learning task (Höffler and Leutner, 2011).
in relation to representation-based information processes, such
Hence, it can be assumed that the higher spatial ability the
as the development of reading skills, spatial ability, or cognitive-
better the learning performance when dealing with static
verbal and visual abilities, were examined so far only little in
images, whereas when learning with animated images, lower
relation to CL or with unclear results (among others Ho et al.,
spatial ability may contribute to deeper understanding (e.g.,
2014; Chen et al., 2018; Lehmann and Seufert, 2020). However,
Huk, 2006). Further determinants of how individual learning
based on theoretical frameworks of multimedia learning (Moreno,
abilities are influenced can be identified using Jäger’s (1984)
2010; Mayer, 2014), human cognitive architecture (Paas and
intelligence model. The recognition of patterns, the agreement
van Merriënboer, 2020), and empirical studies of individual
of similarities, or other linguistic as well as visual contexts
cognitive abilities, assumptions can be made that identify a
allow conclusions to be drawn about logical comprehension
relationship between specific cognitive, representation-based,
as well as verbal and visual thinking skills. A few studies
abilities, the learning material, and the resulting CL. In the
show that here, the expression is also related to learning
following, specific (cognitive) characteristics are identified and
performance (Kuhn et al., 1988; Kaiser and Mayer, 2019).
elicited in terms of a possible causal factor in connection with
Along with verbal reasoning skills, reading skills can have
CL and their possible of measurement within the present study.
an impact on individual learning performance, especially for
text-based instructional materials (Plass and Homer, 2002;
Prior Knowledge Plass et al., 2003; Jäger et al., 2017). In connection with
Studies investigating prior knowledge have shown that it the preference for specific, material-related instructional
significantly influences learning performance (Kalyuga, 2013; formats (see “Learning preference/Representation Preference”),
Chen et al., 2018; Richter et al., 2018; Seufert, 2019). Based reading skills can also have an influence on the processing
on the cognitive architecture of knowledge processing and of the task or the provided material (Peterson et al., 2015;
memorization, increased prior knowledge contributes to chunking Lehmann and Seufert, 2020). On the one hand, reading skills
(new) information and minimizing element interaction in the are also closely related to cognitive (verbal) skills. On the
working memory (Pollock et al., 2002; Kalyuga, 2011). However, other hand, the expression of reading skills can have an
this is only correct up to a certain level of prior knowledge influence on information processing especially when learning
and dependent on the design of learning materials. For example, with text-based (monomodal) materials (Schneider et al., 2017).

Frontiers in Psychology | www.frontiersin.org 5 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

Learning Preference/Representation Preference indicators of CL? To investigate the research question, we made
Preferences with respect to a specific instructional material, modality, different validation assumptions that are differentiated via
or representation fundamentally describe a person’s preferences hypotheses and empirically tested in two studies. However,
to do or use something in a certain way than in another way only assumptions and hypotheses are listed where we expect
(Kürschner et al., 2005; Lehmann and Seufert, 2020). The optional, significant correlations, influences, and effects in respect to
student-specific selection of a learning material according to an the CL derived from the theory.
individual’s preference, may result in positive effects in terms of,
for example, motivation to complete the task (in reference to the Assumption 1) The subjective measurement has an internal consistency
self-regulation: Rheinberg et al., 2000; Baars et al., 2017). Mayer and separates the relevant subscales – ECL, GCL, and ICL.
and Massa (2003) describe learning preference in (multimedia)
H1.1: By adapting the items from previously tested instruments
learning situations as, e.g., visual or verbal learners showing a
of CL (Cierniak et al., 2009b; Leppink et al., 2013; Klepsch
preference for a certain multimedia representation/instruction
et al., 2017), the three subscales ECL, GCL, and ICL can
(representation preference). So, in specific learning situations,
be identified both verbally and by statistical characteristics
learners often prefer one multimedia representation for learning
(internal structure and internal consistency).
instruction (e.g., image and audio vs. image-text) over others
(Mayer and Massa, 2003; Choi and Sadar, 2011). In the study H1.2: The CL subscales show positive and negative correlations
of Lehmann and Seufert (2020), it was investigated whether the to each other. The ECL scale correlates negatively with the
preference of a modality has an influence on learning performance GCL; the ICL scale correlates positively with the GCL (Leppink
and CL. Although this study did not examine multimedia learning et al., 2013; Klepsch et al., 2017).
materials, but only used monomodal text as visual or auditory
stimuli, the results suggest a possible relationship between learning Assumption 2) The subjective instrument leads to different cognitive
preference and CL. Their findings provide insights into the load levels for the respective students in grades 9 and 11.
relationships between the (learning-) preference, the use of the H2.1: Higher prior knowledge, which students should have
respective instructional material (only modality), and the expression in grade 11, should result in lower CL than that of grade 9
of CL. Thus, learners who associated their preferred modality students (Kalyuga, 2013; Chen et al., 2018; Richter et al., 2018;
with visual texts benefited and reported a lower ECL than those Seufert, 2019).
who did not use a modality matching their indicated preference
(Lehmann and Seufert, 2020). Assumption 3) Students’ specific scores in the causal factors –
cognitive-verbal and visual abilities, reading skills, spatial ability,
and representation preference affect the individual CL (ECL,
Aims of the Study, Research Question, GCL, and ICL).
Assumptions, & Hypotheses
The aim of this work is to validate, along the Argument-Based H3.1: There is a correlation between cognitive-verbal ability
Approach to Validation by Kane (2013) and Standards for scales and CL, especially for text-based (monomodal) scaffolds
Educational and Psychological Testing (AERA et al., 2014), a (Kuhn et al., 1988; Mayer and Massa, 2003; Lehmann and
subjective instrument for students of the 9th and 11th grade for Seufert, 2020).
measuring the three different types of CL (ECL, ICL, and GCL):
H3.2: Cognitive-visual ability correlates negatively with ECL
(a) Based on the theoretical construct according to Choi et al.
and positively with GCL (Jäger, 1984; Jäger et al., 2017).
(2014), relationships between causal and assessment factors can
be identified, which need to be tested in order to validate a H3.3: For the selected text-based (monomodal) scaffolds,
subjective measurement instrument for the three different types reading-skill scores correlate negatively with ECL scores and
of CL (Figure 1). In particular, the assessment of cognitive load positively with GCL scores (Plass et al., 2003; Lehmann and
resulting from instructional design (ECL; multimedia scaffolds) Seufert, 2020).
and its interactions/relationships with causal factors (representation
H3.4: Spatial ability may have an impact on CL when using
preference; cognitive abilities) is of great interest for the present study.
static scaffolds (static image). It is possible that a high spatial
(b) A possible convergence between the subjective
ability minimizes the ECL and contributes to the increase of
measurement of CL with an objective instrument (eye-tracking
the GCL (ability-as-compensator hypothesis; Höffler, 2010;
metrics) will be checked. Both measures the load during the
Münzer, 2012).
use of multimedia scaffolds and the related task for the
planning phase of the inquiry process. Furthermore, a deeper H3.5: The individual representation preference will condition
investigation of possible correlations and influences of causal the choice of a multimedia digital scaffold. The choice and
factors (cognitive-verbal and visual abilities, reading skills, the resulting possible fit between representation preference and
and representation preference) through different analysis steps use of the scaffold will be reflected in a low average ECL
will be proofed. (Kürschner et al., 2005; Lehmann and Seufert, 2020).
According to this, the following overall, methodical research
question will be investigated: Do subjective ratings of cognitive Assumption 4) As an objective instrument, visual attention in terms
load (ECL, ICL, and GCL) correlate with other measures or of specific eye-tracking metrics (total fixation count; mean fixation

Frontiers in Psychology | www.frontiersin.org 6 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

duration; and total fixation count) shows a high convergence with day, small groups of three students were first randomly assigned.
the levels of subjectively measured cognitive load (ECL, ICL, and GCL). Following this, the general procedure of the laboratory day
was explained with the participants. In addition, the individual
H4.1: There is a significant correlation between the total
scaffolds (static image-text = image-text; static image-
fixation count and the level of CL. The total fixation count
audio = image-audio, moving image-text = animation; and moving
predicts the ECL (Zu et al., 2020).
image-audio = video) were introduced by content-remote examples
H4.2: The mean fixation duration and total fixation duration since the subsequent selection of the preferred scaffold was
are significantly related to the level of CL. The mean fixation made independently by the students. After examining the
duration and total fixation duration predict the ECL (Zu et al., 2020). phenomena “A mushroom in the pizza dough” in a video jointly
in the classroom/laboratory (study 1) or in groups (study 2),
the independent, experimental work along the inquiry-based
MATERIALS AND METHODS learning process within the small groups followed. After observing
the phenomenon, the students were first invited to independently
General Design of the Study I and II pose a research question (for example: What influence does
To test the subjective measurement of CL, two studies were temperature have on the activity of yeast?). Afterward, the groups
conducted within FLOX, a teaching-learning laboratory of the formulated two well-founded hypotheses in line with the research
University of Kassel. In both studies, the intervention, question, which they would like to test in their experiment.
measurement times, and tasks were identical. Thus, students The hypothesis formulation was followed by the planning phase
experimented in small groups (three students) along the scientific of the experiment, which is also the intervention phase. At
inquiry process on a selected topic of metabolic physiology the beginning of the intervention, the students were initially
(see “Environment (E) & Task (T) as causal factor – scaffolding asked to continue working individually. Along their protocol,
& inquiry learning”; “Procedure”). The experiment as well as in which the students documented all results of the
the experimental components, the scaffolds, the instructional experimentation phases, they were first instructed to individually
material, and group size did not differ in both sub-studies. select one of the four provided scaffolds. In study I, they viewed
However, the difference was in the playback device of the them on a tablet. In study 2, they used a computer with a
scaffolds and in the time duration of the experiment (see stationary eye-tracking system (Tobii-Pro X3-120; 120 HZ; <0.4°;
“Procedure”; “Scaffolds & Task”). Whereas in study I, the 22-inch screen). Before each eye-tracking recording, a short
scaffolds were provided to the participants via a tablet, in technical briefing and an individual 9-point calibration were
study II, the students watched their scaffold on a computer. performed. To optimize the eye-tracking recording and to
In contrast to study I, which was performed with an entire minimize any unnecessary distractions of the participants, some
class (environment), study II could only be conducted with preparations were made, such as the light controlling of the
three students at a time (in a similar total duration of the eye-tracking areas and the separation of working areas (Kastaun
laboratory day in study I), and therefore, the experiment et al., 2020; Kastaun and Meier, 2021). Directly after using the
designed by the students was not performed in study II. The multimedia scaffolds, a subjective measurement of individual
total duration of the laboratory day was 3–4 h. CL was integrated into their protocol. After completing the
CL test, students had the task to operationalize the variables
for their experiment before going into the group work again.
Participants Then, they developed a complete plan in the group to test one
Participants in both studies (n = 250, 53.2% female) were 9th (n = 142)
of their hypotheses. Afterward, the groups conducted their
and 11th (n = 108) grade students from schools in Kassel, Germany
experiment together (study I) or used their planning anticipated
(see Table 1 for details). Divided into the two sub-studies, study
videos that showed the experimental implementation and
I involved 181 students (9th grade: n = 111; 11th grade: n = 70),
matching value tables (study II). After that, the students interpreted
whereas the second study employed a total of 69 students (9th
the results in reference to the research question and hypotheses.
grade: n = 31; 11th grade: n = 38). Both student groups and their
legal guardians were given written and verbal information about
the study. Based on this, both parties had to give their written
Scaffolds & Task
consent to participate in the study. This was voluntary and all
In both studies, the students were required to choose one option
student data during the intervention were recorded anonymously.
from the same list of multimedia scaffolds (Figure 2). These
differed in the combination and modality of their respective
Procedure individual representations. Image-text and animation represent
One week before each experimental-laboratory day, a the monomodal scaffolds, whereas image-audio and video can
pre-instruction including a pre-assessment was conducted within be summarized as multimodal scaffolds with spoken text. The
regular school hours. At this first measurement point, the design of the graphics and the content of all scaffolds are
students’ demographic data, their cognitive abilities (Cognitive identical and comparable according to cognitive-psychological
Ability Test: Heller and Perleth, 2000), and reading skills (Reading principles (among others Mayer, 2014). The scaffolds have about
Speed and Reading Comprehension: Schneider et al., 2017) the same time duration. The length is between 2.5 and 3 min.
were assessed in a written form (Table 2). On the laboratory The content consisted of written and spoken text as well as

Frontiers in Psychology | www.frontiersin.org 7 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

TABLE 1 | Demographic data from the studies.

Total sample (Study I + II) Study I Study II

M SD n M SD n M SD n

Age, years 15.82 1.10 250 15.77 1.13 181 15.95 1.05 69
Sex, % female 53.2 250 48.1 181 66.7 69
Grade 9th 142 111 31
11th 108 70 38

TABLE 2 | Descriptive data of used measurement of assessment and causal factors for the total sample.

M SD Cronbachs α Items per Scale References

ASSESSMENT FACTORS
Extraneous cognitive load
(ECL) 2.17 1.01 0.891 5 Klepsch et al., 2017;
Germane cognitive load
4.04 1.06 0.814 3 Cierniak et al., 2009b;
(GCL)
Intrinsic cognitive load
3.50 1.34 0.696 2 Leppink et al., 2013
(ICL)
CAUSAL FACTORS
Cognitive abilities
Cognitive-visual ability
13.93 8.88 0.895 24 Heller and Perleth, 2000
(Figure classification, N01)
Cognitive-visual ability
12.92 4.22 0.830 24
(Figure Analogy, N02)
Spatial ability (Paper
7.37 3.31 0.790 14
Folding, N03)
Cognitive-verbal ability
7.5 2.6 0.700 19
(Word Analogy, V03)
Reading skills
Reading comprehension 46.75 26.26 0.82 47
Reading speed 41.62 28.36 0.80 Schneider et al., 2017
Reading accuracy 64.72 25.32 0.82 47

graphical representations to summarize the methodological skills/ et al., 2017; see Supplementary Table A). With the aim to
knowledge for the planning phase. Especially, the concept of measure the different load categories, ECL, GCL, and ICL,
variable operationalization and the creation of measurement during learning with multimedia scaffolds, the paper-based
series and associated measurement concept are focused without questionnaire initially consisted of 15 different items (ECL = 7
reference to any specific domain content (Meier and Kastaun, items; GCL = 4 items; and ICL = 4 items). These items were
2021). After the use of a scaffold and the cognitive load composed and adapted in such a way that they can
measurement, the students were asked to identify the dependent be independently reflected upon and assessed by 9th and 11th
and independent variables (see “Environment (E) & Task (T) grade students and have a primary relationship to the scaffolds.
as causal factor – scaffolding & inquiry learning”). To do so, In addition, the subjective instrument is primarily intended
the students needed to use their prior knowledge as well as to assess the CL based on the multimedia scaffolds. Therefore,
the information provided in the scaffold and transfer it to the the items were linguistically adapted and composed in this
specific subject context of the experiment. The use of the scaffold manner. Using a 6-point Likert scale from strongly disagree
as well as the identification of the dependent and independent (1) to strongly agree (6), the students’ self-reports were recorded
variables constitutes the task of the environment (Figure 1). in relation to the respective items. After the survey, we performed
a Principal Component Analysis of the obtained datasets (using
Varimax rotation) and the CL scales were formed by merging
Instruments & Methods discrimination indices. The Cronbach’s α-values were determined
Different instruments and evaluation procedures were used to to verify the internal consistency.
test the validation hypotheses. Table 2 lists the individual test
instruments and associated scales.
Reading Speed and Reading Comprehension Test
Cognitive Load The Reading Speed and Reading Comprehension Test (Schneider
The subjective instrument for measuring individual CL was et al., 2017) is an established test instrument in school
adapted and composed of different items from validated test diagnostics and is used for the differentiated determination
instruments (Cierniak et al., 2009b; Leppink et al., 2013; Klepsch of reading speed, reading accuracy, and reading comprehension

Frontiers in Psychology | www.frontiersin.org 8 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

FIGURE 2 | Image sections of the scaffolds in the different multimedia representation combinations.

from grades 5 to 12. It is a speed-power test in which a Representation Preference

given text must be read as quickly and accurately as possible Representation preference was generated by an independent
within a fixed time (6 min). In the text, items are integrated and free choice of the scaffold. The evaluation is generated
at individual points, which must be used to complete the via a frequency analyses of the individual usage behavior.
sentence by selecting the correct word. By analyzing the
correctly crossed words, the reading accuracy across the text
as well as the reading comprehension can be evaluated by Gaze Behavior
a point system. To record reading speed, the number of all Within study II (n = 69), students’ gaze data were recorded
words read within the prescribed time was counted and using a stationary eye-tracking system. Using the software
evaluated using an evaluation scale. Tobii-Pro Studio, all students with a poor calibration were
eliminated from the sample. In this work, we analyzed the
learning stimuli as a whole, i.e., there was a single AOI
Cognitive Ability Test that covered all information provided in the learning material.
The Cognitive Ability Test (Heller and Perleth, 2000) is also Afterward, the total fixation count, mean fixation
an established test instrument in school diagnostics. It is used duration, and total fixation duration were extracted using
for the differential measurement of cognitive ability dimensions the software. Since the scaffolds have different characteristics
that are particularly relevant for learning at school. Scales of in information retrieval and processing, the data were first
cognitive-verbal (V03) and visual abilities (N01, N02) as well z-standardized.
as spatial ability (N03) were used from this. The individual
instruments primarily aim to measure differential content-based,
processing capacity. The Cognitive Ability Test is also a speed- Data Analysis
power test. Within a prescribed time (8 min), participants had In the following analysis, we excluded datasets that showed
to solve the individual test items of the respective scales. The missing values or too large measurement inaccuracies (outliers
visual, series tasks (N01), visual, and verbal relation tasks (N02 and dropout: n = 23) or calibration errors (eye-tracking
& V03) as well as paper-folding tasks measuring the spatial datasets, n = 4; Gaze Sample under 65%) in the main
ability (N03: paper-folding test) were analyzed by an evaluation intervention. Measurement data from the Reading Speed
scale (correct/wrong), and the expressions of the individual and Reading Comprehension Test and the Cognitive Ability
constructs were identified by the formation of total scores. Test may have different samples to the overall or sub-study

Frontiers in Psychology | www.frontiersin.org 9 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

as these datasets are partially incomplete. Nevertheless, these and differences (t-test) between representation preference and
subjects were included in the main analysis regarding the expression of representation preference were investigated.
representation preference and CL since the complete datasets The quantitative data from the second study were used to test
were available. The IBM SPSS Statistic 27 software was used the fourth validity assumption. For this purpose, the
for the statistical analyses. z-standardized eye-tracking metrics (total fixation count; mean
To investigate the validity assumptions and research fixation duration; and total fixation count) were correlated with
hypotheses, correlation and regression analyses as well as the individual expressions of CL.
methods of inferential statistics (t-test and ANOVA) were
applied for both studies. First, the internal consistency was
examined in more detail by factor analytic procedures and
RESULTS
correlation analyses of the individual scales of the CL test
(validity assumption 1). To test the second validity assumption, Validity Assumption I
different t-tests were conducted to analyze the assumed differences We performed a Principal Component Analysis (Varimax
in the types of cognitive load (ECL; GCL; and ICL) between rotation) to extract the most important independent factors
the two grade levels (9th and 11th grade). In addition, group (KMO = 0.751; the Bartlett Test was highly significant; p < 0.001).
differences (ANOVA) between the selected scaffolds Kaiser’s criteria and the scree-plot yielded empirical justification
(representation preference) and the expressions of the ECLs, for retaining three components (value ≥1) which accounted
GCLs, and ICLs were analyzed for each grade in order to for 65.76% of the total variances. From these dimension-related
exclude possible variance due to the different multimedia items, the individual scales of CL were composed and reduced
representations. Along the third validity assumption, correlations from an initial set of 15 items to 10 (ECL = 5 items, GCL =3
with the individual scales of the Reading Speed and Reading items, and ICL = 2 items; see Supplementary Table A). Items
Comprehension Test as well as the Cognitive Ability Test in were excluded that either had too low factor loadings (<0.3)
relation to the individual expressions of CL were conducted. or loaded too high on another factor that was not integrated
Multiple regression analyses were used to further examine these in the construct. These resulting 10 items and three scales
correlations to identify predictors of CL during the use of the were checked for internal consistency using Cronbach’s α
various scaffolds. No violations of the predictors were identified (Table 2) and correlation analyses of the individual subscales
for any of the regressions. The Durbin-Watson statistics for (Table 3). The examination of the internal consistency by the
all regressions ranged from 1 to 3 and no autocorrelations Cronbach’s α-values of the individual subscales and the total
between variables were present (VIF < 10). Furthermore, possible scale for the CL shows satisfactory characteristic values (Table 2).
correlations (multiple regressions with categorical variables) As expected, statistically significant correlations between the
ECL and GCL (p < 0.001) and ECL and ICL (p < 0.001) were
TABLE 3 | Parallel-test reliability via Pearson-correlations (n = 250).
found. However, the scales of the GCL and ICL did not show
a statistically significant relationship (p = 0.61; Table 3).
GCL ICL

ECL r −0.38 0.23 Validity Assumption II

p 0.00 0.00
To test the second validity assumption stating that CL decreases
GCL r 0.03
p 0.61 in progressive, school-based education due to an increase in
methodological skills, group comparisons (n = 250) were made
p < 0.05. regarding the CL between 9th and 11th grade students.

TABLE 4 | Differences between 9th and 11th grade in cognitive load (t-test, n = 250).

Grade 9th (n = 142) Grade 11th (n = 108)

M SD M SD

ECL 2.41 1.06 1.87 0.87 (95% − CI[0.29,

0.78])/t(248) = 4.26,
p < 0.001, d = 0.552,
r = 0.26
GCL 3.90 1.07 4.22 1.03 (95% − CI[−0.57,
−0.04])/t(248) = −2.27,
p = 0.024, d = −0.295,
r = 0.14
ICL 3.94 1.27 2.92 1.22 (95% − CI[0.70,
1.33])/t(248) = 6.36,
p < 0.001, d = 0.822,
r = 0.38

Frontiers in Psychology | www.frontiersin.org 10 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

TABLE 5 | Pearson correlation between causal and assessment factors for the total sample.

Cognitive-verbal Cognitive-visual abilities Spatial ability Reading Reading speed Reading

ability (V03) (N01/02) (N03) comprehension accuracy

ECL r −0.167 0.066 −0.085 −0.055 −0.112 −0.111 0.023

p 0.013 0.323 0.202 0.409 0.143 0.148 0.766
GCL r 0.143 −0.111 −0.001 0.067 0.163 0.190 0.044
p 0.034 0.095 0.985 0.320 0.034 0.013 0.571
ICL r −0.062 0.118 −0.125 −0.082 −0.107 −0.125 −0.025
p 0.359 0.079 0.063 0.226 0.164 0.102 0.744

p < 0.05.

and the representation preference (choice of scaffold). The

TABLE 6 | Multiple regression with ECL and cognitive abilities (n = 222). results show that only the ECL (F(3,3) = 3.359, p = 0.021,
h 2p = 0.068) in the 9th grade differed significantly between the
Coefficient B SE(B) BETA P VIF
learners who chose an image-text combination (n = 37; M = 2.82;
(Constant) 2.460 0.300 SD = 1.08) and those who, in contrast, chose a video as a
V03 −0.068 0.028 −0.172 0.016 1.155 scaffold (n = 56; M = 2.13; SD = 1.04). Furthermore, there were
N03 −0.007 0.025 −0.024 0.765 1.446 significant differences within the 11th grade with respect to
N02 −0.020 0.020 −0.083 0.310 1.515
the GCL (F(3,3) = 3.187, p = 0.027, h 2p = 0.085) between the
N01 0.040 0.020 0.152 0.046 1.315
groups of animations (n = 29; M = 4.56; SD = 0.86) and picture
R2 = 0.047, Durbin-Watson-Statistic = 1.870. p < 0.05. text (n = 38; M = 3.86; SD = 0.97). For the ICL and between the
other groups regarding the ECL and GCL, no significant
differences could be identified (see Supplementary Material).
TABLE 7 | Multiple regression with GCL and cognitive abilities (n = 222).

Coefficient B SE(B) BETA P VIF

Validity Assumption III
(Constant) 4.081 0.314 To test the third validity assumption, correlation and regression
V03 0.069 0.029 0.168 0.019 1.155 analyses were conducted and evaluated for the total sample
N03 0.033 0.026 0.101 0.203 1.446
(n = 250; Tables 5–11) and the subsamples (study I: n = 181;
N02 −0.005 0.021 −0.021 0.798 1.515
N01 −0.052 0.021 −0.189 0.013 1.315 study II: n = 69; see Supplementary Material). Negative
correlations between cognitive-verbal abilities (V03) and the
R2 = 0.052, Durbin-Watson-Statistic = 1.677. p < 0.05. expressions of the ECL (r = 0.17; Table 5) as well as positive
ones with the GCL (r = 0.14, Table 5) were found. Moreover,
TABLE 8 | Multiple regression with ICL und cognitive abilities (n = 222).
showing a small correlation coefficient, the expressions of reading
speed (r = 0.19) and reading comprehension (r = 0.16) are positively
Coefficient B SE(B) BETA P VIF related to the scale of the GCL’s. Regarding the cognitive-visual
and spatial scales (N01, N02, and N03) as well as the reading
(Constant) 3.471 0.397
accuracy scale, no other statistically significant correlations could
V03 −0.024 0.037 −0.046 0.516 1.144
N03 −0.033 0.032 −0.080 0.314 1.435
be identified (Table 5). Correlation analyses within the sample
N02 −0.052 0.026 −0.162 0.048 1.509 of the first study (n = 181, see Supplementary Material) further
N01 0.079 0.026 0.228 0.003 1.310 revealed statistically significant correlations between the ICL
and the expression of reading comprehension (r = 0.21). Moreover,
R2 = 0.055, Durbin-Watson-Statistic = 2.133. p < 0.05.
significant correlations between reading speed (r = 0.21) and
cognitive-visual ability (N01; r = 0.16) and the expression of
The groups differed statistically significantly in all scales of the GCL were identified. However, for the second study (n = 69,
CL (Table 4). A large effect was shown according to Cohen see Supplementary Material), only one significant correlation
(1988) regarding the ICLs (d = 0.822). The 11th grade students could be found. Cognitive-visual ability (N02) correlated
rated their intrinsic CL about one value lower than the 9th significantly negatively with the ICL scale (r = 0.29). Apart from
grade students. Medium effects were identified in the ECLs that, multiple linear regressions were used to examine the
(d = 0.552). Regarding GCL, the groups differed statistically cognitive abilities (N01, N02, N03, and V03) and reading skill
significantly but only with a small effect size (d = 0.295). Thus, scales as possible predictor variables for CL in the use of
the GCL conducive to learning increases as expected in the multimedia scaffolds. Regarding the reading-skill scales for
higher grade level by a value of 0.3. To support the results reading comprehension, reading accuracy and reading speed
and to strengthen the link to prior knowledge as an influencing have no influence on the individual CL (see
variable, single factor ANOVAs with Bonferroni-correction were Supplementary Material). In contrast, the cognitive-verbal and
conducted on the expression of the individual types of CL visual abilities (V03, N01, and N02) scales showed a significant

Frontiers in Psychology | www.frontiersin.org 11 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

TABLE 9 | Selection frequencies of scaffolds. TABLE 12 | Pearson correlation with eye-tracking metrics and cognitive load
scales (n = 69).
Scaffolds Selection frequencies
Total fixation Mean Total fixation
Total sample study I (n = 180) study II (n = 69) count fixation duration
(n = 249) duration

nj hj nj hj nj hj ECL r 0.247 −0.233 0.065

image-text 75 30.1 53 29.4 22 31.9 p 0.041 0.055 0.596
image- 20 8.0 15 8.3 5 7.2 GCL r −0.079 0.169 0.089
audio p 0.517 0.165 0.467
animation 62 24.9 46 25.6 16 23.2 ICL r −0.020 0.140 0.056
video 92 36.9 66 36.7 26 37.7 p 0.871 0.252 0.646

nj = absolute frequency = sum of used scaffolds in the respective format; hj = relative p < 0.05.
frequency (in percent) = relative share of the selected scaffolds in relation to the total
number of scaffolds used.
TABLE 13 | Linear regression between ECL and total fixation count (n = 69)

Coefficient B SE(B) BETA P VIF

TABLE 10 | Multiple regression between ECL and representation preference
(n = 250).
(Constant) −0.118 0.103
Total fixation count 0.216 0.04 0.247 0.041 1.000
Coefficient B SE(B) BETA P VIF
R2 = 0.061, Durbin-Watson-Statistic = 2.220. p < 0.05.
(Constant) 1.951 0.105
image-text 0.416 0.156 0.188 0.008 1.268
image-audio 0.449 0.248 0.120 0.072 1.120 (95%-CI[0.31, 17.59]/t(167) = 2.05, p = 0.042, d = 0.318) and ECL
animation 0.416 0.165 0.165 0.138 1.257 (95%-CI[0.01, 0.51]/t(247) = 2.01, p = 0.046, d = 0.256, r = 0.132).
Accordingly, users of a monomodal scaffolds exhibited significantly
R2 = 0.033, Durbin-Watson-Statistic = 1.939. p < 0.05.
higher reading speed as well as ECL than those who selected
and used a multimedia one. Furthermore, no statistically
TABLE 11 | Multiple regression between GCL and representation preference significant differences between the multimedia and monomodal
(n = 250). users could be identified (see Supplementary Material). In
addition, group differences between good and poor readers
Coefficient B SE(B) BETA P VIF were analyzed regarding the GCL, ECL, and ICL scores for
the text-based (monomodal) scaffolds. Again, no significant
(Constant) 4.170 0.110
image-text −0.346 0.165 −0.149 0.037 1.268 differences between the individual groups and the expression
image-audio −0.270 0.261 −0.069 0.302 1.120 of cognitive load were found (see Supplementary Material).
animation −0.009 0.174 −0.004 0.959 1.257

R2 = 0.023, Durbin-Watson-Statistic = 1.873. p < 0.05. Validity Assumption IV

To test the fourth validity assumption, correlations (Table 12)
influence on the ECL (Table 6), GCL (Table 7), and ICL were first calculated between the z-standardized eye-tracking
(Table 8) within the total sample. Multiple linear regressions metrics and the expressions of CL. Within the sample (n = 69),
with categorical variables were performed to investigate the a significant correlation between the total fixation count and
influence of representation preference on the different types of the expression of ECL was found (r = 0.247). Although no other
CL. The results show a significant difference regarding the significant correlations were evident, the correlation coefficient
expression of the ECL and GCL between the learners who between the mean fixation duration and the expression of
used scaffolds with static image and text and those who learned ECL, which is just above the significance level, also indicated
with the video (Tables 10, 11). The ECL of students who learned a possible relationship. An examination of the total fixation
with static images and text was significantly higher than for count as a predictor for the individual CL revealed no significant
the video users (p = 0.008), whereas GCL decreased significantly results for GCL and ICL. In contrast, the total fixation count
(p = 0.037). As expected from the validity assumption III, the showed a significant effect on ECL (Table 13).
representation preference that is equivalent to the voluntary
choice of the scaffold had no further significant influence on
the expression of CL (see Supplementary Material and Tables 10, DISCUSSION
11). For an in-depth analysis of possible correlations or differences
between representation preference and the expression of reading The aim of the study was to test a subjective instrument for
skills, cognitive abilities and CL, group differences between measuring differential CL for 9th and 11th grade students across
those who used multimodal and those who used monomodal different validity assumptions. Hereby, possible relationships between
scaffolds in the planning phase were examined. Differences were causal factors (learner characteristics: e.g., cognitive, verbal and
found with respect to the expression of reading speed visual abilities, reading skills, and representation preference) and

Frontiers in Psychology | www.frontiersin.org 12 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

assessment factors of CL were investigated (Figure 1). Furthermore, the instructions from the video-based scaffold were easier to relate
an objective measurement method (eye-tracking: total fixation to each other due to their auditory and visual layout (Paivio,
count, fixation duration, fixation, and mean fixation duration) 2006; Mayer, 2014) or that the learners tended to overestimate
was used. With the constant inclusion of the overarching their own abilities in dealing with the image-text representation
methodological research question (see “Introduction”), different combination. The last is of course also strongly related to the
analyses were carried out for the respective assumptions (see subject-specific complexity as well as the formulation of the
“Aims of the Study, Research Question, Assumptions & Hypotheses”). ECL items, since, for example, “unclear language” (see
In the following, the individual validity assumptions will be discussed Supplementary Table A) can be understood ambiguously.
and combined in a final concluding discussion in relation to the Furthermore, the significant difference in grade 11 in relation to
research question. the GCL among the learners who chose an image-text combination
or an animation can be interpreted in such a way that the students
encountered a hurdle during the linking and extraction of the
Discussion of Validation Assumptions I–IV content from the scaffolds (see “Validity Assumption II”). The
The first assumption targets the internal consistency as well dynamic presentation of text and images in the animations may
as significant correlations between the expected scales of the have minimized the mental interconnection of the information.
ECL, ICL, and GCL for validity testing. The constructed Another explanation could be that the image-text combination
subjective measurement instrument is closely based on existing supports repetitive reading and self-regulated comprehension due
test items (Cierniak et al., 2009b; Leppink et al., 2013; Klepsch to the static presentation (Kastaun and Meier, 2021).
et al., 2017). Even with language and content adaptation, the Regarding the third validity assumption that the individual
evidence-based structure to the CL is confirmed in the dimensions expression of the cognitive-visual and verbal abilities, reading
of the ECLs, ICLs, and GCLs (H1.1). The examination of skills, spatial ability, and representation preference (causal
internal consistency shows satisfactory to very excellent factors) is related to the CL, a differential overall assessment
Cronbachs-alpha values (see “Validity Assumption I”) and, emerges. In comparison with the results of the spatial ability,
except for the values from the ICL, coincides with the results there are correlations to the other causal factors from which
from other studies (Klepsch et al., 2017). In this context, different explanatory approaches can be derived in relation to
possible reasons can be identified in the linguistic construction the overarching question regarding the validity test of the
as these items of the ICL (for example: “I found the identification subjective test instrument (H3.4). Cognitive-verbal and visual
of the dependent and independent variables to be difficult.,” see abilities (N01, N02, and V03) as well as reading skills are
Supplementary Table A) did not directly refer to the use of included as variables in some studies on CL but are not further
scaffold but to the definition of the individual variables, and investigated in a differentiated analysis of CL (e.g., Cierniak
thus, to the complexity of variable operationalization (Leppink et al., 2009b). Furthermore, different test instruments exist
et al., 2013; Klepsch et al., 2017; Klepsch and Seufert, 2021). and can be used to measure this, which makes it difficult to
In contrast, the GCL and ECL items refer almost exclusively transfer the results in respect to the measured construct (e.g.,
to the scaffolds and less to the task within the planning phase verbal reasoning ability; e.g., van de Weijer-Bergsma and van
of the experiment (see Supplementary Table A). This might der Ven, 2021). Thus, it is currently not clear how cognitive
be one reason why the GCL and ECL scales are not significantly abilities relate to the individual scales of CL. In the present
correlated with each other, in contrast to the ECL and ICL study, relationships and influences of cognitive-verbal and visual
scales (Leppink et al., 2013; see “Introduction”). Furthermore, abilities as well as influences on CL and its differentiated
the results confirm the additive relationship between the ICL categories could be identified (Tables 5–11; H3.1; H3.2; and
and ECL (see “Introduction”; Sweller et al., 2011). H3.3). Possible explanations can be found, among others, in
An examination of possible differences between the two grades the construction of the items and the item structure (see
9th and 11th regarding the expression of CL comprises the second “Cognitive Ability Test,” relationship and series tasks) of the
validity assumption. The group comparisons show results in line individual scales of the cognitive ability test used in this study
with expectations that the expression of CL and its categories (Heller and Perleth, 2000). These scales (N01, N02, and V03)
ECL, GCL, and ICL is less pronounced in 11th grade students measure content-based skills by determining logical thinking
than in 9th grade. This can be attributed to differently developed using content-based, here, visual and verbal examples via series/
prior knowledge regarding the content, and it is likely to constitute classification (N01) and ratio/analogy tasks (N02 & V03).
a consolidation of the more developed self-regulation of the learners According to Jäger et al.’s (2006) intelligence model, in which
from 11th grade (among others Richter et al., 2018; Eitel et al., the scales measure verbal and visual abilities under a constant
2020; H2.1). Considering the group differences in the respective reference to the individual thinking ability, content-based
grade level regarding the used scaffold and the expression of the cognitive abilities contribute to a person’s general intelligence.
ECL, ICL, and GCL, the results do not only indicate a difference In terms of CL, this means that the correlations are not only
in subject-related knowledge. In contrast to the older 11th grade due to students’ verbal or visual level, but also due to the
students, who did not differ significantly in ECL, significant ability to easily recognize links between different information.
differences appear among the younger students between those Furthermore, the results can also be explained by the construction
who used an image-text combination and those who chose a of the scaffolds. The primary information carriers of the
video (see “Validity Assumption II”). This suggests that either multimedia scaffolds are the verbal or textual representations,

Frontiers in Psychology | www.frontiersin.org 13 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

i.e., the visual representations also contribute to content scaffold, influences ECL and GCL. Thus, those students who
development but without the verbal or text-based elements show a preference for the image-text representation combination
they do not explain the main content. are more likely to experience significantly higher ECL than those
Considering the multimedia learning theory (Moreno, 2010; who use the video. This is also evident in the GCL, which
Mayer, 2014), which states that verbal and text-based information negatively predicts preference for image-text compared to video
can be reorganized in working memory via images and sounds, users. This can be attributed to the level of complexity of the
high verbal cognitive reasoning ability can help minimize ECL text-based or auditory text (Leahy and Sweller, 2011; Lehmann
and increase GCL. This is also evident from a closer look at and Seufert, 2020) and/or to the multimedia presentation mode
the subsamples (see “Validity assumption III” and of the verbal text in terms of simultaneous, dynamic visualizations
Supplementary Material). The explanatory approach, which (Moreno, 2010; Mayer, 2014) or due to the students’ failure to
implies that those with a higher expression of cognitive-verbal assess what kind of multimedia representation they could best
abilities benefit more than those who show an increased visual handle in the situation (instruction).
ability due to the design characteristics (among others Choi The examination of a possible convergence between the
and Sadar, 2011), is supported by the significant negative subjective and objective instruments for CL (validity assumption
correlations present in study I and II (see “Validity assumption IV) showed results partially confirming to expectations. The
III”) as well as by the findings of the multiple regression positive correlation (Table 12) and the significant influence
analyses. These show a strong to moderate effect of cognitive (Table 13) of the absolute fixation count support the assumed
abilities on the individual load categories. In respect to the convergence (H4.1; Zu et al., 2020). Zu et al. (2020) found a
scaffolds used, cognitive-verbal abilities predicted low ECL and moderate negative correlation between the ECL and the mean
higher GCL, whereas cognitive-visual abilities scales did not fixation duration and a weak correlation between the total visit
predict any of the CL categories (see “Validity assumption III”). duration on an animation and the GCL. In our work, we found
Concerning the study of reading skills as a causal factor to a weak negative correlation between the ECL and the mean
CL, the results show that positive relationships exist between fixation duration that was just above the significance level (p = 0.55)
reading comprehension as well as reading speed and GCL. On and a significant positive correlation between total fixation count
the one hand, this can be attributed to the fact that cognitive- and the ECL. This implies that previously reported relationships
verbal reasoning skills are closely and positively related to the between eye-tracking measures and cognitive load do not translate
individual reading skills (Schneider et al., 2017). On the other to our setting. The examination of other eye-tracking metrics,
hand, it shows that the reading process is an active construction such as transitions between the AOIs, was not possible based
activity (Artelt et al., 2007; Scheerer-Neumann, 2018). Thus, the on the material used here since possible selection and integration
significant correlations may result from the higher load based processes cannot be investigated unambiguously via visual attention
on the selected scaffold (monomodal) but also from the text- when including auditory scaffolds.
based self-assessment of the CL. Accordingly, the text-based test
itself may have elicited a correlation between the GCL and
reading comprehension or reading speed. Contrary to the identified Limitations and Implications for Future
correlations, the results of the multiple regression analyses do Work
not show any significant influences of the reading skills on the Along the individual validity assumptions based on the theoretical
three types of CL (see Supplementary Material). Further analyses interpretation-use argumentation (Kane, 2013), arguments for a
between the use of the monomodal/multimodal scaffolds and possible valid measurement via the test scores of the measurement
the expression of cognitive load in relation to the reading skills instruments are used (causal & assessment factors; objective
also showed differentiated results. It is possible that the reading instruments), and their relationships to each other are identified.
skills do not have a significant influence on the decision of the The central goal of validating a subjective measurement instrument
selected scaffold but still show advantages in the processing of for 9th and 11th grade students was partially achieved by
the information (based on the modality) about the load. confirming individual validity assumptions under study-specific
The relationship between spatial ability and CL (H3.4), which conditions. The results support a valid measurement but also
has been supported and assumed by many studies (e.g., Höffler, indicate that a validity test of a subjective measurement instrument
2010; Anvari et al., 2013), is not evident in the present study. for CL is associated with different difficulties. On the one hand,
As already stated by Höffler and Leutner (2011), among others, CL can be influenced by diverse causal factors in different ways
it also becomes clear with the present results that the spatial (Figure 1; Choi et al., 2014), which is particularly related to
ability only has a significant influence on task performance if the construction of the task and/or environment as well as its
the task itself requires it. Consequently, it can be concluded interaction with CL. On the other hand, a difficulty arises from
that spatial awareness is not required for processing the the fact that there are no comprehensive, tested relationships
information of the scaffolds used here. between the individual subjective and objective measurement
Finally, the assumed correlations between the causal factor instruments for CL (assessment factors) in an evidence-based
of representation preference and CL can be partially confirmed manner yet. Particularly regarding the cognitive abilities for
(H3.5). Independent of the selected scaffold, the ECL shows a visual and verbal reasoning, more detailed investigation is needed
very low expression (see “Procedure”). Multiple regression analyses to determine whether these factors influence CL in a similar
confirm that the representation preference, i.e., the choice of way as prior knowledge (expertise-reversal effect: Cierniak et al.,

Frontiers in Psychology | www.frontiersin.org 14 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

2009a; Kalyuga, 2013). Another limitation of our study is the DATA AVAILABILITY STATEMENT
missing inclusion of learning performance. Since the CLT assumes
that CL also influences learning performance (see Figure 1), it The raw data supporting the conclusions of this article will
would have been appropriate to include this in relation to the be made available by the authors, without undue reservation.
manifestations of cognitive load in relation to the causal factors.
Due to our aims and the corroborating, temporally extensive
design of the study, we were not able to consider learning ETHICS STATEMENT
performance. The focus was on the design of the scaffolds and
the resulting cognitive load. In the future, this should be captured Ethical review and approval was not required for the study
by valid instruments for the methodological skills/knowledge on human participants in accordance with the local legislation
for the planning (content of scaffolds) to extend the analyzed and institutional requirements. Written informed consent to
relationships between causal and assessment factors. However, participate in this study was provided by the participants’ legal
the results of this study demonstrate that performance-based guardian/next of kin.
(cognitive abilities and reading skills) and objective (eye-tracking
metrics) instruments are linked to the subjectively obtained test
scores on individual CL. Therefore, further analyses are needed AUTHOR CONTRIBUTIONS
in the future to examine the relationships between the individual
MM developed the basic idea for the present study and supervised
causal factors in relation to different tasks and the resulting
the project. MK and MM contributed to the conception and
CL. However, not only the findings between the causal and
design of the studies. MK developed the used material and
assessment factors show further approaches of investigation, but
questionnaires and led data collection for studies, analyzed all
also the subjective measurement instrument from its linguistic
the data and interpreted it with the help of MM, SK, and JK,
composition by itself. Thus, the results indicate that especially
and was mainly responsible for writing the paper. SK and JK
the linguistic construction of the items has a great influence
supported in the execution and analysis of the eye-tracking data.
on the actual measurability of the individual constructs. On
This article is part of the Ph.D. thesis of MK supervised by
the one hand, there is the difficulty of phrasing the items
MM. All authors agreed to the final submitted version of the
(especially GCL and ICL) in such a way that they reflect the
manuscript and agreed to be responsible for all aspects of the
construct to be measured (Wording, see “Introduction”). Above work, ensuring that questions regarding the accuracy or integrity
all, the measurement of the GCL must be considered with of any part of the work were adequately investigated and resolved.
reservations. In addition to the term “knowledge,” the word
“understanding” is used in the items. It is questionable whether
students’ understanding and comprehension can actually
FUNDING
be equated with GCL (Ayres, 2018), and how the corresponding
items can be worded differently to express quantify GCL. On The project on which this article is based and the associated
the other hand, there is the challenge of adapting the linguistic supervision by MM was funded by the Deutsche Telekom
level to the learners in such a way that they can understand Stiftung within the framework of the program Fellowship
and assess it independently (Sweller et al., 2011; Klepsch and Fachdidaktik MINT.
Seufert, 2021). In conjunction with the validation approach
according to Kane (2013), which considers the generalizability
of the test instrument as an elementary part of validity, the SUPPLEMENTARY MATERIAL
existing subjective test instruments should be examined in the
future regarding a possible standardization with tested adaptation The Supplementary Material for this article can be found online
possibilities that fit the investigated setting (self-regulated learning at: https://www.frontiersin.org/articles/10.3389/fpsyg.2021.703857
or problem-based learning). /full#supplementary-material

computer graphics programming. Int. J. Inf. Educ. Technol. 3:94. doi: 10.7763/
REFERENCES IJIET.2013.V3.241
AERA, APA, and NCME (2014). Standards for Educational and Arnold, J. C., Kremer, K., and Mayer, J. (2014). Understanding
Psychological Testing. Washington, DC: American Educational Research students’ experiments – what kind of support do they need in
Association. inquiry tasks? Int. J. Sci. Educ. 36, 2719–2749. doi: 10.1080/09500693.
Ainsworth, S. (2006). DeFT: a conceptual framework for considering learning 2014.930209
with multiple representations. Learn. Instruct. 16, 183–198. doi: 10.1016/j. Artelt, C., McElvany, N., Christmann, U., Richter, T., Groeben, N., Köster, J.,
learninstruc.2006.03.001 et al. (2007). Förderung von Lesekompetenz – Expertise (Bildungsforschung,
Alfieri, L., Brooks, P. J., Aldrich, N. J., and Tenenbaum, H. R. (2011). Does 17). Berlin: Bundesministerium für Bildung und Forschung.
discovery-based instruction enhance learning? J. Educ. Psychol. 103, 1–18. Ayres, P. (2018). “Subjective measures of cognitive load: what can they reliably
doi: 10.1037/a0021017 measure?,” in Cognitive Load Measurement and Application: A Theoretical
Anvari, F., Tran, H., and Kavakli, M. (2013). Using cognitive load measurement Framework for Meaningful Research and Practice. ed. R. Z. Zheng (New York:
and spatial ability test to identify talented students in three-dimensional Routledge/Taylor & Francis Group), 9–28.

Frontiers in Psychology | www.frontiersin.org 15 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

Baars, M., Wijnia, W., and Paas, F. (2017). The association between motivation, Huh, D., Kim, J. H., and Jo, I. H. (2019). A novel method to monitoring
affect, and self-regulated learning when solving problems. Front. Psychol. changes in cognitive load in videobased learning. J. Comput. Assist. Learn.
8:1346. doi: 10.3389/fpsyg.2017.01346 35, 721–730. doi: 10.1111/jcal.12378
Bell, R., Smetana, L. K., and Binns, I. (2005). Simplifying inquiry instruction. Huk, T. (2006). Who benefits from learning with 3D models? The case of
Sci. Teach. 72, 30–33. spatial ability. J. Comput. Assist. Learn. 22, 392–404. doi: 10.1111/j.1365-2729.
Chen, S., and Epps, J. (2014). Using task-induced pupil diameter and blink 2006.00180.x
rate to infer cognitive load. Hum. Comput. Interact. 29, 390–413. doi: Jäger, A. O. (1984). Intelligenzstrukturforschung: konkurrierende Modelle, neue
10.1080/07370024.2014.892428 Entwicklungen, Perspektiven. Psychol. Rundsch. 35, 21–35.
Chen, O., Kalyuga, S., and Sweller, J. (2016). Relations between the worked Jäger, A. O., Holling, H., Preckel, F., Schulze, R., Vock, M., Süß, H.-M., et al.
example and generation effects on immediate and delayed tests. Learn. (2006). BIS-HB: Berliner Intelligenzstrukturtest für Jugendliche: Begabungs- und
Instruct. 45, 20–30. doi: 10.1016/j.learninstruc.2016.06.007 Hochbegabungs-diagnostik – Manual. Göttingen: Hogrefe.
Chen, H., Ning, X., Wang, L., and Yang, J. (2018). Acquiring new factual Jäger, D., Itsios, C., Franz, T., and Müller, R. (2017). Cognitive Load und
information: effect of prior knowledge. Front. Psychol. 9:1734. doi: 10.3389/ Aufgabenmerkmale–Verwendung von Zusatzfragen bei authentischen
fpsyg.2018.01734 Problemen. Didaktik der Physik - Beiträge zur DPG-Frühjahrstagung PhyDid.
Choi, J., and Sadar, S. (2011). An empirical investigation of the relationships B, DD, 13.1, 125–130.
Among cognitive abilities, cognitive style, and learning preferences in students Kaiser, I., and Mayer, J. (2019). The long-term benefit of video modeling
enrolled in specialized degree courses at a Canadian college. Can. J. Scholarsh. examples for guided inquiry. Front. Educ. 4:104. doi: 10.3389/
Teach. Learn. 2:5. doi: 10.5206/cjsotl-rcacea.2011.1.5 feduc.2019.00104
Choi, H. H., van Merriënboer, J. J. G., and Paas, F. (2014). Effects of the Kaiser, I., Mayer, M., and Malai, D. (2018). Self-generation in context of inquiry-
physical environment on cognitive load and learning: towards a new model based learning. Front. Psychol. 9:2440. doi: 10.3389/fpsyg.2018.02440
of cognitive load. Educ. Psychol. Rev. 26, 225–244. doi: 10.1007/ Kalyuga, S. (2011). Cognitive load theory: how many types of load
s10648-014-9262-6 does it really need? Educ. Psychol. Rev. 23, 1–19. doi: 10.1007/
Cierniak, G., Gerjets, P., and Scheiter, K. (2009a). Expertise reversal in multimedia s10648-010-9150-7
learning: subjective load ratings and viewing behavior as cognitive process Kalyuga, S. (2013). Effects of learner prior knowledge and working memory
indicators. Proc. Annu. Meet. Cogn. Sci. Soc. 31, 1906–1911. limitations on multimedia learning. Procedia. Soc. Behav. Sci. 83, 25–29.
Cierniak, G., Scheiter, K., and Gerjets, P. (2009b). Explaining the split-attention doi: 10.1016/j.sbspro.2013.06.005
effect: is the reduction of extraneous cognitive load accompanied by an Kalyuga, S., Ayres, P., Chandler, P., and Sweller, J. (2003). The expertise reversal
increase in germane cognitive load? Comput. Hum. Behav. 25, 315–324. effect. Educ. Psychol. 38, 23–31. doi: 10.1207/S15326985EP3801_4
doi: 10.1016/j.chb.2008.12.020 Kane, M. (2006). “Content-related validity evidence in test development,”
Cohen, J. (1988). The effect size. Statistical power analysis for the behavioral in Handbook of Test Development. eds. S. M. Downing and T. M.
sciences, 77–83. Haladyna (Mahwah, New Jersey: Lawrence Erlbaum Associates Publishers),
Cronbach, L. J., and Meehl, P. E. (1955). Construct validity in psychological 131–153.
tests. Psychol. Bull. 52, 281–302. doi: 10.1037/h0040957 Kane, M. T. (2013). Validating the interpretations and uses of test scores.
de Jong, T. (2010). Cognitive load theory, educational research, and instructional J. Educ. Meas. 50, 1–73. doi: 10.1111/jedm.12000
design: some food for thought. Instr. Sci. 38, 105–134. doi: 10.1007/ Kastaun, M., Hunold, C., and Meier, M. (2020). Eye-Tracking – Visuelle
s11251-009-9110-0 Wahrnehmung sichtbar machen. Biologie in unserer Zeit 3, 172–173. doi:
de Jong, T. (2019). Moving towards engaged learning in STEM domains; there 10.1002/biuz.202070311
is no simple answer, but clearly a road ahead. J. Comput. Assist. Learn. 35, Kastaun, M., and Meier, M. (2021). “Eine qualitative Analyse von Blickdaten
153–167. doi: 10.1111/jcal.12337 bei statischen und dynamischen Repräsentationen im naturwissenschaftlichen
DeLeeuw, K. E., and Mayer, R. E. (2008). A comparison of three measures Erkenntnisprozess,” in Eye Tracking als Methode in der Mathematik- und
of cognitive load: evidence for separable measures of intrinsic, extraneous, Naturwissenschaftsdidaktik: Forschung und Praxis. eds. P. Klein, M. Schindler,
and germane load. J. Educ. Psychol. 100, 223–234. doi: 10.1037/0022-0663. N. Graulich and J. Kuhn (Heidelberg: Springer Spectrum).
100.1.223 Kendeou, P., and van den Broek, P. (2007). The effects of prior knowledge
Eitel, A., Endres, T., and Renkl, A. (2020). Self-management as a bridge between and text structure on comprehension processes during reading of scientific
cognitive load and self-regulated learning: the illustrative case of texts. Mem. Cogn. 35, 1567–1577. doi: 10.3758/BF03193491
seductive details. Educ. Psychol. Rev. 32, 1073–1087. doi: 10.1007/s10648- Kind, P., and Osborne, J. (2017). Styles of scientific reasoning—a
020-09559-5 cultural rationale for science education? Sci. Educ. 101, 8–31. doi: 10.1002/
Furtak, E. M., Seidel, T., Iverson, H., and Briggs, D. C. (2012). Experimental sce.21251
and quasi-experimental studies of inquiry-based science teaching: a meta- Kirschner, P. A., Ayres, P., and Chandler, P. (2011). Contemporary cognitive
analysis. Rev. Educ. Res. 82, 300–329. doi: 10.3102/0034654312457206 load theory research: The good, the bad and the ugly. Comput. Hum. Behav.
Heller, K. A., and Perleth, C. (2000). KFT 4–12+ R. Kognitiver Fähigkeitstest 27, 99–105. doi: 10.1016/j.chb.2010.06.025
für 4. bis 12. Klassen, Revision. Göttingen: Beltz Test. Kirschner, P. A., Sweller, J., and Clark, R. E. (2006). Why minimal guidance
Hmelo-Silver, C. E., Duncan, R. G., and Chinn, C. A. (2007). Scaffolding and during instruction does not work: an analysis of the failure of constructivist,
achievement in problem-based and inquiry learning: a response to Kirschner, discovery, problem-based, experiential, and inquiry-based teaching. Educ.
Sweller, and Clark (2006). Educ. Psychol. 42, 99–107. doi: 10.1080/004615 Psychol. 41, 75–86. doi: 10.1207/s15326985ep4102_1
20701263368 Klepsch, M., Schmitz, F., and Seufert, T. (2017). Development and validation
Ho, H. N. J., Tsai, M. -J., Wang, C.-Y., and Tsai, C. C. (2014). Prior knowledge of two instruments measuring intrinsic, extraneous, and germane cognitive
and online inquiry-based science reading: evidence from eye tracking. Int. load. Front. Psychol. 8:1997. doi: 10.3389/fpsyg.2017.01997
J. Sci. Math. Educ. 12, 525–554. doi: 10.1007/s10763-013-9489-6 Klepsch, M., and Seufert, T. (2021). Making an effort versus experiencing load.
Höffler, T. N. (2010). Spatial ability: its influence on learning with visualizations - Front. Educ. 6:645284. doi: 10.3389/feduc.2021.645284
a meta-analytic review. Educ. Psychol. Rev. 22, 245–269. doi: 10.1007/ Koć-Januchta, M. M., Höffler, T. N., Eckhardt, M., and Leutner, D. (2019).
s10648-010-9126-7 Does modality play a role? Visual-verbal cognitive style and multimedia
Höffler, T. N., and Leutner, D. (2011). The role of spatial ability in learning learning. J. Comput. Assist. Learn. 35, 747–757. doi: 10.1111/jcal.12381
from instructional animations – evidence for an ability-as- Korbach, A., Brünken, R., and Park, B. (2018). Differentiating different types
compensator hypothesis. Comput. Hum. Behav. 27, 209–216. doi: 10.1016/j. of cognitive load: a comparison of different measures. Educ. Psychol. Rev.
chb.2010.07.042 30, 503–529. doi: 10.1007/s10648-017-9404-8
Höffler, T. N., and Schwartz, R. N. (2011). Effects of pacing and cognitive Krell, M. (2017). Evaluating an instrument to measure mental load and mental
style across dynamic and non-dynamic representations. Comput. Educ. 57, effort considering different sources of validity evidence. Cogent Educ. 4:1280256.
1716–1726. doi: 10.1016/j.compedu.2011.03.012 doi: 10.1080/2331186X.2017.1280256

Frontiers in Psychology | www.frontiersin.org 16 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

Kuhn, D., Amsel, E., and O’Loughlin, M. (1988). The Development of Scientific Plass, J. L., Chun, D. M., Mayer, R. E., and Leutner, D. (2003). Cognitive load
Thinking Skills. San Diego, CA: Acadermic Press. in reading a foreign language text with multimedia aids and the influence
Kürschner, C., Schnotz, W., Eid, M., and Hauck, G. (2005). Individuelle of verbal and spatial abilities. Comput. Hum. Behav. 19, 221–243. doi: 10.1016/
Modalitätspräferenzen beim Textverstehen: Präferenzen für auditive oder S0747-5632(02)00015-8
visuelle Sprachverarbeitung in unterschiedlichen Bevölkerungsgruppen. Z. Plass, J. L., and Homer, B. D. (2002). “Cognitive load in multimedia learning:
Entwicklungspsychol. Padagog. Psychol. 37, 2–16. doi: 10.1026/0049-8637.37.1.2 the role of learner preferences and abilities,” in Proceedings of the International
Leahy, W., and Sweller, J. (2011). Cognitive load theory, modality of presentation Conference on Computers in Education. ed. B. Werner (USA: IEEE Computer
and the transient information effect. Appl. Cogn. Psychol. 25, 943–951. doi: Society), 564–568.
10.1002/acp.1787 Pollock, E., Chandler, P., and Sweller, J. (2002). Assimilating complex information.
Lehmann, J., and Seufert, T. (2020). The interaction between text modality Learn. Instruct. 12, 61–86. doi: 10.1016/S0959-4752(01)00016-0
and the learner’s modality preference influences comprehension and cognitive Quaiser-Pohl, C., Lehmann, W., and Schirra, J. (2001). Sind Studentinnen
load. Front. Psychol. 10:2820. doi: 10.3389/fpsyg.2019.02820 der Computervisualistik besonders gut in der Raumvorstellung?
Leppink, J., Paas, F., van der Vleuten, C. P. M., van Gog, T., and Van Psychologische Aspekte bei der Wahl eines Studienfachs. FIfF Kommunikation
Merriënboer, J. J. G. (2013). Development of an instrument for measuring 18, 42–46.
different types of cognitive load. Behav. Res. Ther. 45, 1058–1072. doi: Rheinberg, F., Vollmeyer, R., and Burns, B. D. (2000). “Motivation and self-
10.3758/s13428-013-0334-1 regulated learning,” in Motivational Psychology of Human Development:
Leutner, D., Funke, J., Klieme, E., and Wirth, J. (2005). “Problemlösefähigkeit Developing Motivation and Motivating Development. ed. J. Heckhausen
als fächerübergreifende Kompetenz,” in Problemlösekompetenz von Schülerinnen (Amsterdam: Elsevier Science), 81–108.
und Schülern. eds. E. Klieme, D. Leutner and J. Wirth (Wiesbaden: VS Richter, J., Scheiter, K., and Eitel, A. (2018). Signaling text–picture relations
Verlag für Sozialwissenschaften), 11–19. in multimedia learning: the influence of prior knowledge. J. Educ. Psychol.
Mayer, R. E. (2014). “Introduction to multimedia learning,” in The Cambridge 110, 544–560. doi: 10.1037/edu0000220
Handbook of Multimedia Learning. ed. R. E. Mayer (Cambridge MA: Cambridge Sanchez, C. A., and Wiley, J. (2014). The role of dynamic spatial ability in
University), 1–24. geoscience text comprehension. Learn. Instruct. 31, 33–45. doi: 10.1016/j.
Mayer, R. E., and Massa, L. J. (2003). Three facets of visual and verbal learners: learninstruc.2013.12.007
cognitive ability, cognitive style, and learning preference. J. Educ. Psychol. Scheerer-Neumann, G. (2018). Lese-Rechtschreib-Schwäche und Legasthenie.
95, 833–846. doi: 10.1037/0022-0663.95.4.833 Grundlagen, Diagnostik und Förderung. Stuttgart: Verlag W. Kohlhammer.
Meier, M., and Kastaun, M. (2021). “Lernunterstützungen als Werkzeug Scheiter, K., Ackerman, R., and Hoogerheide, V. (2020). Looking at mental
individualisierter Förde-rung im naturwissenschaftlichen Erkenntnisprozess,” effort appraisals through a metacognitive lens: are they biased? Educ. Psychol.
in Vielfältige Wege biologiedidaktischer Forschung. eds. M. Meier, C. Wulff Rev. 32, 1003–1027. doi: 10.1007/s10648-020-09555-9
and K. Ziepprecht (Münster: Waxmann), 95–116. Schneider, W., Schlagmüller, M., and Ennemoser, M. (2017). LGVT 5–12+.
Messick, S. (1995). Validity of psychological assessment: validation of Lesegeschwindigkeits- und Verständnistest für die Klassen 5–12. Göttingen:
inferences from persons’ responses and performances as scientific inquiry hogrefe.
into score meaning. Am. Psychol. 50, 741–749. doi: 10.1037/0003-066X.50.9.741 Seufert, T. (2018). The interplay between self-regulation in learning and cognitive
Minkley, N., Kärner, T., Jojart, A., Nobbe, L., and Krell, M. (2018). Students’ load. Educ. Res. Rev. 24, 116–129. doi: 10.1016/j.edurev.2018.03.004
mental load, stress, and performance when working with symbolic or Seufert, T. (2019). Training for coherence formation when learning from text
symbolic–textual molecular representations. J. Res. Sci. Teach. 55, 1162–1187. and picture and the interplay with learners’ prior knowledge. Front. Psychol.
doi: 10.1002/tea.21446 10:193. doi: 10.3389/fpsyg.2019.00193
Minkley, N., Xu, K. M., and Krell, M. (2021). Analyzing relationships between Shrager, J., and Siegler, R. S. (1998). SCADS: A model of children’s strategy
causal and assessment factors of cognitive load: associations between objective choices and strategy discoveries. Psychol. Sci. 9, 405–410. doi:
and subjective measures of cognitive load, stress, interest, and self-concept. 10.1111/1467-9280.00076
Front. Educ. 6:632907. doi: 10.3389/feduc.2021.632907 Siegler, R. S. (2005). Children’s learning. Am. Psychol. 60, 769–778. doi:
Moreno, R. (2010). Cognitive load theory: more food for thought. Instr. Sci. 10.1037/0003-066X.60.8.769
38, 135–141. doi: 10.1007/s11251-009-9122-9 Skuballa, I. T., Xu, K. M., and Jarodzka, H. (2019). The impact of co-actors
Münzer, S. (2012). Facilitating spatial perspective taking through animation: on cognitive load: when the mere presence of others makes learning
evidence from an aptitude–treatment-interaction. Learn. Individ. Differ. 22, more difficult. Comput. Hum. Behav. 101, 30–41. doi: 10.1016/j.chb.
505–510. doi: 10.1016/j.lindif.2012.03.002 2019.06.016
Nitz, S., Ainsworth, S. E., Nerdel, C., and Prechtl, H. (2014). Do student Solhjoo, S., Haigney, M. C., McBee, E., van Merriënboer, J. J. G., Schuwirth, L.,
perceptions of teaching predict the development of representational competence Artino, A. R., et al. (2019). Heart rate and heart rate variability correlate
and biological knowledge? Learn. Instruct. 31, 13–22. doi: 10.1016/j. with clinical reasoning performance and self-reported measures of cognitive
learninstruc.2013.12.003 load. Sci. Rep. 9, 1–9. doi: 10.1038/s41598-019-50280-3
OECD (2007). PISA 2006: Science Competencies for Tomorrow’s World. Vol. 1. Sweller, J. (1994). Cognitive load theory, learning difficulty, and instructional
Paris: OECD Publishing. design. Learn. Instruct. 4, 295–312. doi: 10.1016/0959-4752(94)90003-5
Paas, F., Tuovinen, J. E., Tabbers, H., and van Gerven, P. W. M. (2003). Cognitive Sweller, J., Ayres, P. L., and Kalyuga, S. (2011). Intrinsic and extraneous cognitive
load measurement as a means to advance cognitive load theory. Educ. load. In Cognitive Load Theory. New York, NY: Springer. 57–69.
Psychol. 38, 63–71. doi: 10.1207/S15326985EP3801_8 Sweller, J., van Merriënboer, J. J. G., and Paas, F. (2019). Cognitive architecture
Paas, F., van Gog, T., and Sweller, J. (2010). Cognitive load theory: new and instructional design: 20 years later. Educ. Psychol. Rev. 31, 261–292.
conceptualizations, specifications, and integrated research perspectives. Educ. doi: 10.1007/s10648-019-09465-5
Psychol. Rev. 22, 115–121. doi: 10.1007/s10648-010-9133-8 Thees, M., Kapp, S., Altmeyer, K., Malone, S., Brünken, R., and Kuhn, J.
Paas, F., and Van Merriënboer, J. J. G. (1994). Instructional control of cognitive (2021). Comparing two subjective rating scales assessing cognitive load
load in the training of complex cognitive tasks. Educ. Psychol. Rev. 6, During technology-enhanced STEM laboratory courses. Front. Educ. 6:705551.
351–371. doi: 10.1007/BF02213420 doi: 10.3389/feduc.2021.705551
Paas, F., and van Merriënboer, J. J. G. (2020). Cognitive-load theory: methods van de Weijer-Bergsma, E., and van der Ven, S. H. G. (2021). Why and for
to manage working memory load in the learning of complex tasks. Curr. whom does personalizing math problems enhance performance? Testing
Dir. Psychol. Sci. 29, 394–398. doi: 10.1177/0963721420922183 the mediation of enjoyment and cognitive load at different ability levels.
Paivio, A. (2006). Mind and its Evolution; A Dual Coding Theoretical Interpretation. Learn. Individ. Differ. 87:101982. doi: 10.1016/j.lindif.2021.101982
Mahwah: Lawrence Erlbaum Associates, Inc. van Merriënboer, J. J. G., Kester, L., and Paas, F. (2006). Teaching
Peterson, K., DeCato, L., and Kolb, D. A. (2015). Moving and learning: expanding complex rather than simple tasks: balancing intrinsic and germane load
style and increasing flexibility. J. Exp. Educ. 38, 228–244. doi: 10.1177/10538259 to enhance transfer of learning. Appl. Cogn. Psychol. 20, 343–352. doi:
14540836 10.1002/acp.1250

Frontiers in Psychology | www.frontiersin.org 17 August 2021 | Volume 12 | Article 703857

Kastaun et al. Cognitive Load in Scientific Inquiry

Zu, T., Hutson, J., Loschky, L. C., and Rebello, N. S. (2020). Using eye or those of the publisher, the editors and the reviewers. Any product that may
movements to measure intrinsic, extraneous, and germane load in a multimedia be evaluated in this article, or claim that may be made by its manufacturer, is
learning environment. J. Educ. Psychol. 112, 1338–1352. doi: 10.1037/ not guaranteed or endorsed by the publisher.
edu0000441
Copyright © 2021 Kastaun, Meier, Küchemann and Kuhn. This is an open-access
Conflict of Interest: The authors declare that the research was conducted in article distributed under the terms of the Creative Commons Attribution License
the absence of any commercial or financial relationships that could be construed (CC BY). The use, distribution or reproduction in other forums is permitted, provided
as a potential conflict of interest. the original author(s) and the copyright owner(s) are credited and that the original
publication in this journal is cited, in accordance with accepted academic practice.
Publisher’s Note: All claims expressed in this article are solely those of the No use, distribution or reproduction is permitted which does not comply with
authors and do not necessarily represent those of their affiliated organizations, these terms.

Frontiers in Psychology | www.frontiersin.org 18 August 2021 | Volume 12 | Article 703857