Computers in Human Behavior 26 (2010) 140–150
Contents lists available at ScienceDirect
Computers in Human Behavior
journal homepage: www.elsevier.com/locate/comphumbeh
The influence of leads on cognitive load and learning in a hypertext environment
Pavlo D. Antonenko a,*, Dale S. Niederhauser b
a
b
210 Willard Hall, Oklahoma State University, OK 74078, USA
N155 Lagomarcino Hall, Iowa State University, Ames, IA 50014, USA
a r t i c l e
i n f o
Article history:
Available online 30 November 2009
Keywords:
Cognition
Learning
Technology
Text structure
Psychophysiology
a b s t r a c t
The purpose of this study was to determine the effect of leads (or hypertext node previews) on cognitive
load and learning. Leads provided a brief summary of information in the linked node, which helped orient
the reader to the linked information. Dependent variables included measures of cognitive load: selfreport of mental effort, reading time, and event-related desynchronization percentage of alpha, beta
and theta brain wave rhythms; and learning performance: a recall task, and tests of domain and structural knowledge. Results indicated that use of leads reduced brain wave activity that may reflect split
attention and extraneous cognitive load, and improved domain and structural knowledge acquisition.
Further, findings provide insights into differentiating the types of cognitive load apparent in hypertext-assisted learning environments. Use of EEG measures allowed examination of instantaneous cognitive load, which showed that leads may be influencing germane load—reducing mental burden
associated with creating coherence between two linked node. The self-report of mental effort measure
appears more closely associated with overall and intrinsic load.
Published by Elsevier Ltd.
1. Introduction
2. Theoretical framework
Hypertext systems, like the World Wide Web, provide a linked
computer-based information storage and retrieval system, in
which massive amounts of information are organized into a vast
semantic network (Rada, 1989). Since its very inception, psychologists and educators have been enthusiastic about the potential of
hypertext to serve as an intellectual partner for the reader (e.g.,
Jonassen, 2000), thereby enriching the learning experience. The
unique characteristics of hypertext provide a mechanism for elaborating topics via hyperlinks, leading researchers to believe that
this format of text presentation requires learners to take a more active approach to reading by interacting more directly with the content (Landow, 1992), enhances conceptual and structural
knowledge acquisition (Jonassen & Wang, 1993), and results in
more flexible knowledge representations that enhance transfer of
learning (Spiro, Vispoel, Schmitz, Samarapungavan, & Boerger,
1987). However, despite initial hopes that using hypertext would
enhance learning, little empirical evidence exists to supports these
claims, and the cognitive consequences of hypertext-assisted
learning continue to be debated (Calandra & Barron, 2005; DeStefano & LeFevre, 2007).
Reading hypertext is a task of exploration. Unlike print text, that
is typically read in a sequence prescribed by the author, hypertext
is presented in brief nodes that contain one or more concise expository paragraphs, for which the reader determines access sequence.
Each node must be self-contained because hypertext authors cannot make assumptions about which nodes have already been read.
The unique characteristics of hypertext allow hypertext authors to
create connections to other related topics that are not easily
accomplished in traditional print text presentation. Hyperlinks
form a more intricate web of connected information nodes than
is permitted by the straight-forward flow of a print text, even texts
with elaborate annotations and footnotes. Thus, hypertext readers
can determine their own path through the content, supporting creation of their own unique knowledge representations by selecting
among the various linked concepts presented in hypertext nodes
(Rouet, Levonen, Dillon, & Spiro, 1996).
* Corresponding author. Tel.: +1 405 744 8003; fax: +1 405 744 7758.
E-mail address: [email protected] (P.D. Antonenko).
0747-5632/$ - see front matter Published by Elsevier Ltd.
doi:10.1016/j.chb.2009.10.014
2.1. Role of cognitive load in reading hypertext
Both traditional print text and hypertext provide contexts for
learning and organizing complex conceptual information. Processing print text and hypertext occurs on many levels, from low-level
processes like decoding characters, recognizing words, and parsing
sentences; to higher-level comprehension processes through
which readers develop a situation model of the content by
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
integrating new information into their existing knowledge base
(Niederhauser, 2008; Shapiro & Niederhauser, 2004; Alexander,
Kulikowich & Jetton, 1994). In addition to cognitive processes,
reading involves metacognitive activity. Skilled readers continuously monitor comprehension to identify possible information
gaps in their situation model (Azevedo & Jacobson, 2008), and
selectively allocate attentional resources based on perceived
importance and relevance of information (Reynolds, 2000).
While learning from print text and hypertext is grounded in
the same basic reading processes, making meaning in a hypertext-assisted learning environment places additional cognitive demands on the reader (Niederhauser, Reynolds, Salmen, &
Skolmoski, 2000). After reading a given node, hypertext readers
must actively decide which informational node to read next, given
their interests and learning goals on one hand, and the potential
relational-linking options provided by the interface links on the
other (Landow, 1992). Through this process of selecting nodes
and monitoring comprehension, readers engage in cognitive activity to integrate concepts from spatially distinct nodes (DeStefano
& LeFevre, 2007), and metacognitive activity to consider various
paths available to them, and choose some paths over others
(Charney, 1994). The additional cognitive and metacognitive processes involved in navigating and making meaning from linked
hypertext nodes appears to increase cognitive demands on the
reader.
For example, in a hypertext study that was grounded in cognitive flexibility theory (see Spiro, Coulson, Feltovich & Anderson,
1988), Niederhauser et al. (2000) hypothesized that students
who took advantage of compare and contrast links within a large
hierarchical hypertext on behaviorism and constructivism would
develop a deeper, more integrated understanding of this complex
domain than would those who chose to sequentially read through
all of the information on one learning theory, then read the information in the second theory in the same manner. Unexpectedly,
results indicated that learners who used links to compare and contrast concepts tended to have lower scores on learning measures
than did those who employed a more sequential approach characteristic of reading traditional print text. The latter group adopted a
strategy for choosing which node to read next that minimized their
use of cognitive resources associated with constructing an individualized path through the text. The authors concluded that the benefits these students might have gained by engaging in deeper
processing to compare and contrast the content may have been
mitigated by increased cognitive load (see Sweller, 1988, 1994)
associated with navigating the hypertext. Consistent with these results, other researchers have reported better factual recall from
reading linear texts when compared to hypertexts (Barab, Young,
& Wang, 1999; Eveland, Cortese, Park, & Dunwoody, 2004). Similarly, Zhu (1999) demonstrated that learning performance on a
multiple-choice test and written summary, as well as subjective
ratings of the hypertext system, were better when linking options
were limited to 3–7 links, when compared to a comparable system
containing 8–12 links. Cognitive load theory offers a useful framework to interpret these findings.
Cognitive load refers to the mental burden that performing a
task imposes on the learner (Sweller, 1988). Three types of cognitive load have been identified in the literature (Sweller, van
Merrienboer, & Paas, 1998). Intrinsic load is imposed by the inherent complexity of the content, which relates to the extent to which
various information elements interact. When information interactivity is low, content can be understood and learned one element at
a time. Conversely, highly interactive information is more difficult
to learn. For example, learning individual words of a foreign language is intrinsically less demanding than learning grammar because, while vocabulary can be learned word-by-word, learning
to construct a sentence requires understanding of both the words
141
that make up the sentence, and the rules that govern things like
word order and tense agreement.
While intrinsic load is generally thought to be immutable to
instructional manipulations due to the inherent complexity of
content (Sweller et al., 1998), learning difficulty can be manipulated by controlling the contributions of germane and extraneous
load. Unlike intrinsic load, which is caused by the number of
interacting information elements and the extent to which these
elements interact, germane load is mediated by the learner’s prior
knowledge of the domain, cognitive and metacognitive skills, and,
therefore, depends on the individual differences and learning
characteristics of each individual learner. Learners experience germane load, when learning activities and/or materials encourage
higher-order thinking and challenge the learner at an appropriate
level—within what Vygotsky (1978) called their zone of proximal
development. In hypertext-based learning materials, a potential
source of germane load is the cognitive flexibility that is involved
in the development of a situation model of the content. According
to Spiro et al. (1988), this criss-crossing of the conceptual landscape allows learners to achieve deeper, more connected understandings as they arrive at informational nodes from various
intellectual perspectives. Each user-control decision about
whether to follow a hyperlink requires cognitive resource investment and creates germane load when making this decision activates learners’ prior knowledge of the domain and engages
them in deeper semantic processing that encourages restructuring and expanding one’s knowledge base (Niederhauser et al.,
2000).
Although germane load is necessary for learning, managing its
levels is critical to effective instructional design. When learning
activities or materials are not cognitively challenging for the learner, germane load is low, which may result in a decrease in interest
and learning. Conversely, content that is so challenging for the
learner that it lies beyond his or her zone of proximal development
may cause inappropriately high levels of germane cognitive load,
which may also impair learning. Thus, germane load can be managed through selection of tasks and content that is of appropriate
difficulty for the learner.
Cognitive investment that does not contribute to (or interferes
with) understanding of the content to-be-learned is termed extraneous load (Chandler & Sweller, 1992). Potential extraneous load
may be present at several levels in the context of hypertext-assisted learning. Since learning from hypertext involves reading,
extraneous load associated with basic reading processes like
decoding, comprehension monitoring, and situation model development have cognitive consequences for learning. Further, as with
any computer-based information presentation system, extraneous
load may occur when readers lack facility with computers or familiarity with the program, and they must invest cognitive resources
to figure out the basic operation of the computer and software,
which, in the case of hypertext, involves actions like navigating
links, menus, and lists. Negative effects of extraneous load associated with navigating a hypertext can result in a sense of disorientation (Dias & Sousa, 1997; Kim & Hirtle, 1995) and finding oneself
to be ‘‘lost in hyperspace” (Otter & Johnson, 2000; Shapiro &
Niederhauser, 2004).
Intrinsic, germane, and extraneous load are additive (Paas,
Tuovinen, Tabbers, & van Gerven, 2003), and if the sum of the various loads exceeds the learner’s cognitive capacity, learning will be
impaired. Since intrinsic load is thought to be immutable and germane load is necessary for learning, managing cognitive load in
instructional settings involves selecting content that is appropriately challenging for the learner, and structuring the learning environment to reduce extraneous load—thereby freeing cognitive
capacity to allow a greater proportion of germane load to be used
for deeper processing of the information.
142
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
2.2. Working memory limitations and split attention in hypertext
reading
Dynamics of cognitive load are directly influenced by the characteristics of human cognitive architecture (Sweller, 2008). In particular, the limited capacity of working memory may create
challenges for hypertext readers. With print text the author typically tries to present a series of logical internally consistent points
in the text in an effort to conceptually guide the reader. Reading
linked information in hypertext, however, requires the reader to
assume responsibility for developing a coherent representation of
the textbase. It is up to the reader to develop a coherent understanding of the content by integrating information from the text
with prior knowledge, and creating a more sophisticated situation
model. To accomplish this integration, the reader must hold conceptual representations encountered in a given node in working
memory while considering how the information from a new node
might relate. The reader must then decide what to read next
among the available links, and select a new node that will, in the
reader’s opinion, lead to additional promising information. This
process continues until the information is rendered sensible and
the reader’s situation model of the text has been created.
A fundamental problem with processing information in this
manner lies in the fact that human cognitive architecture constrains readers’ ability to optimally integrate concepts from working memory into long-term memory when reading is interrupted
(Glanzer, Fischer, & Dorfman, 1984). For each disruption, critical
text information must be reinstated in working memory if the
reader is to successfully continue the situation model building process. When learning from hypertext, situation model development
is continually disrupted because information is presented in separate nodes, leading to attentional division between concepts from
the current node being held in working memory, and the novel
information encountered in a new node. This split attention effect
clearly has implications for hypertext-assisted learning.
The debate on whether it is possible to attend to multiple
sources of information at once has been of interest to psychologists
since the 1890s (Benjamin, 1968). Cognitive psychologists have
suggested that split attention is caused by intensive search-andmatch processes involved in cross-referencing units of information
in working memory (Chandler & Sweller, 1992). Additional cognitive processing associated with split attention may function as a
source of extraneous load on limited working memory resources,
resulting in acquisition of knowledge fragments instead of coherent knowledge structures (Kalyuga, Ayres, Chandler, & Sweller,
2003).
Physical integration of related information elements has been
proposed as a means to address extraneous load associated with
the split attention problem (Chandler & Sweller, 1992; Mayer &
Anderson, 1991). Taken to an extreme, this would involve the elimination of links to allow spatial integration of the hypertext nodes
into one undivided unit—essentially making it a scrolling print
text. While this might solve the hypertext linking split attention
problem, efforts to reduce extraneous load by using a linear ‘‘linkless” format removes the opportunity to enhance germane load
that is afforded in a hypertext environment. The comparison and
elaboration processes that facilitate complex situation model construction and automation of flexible, adaptive schemas represented in the cognitive flexibility viewpoint are less likely to
occur when reading sequential text. Thus, such conversion would
eradicate the primary advantage of hypertext (a flexible access
information structure) relative to key aspects of both cognitive
flexibility theory and cognitive load theory.
In an effort to make hypertext-assisted learning environments
more comprehensible, designers of hypertext-based learning
materials have proposed learner-support tools like guided tours,
hierarchies, outlines, and graphical overviews, which have been
used with varying success across different types of learners and
content areas (Boechler & Dawson, 2002). Researchers have also
proposed guidelines for developers to help learners recognize the
relationships among ideas scattered across a hypertext (e.g.,
McNamara & Shapiro, 2005; Shapiro, 2008). Although using these
guidelines may help learners develop an overview of the scope
and sequence of the hypertext, these efforts are focused on reducing extraneous load at the global organizational level, providing little support for split attention effects that take place locally–when
linking between individual hypertext nodes. However, navigational decisions are made, and potential situations for split attention occur, when the reader is selecting an embedded link (or
hotword) to a new node. In many cases these hotwords are the primary navigation method available in a hypertext system, and since
they provide little contextual support, there appears to be a need to
provide additional structural cues that address split attention, reduce extraneous load, and support textbase construction in hypertext-assisted learning environments.
2.3. Leads
One mechanism that may reduce split attention and help learners develop better situation models from hypertext is a lead. Initially devised by journalists to provide readers with a preview of
a newspaper article, leads may function as what Ausubel (1960) referred to as an advance organizer. An advance organizer is ‘‘introductory material at a higher level of abstraction, generality, and
inclusiveness than the learning passage itself” (Ausubel, 1978,
p.252). Ausubel’s original research (and many studies thereafter)
studied the use of textual organizers on verbal learning. Ausubel
(1960) asked college students to read a 2500-word passage about
metallurgy after reading either a 500-word advance organizer that
presented the underlying concepts ora 500-word historical passage. The advance organizer group performed better on a subsequent test of retention, presumably because the learners were
able to tie the information to knowledge structures presented in
the organizer. Mayer (1979) reviewed 27 empirical studies that
compared an advance organizer group with a control group, and
found a small but consistent advantage for the advance organizer
participants on tests of retention and learning. This advantage
was more likely when: (a) the content was unfamiliar and (b)
the learners were less experienced in the subject matter.
With respect to hypertext, advance organizers like leads may
serve to orient and prepare the reader for information contained
in the linked node while the current node is still visible. Leads
may be implemented in the form of a mouse-over balloon that
pops up next to the relevant hotword link. Thus, in a lead-augmented hypertext system, the reader has the opportunity to get
an idea of what is coming in the next node without having to leave
the current node.
While the idea of using leads may be relatively new in educational hypertexts, there is evidence that demonstrates their potential value. Empirical research has shown the value of link comments
(which are conceptually equivalent to leads) in several areas. For
example, they have proven useful in hypertext-assisted learning
by improving appropriateness of link selection during a browsing
session (Schweiger, 2001), helping readers integrate textual and
pictorial information (Betrancourt & Bisseret, 1998), by enhancing
searching and knowledge acquisition (Cress & Knabel, 2003), and
supporting navigational decisions (Maes, van Geel, & Cozijn,
2006). Although this research has furthered our understanding of
the potential role of leads in enhancing some aspects of hypertext-assisted learning, we know little about the effects of leads
on helping learners overcome split attention, optimizing cognitive
load, and improving learning outcomes.
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
The present study was designed to examine the effect of leads
on extraneous cognitive load associated with split attention, and
subsequent effects on learning in a conceptually rich hypertext
environment. Subjective, behavioral, and psychophysiological
methods, including self-reported mental effort, reading time, and
electroencephalogram, were employed to assess cognitive load
while participants read traditional and lead-augmented hypertext.
Subjects’ learning performance in the two experimental conditions
was assessed and compared based on measures of recall, domain,
and structural knowledge. This research was driven by the following research questions:
1. What are the effects of leads on cognitive load in a hypertextassisted learning environment?
2. To what extent does the use of leads influence learning
performance?
3. Method
3.1. Subjects
The initial subject pool included 22 teacher education students
from a large Midwestern university. An invitation to participate
was e-mailed to all undergraduate education majors (n = 687). Fifteen dollars and an opportunity to win an iPod Nano™ served as
incentives. The first 22 respondents who met the study criteria
were enrolled as subjects in the study; however, two subjects did
not complete one or more of the experimental tasks and were
dropped from the study, yielding a final subject pool of 20 participants. Because EEG studies of brain wave patterns are conducted
in the within-subject format, where each subject as his or her
own control (Pivik et al., 1993), the sample size is typically limited
to about 10 participants, with the average being 13 (e.g., Gevins
et al., 1998; Kaiser, 1994).
Subject selection was used to help control for some of the variance associated with interpersonal differences. Researchers have
suggested that gender, handedness, and age can differentially affect brain wave activity (Andreassi, 2007; Fisch, 1999). Further,
EEG patterns can be influenced by brain disorders, or medications
to treat brain disorder conditions (Andreassi, 2007). Thus, the subject pool was limited to right-handed, 18–23 year-old females with
no known brain disorders.
Based on an initial demographic questionnaire, all subjects described themselves as native English speakers and experienced
computer and hypertext users. Ten subjects were raised in a family
with low to medium income ($30,000–$60,000), while the remainder came from more affluent families (above $60,000). The sample
consisted of ten seniors, three juniors, two sophomores, and five
freshmen. Most (n = 11) were majoring in Elementary Education.
Of the remainder, five were in Early Childhood Education, two in
secondary Mathematics, one in secondary Biology, and one in Family and Consumer Science. All subjects characterized themselves as
novices with regard to the learning theories presented in the
instructional materials, even though 13 of them had taken an
introductory educational psychology course. Seven subjects indicated that they preferred reading websites to reading books. Eleven
reported taking courses, which required them to read text from
websites at least once a week, seven reported reading course information from websites occasionally, and two indicated they always
printed web-based materials before reading them.
3.2. Materials
Materials for the study included four hypertexts, each on a different learning theory, a hypertext presentation system, and mea-
143
sures to assess reading ability, metacognitive awareness, prior
knowledge, cognitive load, and learning performance.
3.2.1. Learning theory hypertexts
Four instructional hypertexts were developed to present background information on each of four learning theories: Behaviorism,
Information Processing, Cognitive Constructivism, and Social Constructivism. Structure was consistent with what is referred to as
‘‘schema-driven design” (Lawless & Brown, 1997). Each text included a primary passage and seven linked nodes. The primary
passages contained 500 ± 5 words. Each of these passages contained seven links to subordinate nodes describing the theory’s social context, an influential theorist, the view of knowledge inherent
in the theory, and other defining concepts. Nodes contained 90 ± 5
words.
Information Processing and Cognitive Constructivism hypertexts were augmented with leads. Hypertexts on Social Constructivism and Behaviorism did not contain leads. Leads included
10 ± 2 words. To reduce potential text-driven interactions in this
study, experimental texts were normalized and compared by a linguist and four educational psychology experts. All texts were measured and adjusted to have equivalent Flesch reading ease scores of
35, which is considered an appropriate difficulty level for high
school and college students (Flesch, 1948). Conceptual difficulty
equivalence and content validity were established through a
read-and-revise process by four educational psychology experts
and a linguist who reviewed and corrected the texts to ensure that
(a) the same number of concepts (n = 7) were covered in the linked
nodes of each text, (b) the concepts were of the same level of domain-specificity (e.g., influential scholar, view of knowledge), and
(c) identical grammatical and syntactical structures were used
across all texts (e.g., sentence length, number of clauses, stylistic
devices).
3.2.2. Hypertext presentation system
The hypertext presentation system was designed using guidelines suggested by current web usability research (Nielsen, 2006).
Nodes were presented as a single frame using Georgia serif font,
with black lettering on a white background and no tracing images
or watermarks. Font size was set to 100 percent of the default
browser font size, which translated to approximately 16 point font
on the monitor used for the study. Three styles were used to identify links to subordinate nodes: (a) blue, underlined for all new,
non-active links; (b) purple, underlined for all visited, non-active
links; and (c) red, underlined for active links. Both primary passages and subordinate nodes included a banner indicating the
name of the learning theory, and the informational text in the body
of the frame. A sample primary passage, with rollover lead activated, can be seen in Fig. 1.
Four restrictions were used to guide the development of the
hypertext system to avoid contamination from variables other than
those of interest in the present study: (a) no images were used to
remove variance associated with media other than text; (b) only
contextual (embedded hotword) links were available as a navigation method to remove variance associated with different navigational options (e.g., menus, maps, lists etc.); (c) no links to
external websites were included to limit access to information;
(d) all new nodes opened in the same browser window, so subjects
could view only one informational frame at a time.
Leads were implemented as mouse-over popup balloons which
introduced the content that would be presented in each of the seven subordinate nodes in the given passage (see Fig. 1). Presentation of leads was based on a mouse-over event (rather than a
mouse click). Subjects were able to hover the cursor over the link
and read the lead, before deciding whether to click on the link
and proceed to the subordinate node. Balloons covered the
144
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
Fig. 1. Screenshot of the primary passage of a lead-augmented hypertext system.
remainder of the current line of text and part of the two lines of
text following the link, essentially preventing subjects from continuing to read the text that followed the hotword while the popup
was active. Clicking the hotword accessed the subordinate node
using the same window, and a return link in the subordinate node
took the reader back to the primary passage.
3.3. Measures
3.3.1. Individual difference measures
3.3.1.1. Demographic questionnaire. Subjects completed a 17-item
demographic questionnaire to provide information on major, year
in school, and whether they had taken any classes in educational
psychology. They also provided information on their experience
and confidence in using computers and the Internet, and their habits and preferences associated with reading texts from books or
websites.
3.3.1.2. Hypertext navigation and reading time. Navigation path and
reading time data were recorded using Camtasia™ screen-capture
software. Each time a subject clicked on a link in the traditional
hypertexts—or clicked a link after reading a lead in the lead-augmented hypertexts—a marker was placed on the EEG recording.
Subjects were instructed to read each hypertext at their normal
reading pace, and to keep reading and reviewing until they felt that
they thoroughly understood the content. Total reading time was
computed for each hypertext, and amount of time spent reading
in the lead and no-lead conditions served as the reading time
measure.
3.3.1.3. Reading comprehension. Subjects’ ability to understand
what they read was assessed using the Nelson-Denny advanced
reading comprehension test, Form E (Brown, Bennett, & Hanna,
1981). Subjects read eight passages on various topics and answered a total of 24 five-option multiple-choice questions about
information presented in the passages.
3.3.1.4. Domain knowledge pretest. Prior knowledge of the educational psychology content was assessed using a four alternative
12-item multiple-choice test—three questions for each of the four
learning theories presented in the hypertext. For example, subjects
were asked:
How do behaviorists believe we learn?
a. Associations between stimuli and
automatic.
b. Through fear of punishment.
c. By making sense of experiences.
d. From trial and error.
responses
become
3.3.1.5. Metacognitive awareness. Metacognitive awareness was assessed using the Metacognitive Awareness Inventory (Schraw &
Dennison, 1994). In this instrument items were classified into
two metacognitive factors–knowledge of cognition and regulation
of cognition. Knowledge of cognition included items like ‘‘I know
what kind of information is most important to learn,” and ‘‘I am
a good judge of how well I understand something;” while regulation of cognition included items like ‘‘I think about what I really
need to learn before I begin a task,” and ‘‘I find myself pausing regularly to check my comprehension.” Subjects responded to the 52
items using a True/False response scale. In a study analyzing use of
comprehension aids in a hypermedia environment with 116 college students, Cronbach’s alpha for the MAI was estimated at
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
0.86–indicating acceptable internal consistency (Bendixen & Hartley, 2003).
3.3.2. Cognitive load measures
3.3.2.1. Self-report of mental effort. The subjective measure of selfreported mental effort used in this experiment was a version of
the Bratfisch, Borg, and Dornic (1972) scale for measuring perceived task difficulty as modified by Paas (1992). Numerical values
and labels assigned to the categories ranged from very, very low
mental effort (1) to very, very high mental effort (9). Reliability of
the scale in previous research with college students was estimated
at 0.90 using Cronbach’s coefficient alpha (Paas & van Merriënboer,
1994). The scale was discussed with the subjects before beginning
experimental sessions. Average subject ratings of mental effort for
lead-augmented and non-lead-augmented hypertext passages
served as the measure of self-report of mental effort.
3.3.2.2. Electroencephalogram. EEG data were acquired using a Biopac MP30 connected to a Macintosh G4 MiniMac computer. Electrode placement was confined to one set of electrodes over the
pre-frontal cortex (F7) and one-over the parietal lobe (P3) in the
left hemisphere of the right-handed subjects. EEG data were collected at a sampling rate of 500 Hz as each subject read each of
the four hypertexts. Electrode placement followed the Modified
Combinatorial Nomenclature expanded 10–20 system, as proposed
by the American Clinical Neurophysiology Society (Jasper, 1958).
The EEG software recorded brain wave rhythms as separate channels, allowing identification of the following wave components: (a)
raw EEG signal, (b) alpha rhythm, (c) beta rhythm, and (d) theta
rhythm.
Event-Related Desynchronization percentage (ERD%) for alpha,
beta, and theta rhythms were used as online measures of brain
activity (Pfurtscheller & Lopes de Silva, 1999). Increased cognitive
load is associated with higher brain wave desynchronization for alpha and beta rhythms, and higher brain wave synchronization for
the theta rhythm, when subjects move from a relaxed, eyes-closed
state (baseline) to an eyes-open, active-reading state (Basar, 2004;
Klimesch, 2005). Consequently, ERD%, which compares brain wave
power in the test condition with the brain wave power in the baseline condition, is represented by a positive number for the subjects’
alpha and beta rhythms (reflecting wave desynchronization), and a
negative number for the theta rhythm (reflecting synchronization).
Thus, larger desynchronization percent values for alpha and beta
waves, and larger synchronization percent values for theta waves
indicate increased cognitive load.
The following formula was used to compute ERD% (Pfurtscheller
& Lopes de Silva, 1999):
ERD% ¼
baseline intrerval band power test interval band power
baseline intrerval band power
100
Band power values of the subject’s alpha, beta, and theta brain
waves were estimated with the Biopac psychophysiometrical system software. The Area function was used to calculate alpha, beta,
and theta wave power. This function computes the total area of
waveform under a straight line drawn between the endpoints.
Thus, this measure takes into account both wave frequency (Hz)
and amplitude (lV). Area under the curve for alpha, beta, and theta
brain wave rhythms was computed for a 20-s segment of the baseline condition, which was obtained before a given subject started
reading each of the four texts. This yielded a unique baseline interval band power of brain activity for the time period immediately
preceding reading each text that was used in the ERD% formula.
Area was then computed for the test interval band power. Using
the marker placed on the EEG recording as a referent, a 20-s test
145
interval was established which included the 10 s spent reading the
primary passage immediately before selection of given link to a subordinate node, and the 10 s spent reading immediately after clicking
the link to the subordinate node. Test interval band power was calculated using this 20-s interval because subjects read and processed
lead information before they clicked on the link to open the subordinate node in the lead-augmented hypertext condition. ERD% values were computed for alpha, beta, and theta brain wave rhythms
on each of the seven links within each of the four hypertexts.
Average ERD% value was then computed for each of the three
brain wave rhythms under each of the two experimental conditions. For example, the ERD% values for the 20-s intervals for the
alpha rhythm were averaged for each subject, yielding one alpha
ERD% value for each text. The alpha ERD% values for the two nolead texts were then averaged, as were the alpha ERD% values for
the two lead-augmented texts. This yielded one grand alpha
ERD% value for the lead-augmented hypertext condition, and a second alpha ERD% value for the no-lead condition. This procedure
was then used to compute grand averages for lead and no-lead
conditions for beta and theta rhythms. Since cognitive load is reflected by more negative ERD% for the theta rhythm (Klimesch,
2005), and more positive ERD% for alpha and beta rhythms, we
used absolute values for measures of theta waves to help make
interpretation of results more intuitive. Grand average ERD% values
for alpha, beta, and theta rhythms served as dependent measures.
3.3.3. Learning measures
3.3.3.1. Recall. Prompted verbal recall was used to assess what students remembered after reading the texts. Each subject was shown
a visual prompt containing the name of a learning theory and
asked to verbally recount everything she could remember about
that text. This procedure was repeated with the remaining text
topics, providing an opportunity for subjects to recall information
about each theory in the order in which they had been presented.
Student responses were digitally recorded and audio files were
transcribed. Recall was determined by counting the number of discrete concepts subjects mentioned for each theory. Proportion of
concepts recalled served as the recall measure.
3.3.3.2. Structural knowledge. Subjects’ structural knowledge was
assessed using a concept-sorting task. Each subject was presented
with an Inspiration™ concept-mapping software template containing (a) columns with names of the four learning theories and (b) 28
key concepts (one concept for each of the seven subordinate node
for each of the four texts) presented in a randomized cluster on the
bottom of the screen. Subjects were instructed to drag and drop
each of the 28 concepts into the appropriate learning theory column. Proportion of correctly placed concepts served as the structural knowledge measure.
3.3.3.3. Domain knowledge post-test. Domain knowledge was assessed using a multiple-choice test. Twenty-eight new questions
(seven per text) were added to the twelve pretest items to create
the 40-item four-alternative domain knowledge post-test. Randomly presented items were distributed evenly across the four
texts to assess content presented in both primary passages and
subordinate nodes. Post-test items were identical to pretest items
in form. Proportion of correct responses served as the domain
knowledge measure.
3.4. Procedure
3.4.1. Pre-treatment phase
The entire group of subjects attended one initial session as a
group before participating in the treatment phase of the study.
Those who wore corrective lenses were instructed to wear them
146
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
when participating in all research activities. Each subject sat at a
desk with a pencil and the reading comprehension measure, domain knowledge pretest, metacognitive awareness inventory, and
demographic questionnaire placed on the writing surface in the order in which they were to be completed. A researcher read a prepared script to direct the students through the activities and
answered any questions that arose. Each subject then scheduled
an individual time to complete the treatment phase during the
immediately following 2-week period.
open her eyes and read the second text, complete the mental effort
scale, and close her eyes until she returned to the baseline condition. This procedure was repeated for the remaining two texts.
The sequence of presenting texts to subjects was counterbalanced:
lead/no-lead/lead/no-lead for half the subjects, and no-lead/lead/nolead/lead for the remainder. When a subject had finished reading
the fourth hypertext, she was instructed to close her eyes to return
to an alpha state, which provided an end point for the brain wave
recording.
3.4.2. Treatment phase
The treatment phase took place in a 12 by 16 foot physiological
testing laboratory. The lab was located in a quiet area removed
from hallways and had a heavy door to guard against distracting
audio stimuli. All walls were painted white and the ceiling was a
white drop-panel design with recessed fluorescent lights. Each
lighting fixture contained three bulbs on separate switches. All
room lights were on during treatment sessions and no external
light was allowed to enter the room. This laboratory setting provided consistent environmental conditions for all subjects.
Subjects participated in the treatment phase one at a time. Each
sat at a table with an adjustable-height high-back chair with armrests. Subjects faced a blank white wall, which extended beyond
peripheral vision on both sides. A full-size 110-key keyboard, wireless mouse, and 17-inch LCD monitor were on the table in front of
the subject. The same computer and monitor were used for all
subjects.
The monitor surface was approximately 50 cm from the subject,
and was set to 1280 by 854 pixels (the highest available resolution)
and maximum level of brightness. Luminescence was measured at
eye-level and at a distance of approximately 50 cm from the computer screen using a Tenma Digital Lux meter. Luminosity of
hypertext displayed on the LCD monitor was equal across text
conditions.
With the subject comfortably seated at the computer, the researcher read a scripted verbal overview of treatment procedures
to begin the session. He then attached disposable vinyl electrodes
(Ag/AgCl) to two recording sites on the subject’s skull: (a) pre-frontal lobe (F7) to collect data on the power of beta and theta waves
and (b) parietal lobe (P3) to collect data on the power of alpha
waves. Measurements were referenced to the left mastoid with
the earlobe serving as ground. Electrode impedance was below
10 kX. The EEG signal passed through an Infinite Impulse Response
(IIR) bandpass filter to remove unintended artifacts of movement,
allowing us to retain only the frequency components that were
of interest in the present study: theta (4–7 Hz), alpha (8–13 Hz),
and beta (14–30 Hz). Two more sets of electrodes were attached
to collect the subject’s electrooculogram and electromyogram of
the dominant (right) hand, which was used to filter artifacts associated with eye movement, blinking, hand movement and mouse
clicking. Subjects were instructed to minimize unnecessary movement during the hypertext reading and browsing task.
After all electrodes were placed and the EEG equipment was
activated, the subject sat in a relaxed state with her eyes closed until an extended alpha pattern was noted. At that point the 20-s
baseline brain wave rhythm sample was recorded and the subject
was instructed to open her eyes (blocking the alpha rhythm) and
read the first experimental hypertext. When the subject had finished reading she said ‘‘done,” completed a self-report of mental
effort measure, and closed her eyes and relaxed until her brainwave patterns returned to the baseline condition (extended alpha).
Returning to baseline helped circumvent carryover effects from
one treatment to the next.
The researcher brought up the next experimental text while the
subject was returning to baseline. After establishing and recording
the next baseline brain wave sample, the subject was instructed to
3.4.3. Post-treatment phase
When the subject had finished reading all four texts, she was
asked to complete four ‘‘word jumble” anagrams as an interpolated
task to clear working memory. Learning performance was then assessed in the following sequence: free recall task, structural knowledge concept-sorting task, and domain knowledge multiple-choice
post-test. Subjects were then paid $15 and allowed to leave.
4. Results
We employed a repeated measures design, enabling each subject to serve as her own experimental control. This design, however, obviated the use of covariance to control for systematic
between-subject differences. We used subject selection (as described in the Section 3) to help control some differences, and report here variance associated with prior domain knowledge
(M = 0.91, SD = 0.08), reading ability (M = 0.40, SD = 0.07), and
metacognitive awareness (M = 0.74, SD = 0.11)—three interpersonal factors that have been associated with hypertext-assisted
learning (Shapiro & Niederhauser, 2004). Small standard deviations
indicate that subjects were fairly similar across these measures.
Results of this study are presented in the following sections: (a)
cognitive load, (b) learning performance, and (c) hypertext browsing behavior.
4.1. Cognitive load
A 2 5 repeated measures MANOVA was conducted to determine the effect of leads on learners’ cognitive load. Presence of
leads (Lead vs. No-lead) served as a within-subject factor, and
the five measures of cognitive load: (a) reading time; (b) self-reported mental effort; and Event-Related Desynchronization percentages of (c) alpha; (d) beta, and (e) theta brain wave rhythms
were used as dependent measures.
The MANOVA for cognitive load measures was significant
(F(5,12) = 58.94, p < 0.01), prompting further analysis through a series of univariate repeated measure ANOVAs. For each of the ensuing ANOVAs, presence of leads served as the independent variable,
with each of the cognitive load measures serving as a dependent
measure. A main effect was found for reading time (F(1,16) = 5.55,
p < 0.05, MSE = 427.63). Subjects spent more time reading in the
lead condition than in the no-lead condition (X = 587.23,
SD = 116.06 and X = 571.83, SD = 126.20 s, respectively). Main effects were also found for alpha, beta, and theta ERD%
(F(1,16) = 103.47, p < 0.01, MSE = 7.12; F(1,16) = 35.71, p < 0.01,
MSE = 15.34; F(1,16) = 252.56, p < 0.01, MSE = 1.03, respectively). Table 1 shows that mean alpha, beta, and absolute value of theta
ERD% in the no-lead condition was higher than in the lead condition. These findings reveal lower cognitive load in the Lead condition. No other results reached significance (p > 0.08).
4.2. Learning performance
A 2 3 repeated measures MANOVA was conducted to determine the effect of leads on subjects’ learning performance.
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
Table 1
Descriptive statistics for measures of cognitive load.
Alpha ERD%
M
SD
Beta ERD%
Theta ERD%
No-lead
Lead
No-lead
Lead
No-lead
Lead
44.67
9.12
36.08
9.24
33.67
8.52
26.27
7.87
16.90
3.43
11.80
3.55
Presence of leads (Lead vs. No-lead) was used as a within-subjects
factor, and the three measures of learning: (a) recall, (b) structural
knowledge, and (c) domain knowledge served as dependent
measures.
The MANOVA for learning was significant (F(3,14) = 6.13,
p < 0.05), prompting further analysis through a series of univariate
repeated measure ANOVAs. For each of the ensuing ANOVAs, presence of leads served as the independent variable, and each of the
three learning performance measures served as a dependent
variable.
Main effects were found for both structural knowledge and domain knowledge (F(1,16) = 9.26, p < 0.01, MSE = 0.10; and
F(1,16) = 5.79, p < 0.05, MSE = 0.12, respectively). Higher means for
learning in the Lead condition on structural and domain knowledge measures indicate that leads produced a positive effect on
these indices of learning performance (see Table 2). No other results reached significance (p > 0.2).
4.3. Hypertext browsing behavior
As reported earlier, the reading time analysis showed that subjects spent more time reading the hypertexts that were augmented
with leads. Analysis of the screen-capture data revealed that on
average subjects spent 6.77 s reading the leads (SD = 1.47), based
on the first time they accessed the lead. Fourteen subjects only
opened the subordinate nodes one time during their initial reading
of the text, following the order of link presentation from top to bottom of the main node and then exited from the browser. The
remaining six subjects exhibited a different reading behavior.
When reading and browsing traditional hypertext without leads,
they first read the entire text in the main node, then went back
and accessed subordinate nodes nonlinearly. In the lead-augmented hypertext condition, these students accessed the leads
sequentially, processing the information in the leads one by one,
and then returning to the links and opening the linked nodes in
the order that they were presented on the screen. Before exiting
from the browser, these subjects reviewed the main node, and
re-read each of the leads without actually clicking the link to access the subordinate node.
Bivariate Pearson product-moment correlation analysis was
used to examine this browsing behavior. A positive relationship
was found between subjects’ metacognitive awareness scores
and whether they used leads to review content in the subordinate
nodes (r = 0.63, p < 0.01). This finding suggests that subjects with
better metacognitive skills tended to use leads as a tool to review
information and regulate their cognition in the lead-augmented
hypertext.
Table 2
Descriptive statistics for measures of learning performance.
M
SD
Structural knowledge
Domain knowledge
No-lead
Lead
No-lead
Lead
0.53
0.18
0.63
0.23
0.58
0.18
0.66
0.14
147
A battery of Pearson correlation analyses was conducted to
determine whether the effects on cognitive load and learning performance were due to the increased reading time in the lead condition. First, the researchers explored the relationship between
reading time in the lead condition and the results of each cognitive
load and learning performance measure in the lead condition.
None of the results reached significance (r < 0.40). Another battery
of correlation analyses was conducted for reading time in the nolead condition and each cognitive load and learning performance
dependent variable in the no-lead condition, yielding no significant
results. No significance was also observed when running Pearson
correlation tests for reading time and each of the individual differences variables (i.e., prior knowledge, metacognitive awareness,
and reading ability). Non-significant correlations between reading
time and results of cognitive load and learning performance measures indicate that increased reading time in the lead condition
was not related to the participants’ learning performance scores
and cognitive load. A similar use of correlation analyses to compare
the utility of several cognitive load measures was reported recently
by DeLeeuw and Mayer (2009).
5. Discussion
The aim of this study was to determine the influence of leads on
cognitive load and learning in hypertext. The study produced several important findings. First, EEG-based cognitive load measures
showed that subjects’ brain wave activity was less intense when
they were accessing hypertext nodes via leads. Conversely, the
self-report of mental effort measure did not detect significant differences in cognitive load between the two experimental conditions, and subjects tended to spend more time reading leadaugmented hypertext. Second, use of leads appeared to produce a
positive effect on learning outcomes relative to domain and structural knowledge acquisition.
The discrepancy between the results of EEG-based measures of
cognitive load, self-report of mental effort, and the reading time
measure prompted further examination of the literature on cognitive load measurement. As a result, we offer two possible explanations as to why the results of cognitive load measures used in this
study were inconsistent. These measures may vary in their facility
to assess (a) temporal dimensions of cognitive load, and (b) contributions of the three structural load types (intrinsic, extraneous,
and germane).
5.1. Effects of leads relative to temporal differentiation of cognitive
load
Cognitive load has been described as a dynamic process, rather
than a static property of task situations (Madni & Lyman, 1983).
Thus, results of cognitive load measurement should be interpreted
in the context of temporal dimensionality of this cognitive process.
According to Xie and Salvendy (2000), cognitive load should be defined as a multi-dimensional construct that represents instantaneous load, peak load, average load, accumulated load, and
overall load. Further, a meaningful interpretation of cognitive load
should involve the associated performance level, learning effort,
test effort, test performance, time on task, and integration of other
measures of learning and cognitive load that are relevant in a given
context (Paas et al., 2003).
Since the EEG-based measures addressed the 20-s period of
time when subjects were following links or leads to a new hypertext node, they assessed instantaneous cognitive load. Measurement of instantaneous load was of particular interest in this
study because it reflected subtle fluctuations in brain activity that
were presumably associated with the negative effects of split
148
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
attention during hypertext reading. During these 20-s time periods, brain activity was heightened in the no-lead, traditional
hypertext condition, as reflected by uniformly higher alpha, beta,
and theta ERD% values. While following links in the no-lead hypertext may have resulted in an interruption of conceptual flow and
mental separation of content from two or more nodes, leads appeared to help learners overcome split attention, and use the
now-available cognitive resources for more effective conceptual
integration of the new content. This conclusion is corroborated
by the results of the tests of learning.
Unlike the EEG-based measures, which appeared to be useful
for detecting instantaneous cognitive load, self-reported mental effort may serve as an index of overall cognitive load. Overall cognitive load reflects the individual’s judgment of cognitive investment
based on the whole task. In the present study, the self-report measure was completed after reading each hypertext. This summative
rating of the mental effort invested in reading the entire hypertext
was necessary because administrating the self-report instrument
every time subjects accessed a link or lead would have disrupted
and altered the primary cognitive activity-developing the situation
model of the text. Thus, while leads may have reduced split attention and the resulting extraneous load during the 20 s when subjects opened new nodes (according to the EEG data), subjects did
not appear to perceive this as a reduction in the overall cognitive
investment in reading the entire hypertext.
5.2. Effects of leads relative to structural differentiation of cognitive
load
Considering cognitive load theory’s distinction of intrinsic,
extraneous, and germane load, it is important to note that, to date,
researchers have been unable to advance the theory by using any
of the known measurement techniques to differentiate between
the three cognitive load components. Intrinsic, extraneous, and
germane load are additive in nature, and structurally the sum of
contributions of each load type makes up the total load imposed
by a learning material or activity. Further, the main distinction between germane load and extraneous load is that while the former
is associated with improved cognition and learning, the latter hinders comprehension due to limitations of content presentation.
Keeping this distinction in mind, we propose that the contributions
of germane and extraneous load can be deduced by analyzing results of cognitive load measures relative to learning performance.
The results of learning assessments used in this study demonstrate better domain and structural understanding of the content
in the lead-augmented hypertext condition. Learning from texts
with leads allowed subjects to perform better at identifying the
meaning of complex learning theory concepts and sorting concepts
based on their associations with specific learning theories. It is
likely that leads decreased the semantic space between linked
nodes and improved the conceptual integration of information,
which ultimately had a positive effect on the acquisition of domain
and structural knowledge.
Since subjects exhibited better learning performance after reading lead-augmented hypertext, decreased brain activity revealed
by the EEG measures in this experimental condition may signify
a reduction in extraneous load. This result is consistent with prior
research (Chandler & Sweller, 1992), which indicates that split
attention results in extraneous load rather than germane or intrinsic load. In the present study, leads were introduced in instructional hypertext to reduce the amount of detrimental searchand-match processes that can cause extraneous cognitive load,
and therefore it is likely that learners in the lead-augmented condition acquired more coherent knowledge structures. This interpretation is consistent with the construction integration model
(Kintsch, 1988), which posits that the mere presence of links will
not aid reading comprehension because links do not help relate
text to the situation model. This model further predicts that labeled
links (like leads), which indicate the type of information available
in the linked node, might support development of the situation
model because labels provide a clue to the reader that the linked
information is relevant and should be incorporated in the current
situation model, or alternatively, that the link leads to a new topic
and thus requires development of a new situation model (DeStefano & LeFevre, 2007).
Reading time data revealed that subjects took more time to process hypertext that was augmented with leads. While the time
necessary to process the lead (the average of seven seconds) certainly explains that there was more content to read in the leadaugmented condition, another possible explanation is that longer
reading times in this condition may indicate increased levels of
engagement with the content and greater germane load. If leads reduced the negative effects of split attention and optimized extraneous load during the 20 s of accessing leads and initial processing of
the new node, the learners’ cognitive capacity was effectively increased to allow for deeper integration of new information on
the germane level after these 20 s expired. More cognitive resources were then available, enabling learners to assimilate or
accommodate new concepts with the existing knowledge base
and update the current situation model of the text. Thus, it seems
likely that node summaries contained in the leads encouraged
learners to relate new information to their prior knowledge and allowed activation of relevant knowledge representations from longterm memory, which resulted in increased germane processing
and longer reading times for the entire hypertext.
Finally, structural differentiation of cognitive load can inform
analysis of the results of self-reports, which showed no effect of
leads on cognitive load. To isolate the potential effect of leads, texts
were normalized so they were the same level of complexity, and all
participants had little prior knowledge of the domain. Lack of significant differences on self-reported mental effort supports that,
overall, the four texts were equivalent in terms of conceptual difficulty, supporting the normalization efforts of the educational psychology experts. Hence, from a structural point of view, the lack of
differences in cognitive load as reflected by self-reports might
actually indicate no differences in intrinsic load. The type of self-report of mental effort measure used in this study seems to rely on
the intrinsic difficulty of the learning materials, and is, in fact, a
modification of the Bratfisch et al. (1972) scale that was developed
to measure perceived task difficulty.
6. Conclusion
Results of this study have led us to argue that single-attribute
cognitive load assessment, such as the widely used practice of
measuring overall load with self-reports may not be adequate
for systematically describing the causes and effects of cognitive
load. Instead, cognitive load should be conceptualized as a dynamic process and assessed using a comprehensive analytical
framework that integrates measures to target both the temporal
dimensions of cognitive load like instantaneous load, peak load,
average load, accumulated load, and overall load, and the contributions of structural load including intrinsic, extraneous, and germane load.
Findings build on prior research on the use of node previews
and similar tools. Earlier studies reported a positive effect of
lead-like features on improving learning in a search task (e.g.,
Betrancourt & Bisseret, 1998), and a knowledge acquisition task
(e.g., Cress & Knabel, 2003; Schweiger, 2001). Our findings showed
a similar effect of leads on acquisition of domain and structural
knowledge and extended our understanding of this phenomenon
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
by using EEG-based measures to reveal decreased brain activity
that may be associated with split attention and extraneous load
when subjects were accessing new nodes that included leads.
These findings suggest that leads appear to be a potentially useful
tool for reducing extraneous load caused by split attention, and
improving structural and domain learning. Insights gained from
this work help us better understand cognitive processes associated
with hypertext-assisted learning, and can help guide the development of these kinds of systems.
References
Alexander, P. A., Kulikowich, J. M., & Jetton, T. L. (1994). The role of subject matter
knowledge and interest in the processing of linear and non-linear texts. Review
of Educational Research, 64, 201–252.
Andreassi, J. (2007). Psychophysiology: Human behavior and physiological response.
Mahwah, NJ: Lawrence Erlbaum Associates.
Ausubel, D. P. (1960). The use of advance organizers in the learning and retention of
meaningful verbal material. Journal of Educational Psychology, 51, 267–272.
Ausubel, D. (1978). In defense of advance organizers: A reply to the critics. Review of
Educational Research, 48, 251–257.
Azevedo, R., & Jacobson, M. (2008). Advances in scaffolding learning with hypertext
and hypermedia: a summary and critical analysis. Educational Technology
Research and Development, 56, 93–100.
Barab, S. A., Young, M. F., & Wang, J. (1999). The effects of navigational and
generative activities in hypertext learning on problem solving and
comprehension. International Journal of Instructional Media, 26(3), 283–309.
Basar, E. (2004). Memory and brain dynamics: Oscillations integrating attention,
perception, learning and memory. CRC Press LLC.
Bendixen, L. D., & Hartley, K. (2003). Successful learning with hypermedia: The role
of epistemological beliefs and metacognitive awareness. Journal of Educational
Computing Research, 28, 15–30.
Benjamin, W. (1968). The work of art in the age of mechanical reproduction. In
Illuminations (pp. 219–253). New York: Harcourt, Brace and World.
Betrancourt, M., & Bisseret, A. (1998). Integrating textual and pictorial information
via pop-up windows: An experimental study. Behavior and Information
Technology, 17, 263–273.
Boechler, P. M., & Dawson, M. R. W. (2002). The effects of navigation tool
information on hypertext navigation behavior: A configural analysis of pagetransition data. Journal of Educational Multimedia and Hypermedia, 11(2),
95–115.
Bratfisch, O., Borg, G., & Dornic, S. (1972). Perceived item-difficulty in three tests of
intellectual performance capacity. Reports from the institute of applied psychology
(Vol. 29). Stockholm: University of Stockholm.
Brown, J. I., Bennett, F. M., & Hanna, G. (1981). The Nelson Denny reading test (form E).
Chicago: Riverside Publishing.
Calandra, B., & Barron, A. E. (2005). A preliminary investigation of advance
organizers for a complex educational website. Journal of Educational
Multimedia and Hypermedia, 14(1), 5–23.
Chandler, P., & Sweller, J. (1992). The split-attention effect as a factor in the design
of instruction. British Journal of Educational Psychology, 62, 233–246.
Charney, D. (1994). The effect of hypertext on processes of reading and writing. In C.
L. Selfe & S. Hilligoss (Eds.), Literacy and computers: The complications of teaching
and learning with technology (pp. 238–263). New York: Modern Language
Association.
Cress, U., & Knabel, O. B. (2003). Previews in hypertext: effects on navigation and
knowledge acquisition. Journal of Computer-Assisted Learning, 19, 517–527.
DeLeeuw, K., & Mayer, R. (2009). A comparison of three measures of cognitive load:
Evidence for separable measures of intrinsic, extraneous, and germane load.
Journal of Educational Psychology, 100(1), 223–234.
DeStefano, D., & LeFevre, J. (2007). Cognitive load in hypertext reading: A review.
Computers in Human Behavior, 23, 1616–1641.
Dias, P., & Sousa, P. (1997). Understanding navigation and disorientation in
hypermedia learning environments. Journal of Educational Multimedia and
Hypermedia, 6(2), 173–185.
Eveland, W. P., Cortese, J., Park, H., & Dunwoody, S. (2004). How web site
organization influences free recall, factual knowledge, and knowledge structure
density. Human Communication Research, 30(2), 208–233.
Fisch, B. (1999). Fisch and Spehlmann’s EEG primer: Basic principles of digital and
analog EEG. Amsterdam: Elsevier.
Flesch, R. (1948). A new readability yardstick. Journal of Applied Psychology, 32,
221–233.
Gevins, A., Smith, M. E., Leong, H., McEvoy, L., Whitfield, S., Du, R., et al. (1998).
Monitoring working memory load during computer-based tasks with EEG
pattern recognition methods. Human Factors, 40, 79–91.
Glanzer, M., Fischer, B., & Dorfman, D. (1984). Short-term storage in reading. Journal
of Verbal Learning and Verbal Behavior, 23, 467–486.
Jasper, H. A. (1958). The ten–twenty system of the International Federation.
Electroencephalography and Clinical Neurophysiology, 10, 371–375.
Jonassen, D. H. (2000). Computers as mindtools for schools: Engaging critical thinking.
Columbus, OH: Prentice-Hall.
149
Jonassen, D. H., & Wang, S. (1993). Acquiring structural knowledge from
semantically structured hypertext. Journal of Computer-Based Instruction,
20(1), 1–8.
Kaiser, D. A. (1994). Interest in films as measured by subjective and behavioral ratings
and topographic EEG. Los Angeles: University of California.
Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The expertise reversal effect.
Educational Psychologist, 38(1), 23–31.
Kim, H., & Hirtle, S. C. (1995). Spatial metaphors and disorientation in hypertext
browsing. Behaviour and Information Technology, 4(14), 239–250.
Kintsch, W. (1988). The role of knowledge in discourse comprehension: A
construction-integration model. Psychological Review, 95, 163–182.
Klimesch, W. (2005). The functional significance of theta and upper alpha
oscillations. Experimental Psychology, 52(1), 1–10.
Landow, G. P. (1992). Hypertext: the convergence of contemporary critical theory and
technology. Baltimore and London: Johns Hopkins University Press.
Lawless, K. A., & Brown, S. W. (1997). Multimedia learning environments: Issues of
learner control and navigation. Instructional Science, 25, 117–131.
Madni, A. M., & Lyman, J. (1983). Model-based estimation and prediction of
task-imposed mental workload. In Proceedings of 1983 IEEE international conference on systems, man, and cybernetics symposium. Norfolk, VA.
Maes, A., van Geel, A., & Cozijn, R. (2006). Signposts on the digital highway: The
effect of semantic and pragmatic hyperlink previews. Interacting with
Computers, 18(2), 265–282.
Mayer, R. L. (1979). Can advance organizers influence meaningful learning? Review
of Educational Research, 49, 371–383.
Mayer, R. E., & Anderson, R. B. (1991). Animations need narrations: An experimental
test of a dual coding hypothesis. Journal of Educational Psychology, 83(4),
484–490.
McNamara, D. S., & Shapiro, A. M. (2005). Multimedia and hypermedia solutions for
promoting metacognitive engagement, coherence, and learning. Journal of
Educational Computing Research, 33(1), 1–29.
Niederhauser, D. (2008). Educational hypertext research. In J. M. Spector, M. D.
Merrill, J. van Merrienboer, & M. P. Driscoll (Eds.), Handbook of research on
educational communications and technology (3rd ed., pp. 199–210). New York:
Lawrence Erlbaum Associates.
Niederhauser, D. S., Reynolds, R. E., Salmen, D. J., & Skolmoski, P. (2000). The
influence of cognitive load on learning from hypertext. Journal of Educational
Computing Research, 23(3), 237–255.
Nielsen, J. (2006). Prioritizing web usability. Berkeley, CA: New Riders Press.
Otter, M., & Johnson, H. (2000). Lost in hyperspace: Metrics and mental models.
Interacting with Computers, 13(1), 1–40.
Paas, F. (1992). Training strategies for attaining transfer of problem-solving skill in
statistics: A cognitive-load approach. Journal of Educational Psychology, 84,
429–434.
Paas, F., Tuovinen, J., Tabbers, H., & van Gerven, P. W. M. (2003). Cognitive load
measurement as a means to advance cognitive load theory. Educational
Psychologist, 38, 63–72.
Paas, F., & van Merriënboer, J. J. G. (1994). Instructional control of cognitive load
in the training of complex cognitive tasks. Educational Psychology Review, 6,
51–71.
Pfurtscheller, G., & Lopes de Silva, F. H. (1999). Event-related EEG/MEG
synchronization
and
desynchronization:
Basic
principles.
Clinical
Neurophysiology, 110, 1842–1857.
Pivik, R. T., Broughton, R. J., Coppola, R., Davidson, R. J., Fox, N., & Nuwer, M. R.
(1993). Guidelines for the recording and quantitative analysis of
electroencephalographic activity in research contexts. Psychophysiology, 30,
547–558.
Rada, R. (1989). Writing and reading hypertext: An overview. Journal of the American
Society for Information Science, 40, 164–171.
Reynolds, R. E. (2000). Attentional resource emancipation: Toward understanding
the interaction of word identification and comprehension processes in reading.
Scientific Studies of Reading, 4(3), 169–195.
Rouet, J.-F., Levonen, J. J., Dillon, A., & Spiro, R. J. (1996). Hypertext and cognition.
Mahweh, NJ: Lawrence Erlbaum.
Schraw, G., & Dennison, R. S. (1994). Assessing metacognitive awareness.
Contemporary Educational Psychology, 19(4), 460–475.
Schweiger, W. (2001). Hypermedien im Internet. München: Verlag Reinhard
Fischer.
Shapiro, A. (2008). Hypermedia design as learner scaffolding. Educational Technology
Research and Development, 56, 29–44.
Shapiro, A., & Niederhauser, D. (2004). Learning from hypertext: Research issues
and findings. In D. Jonassen (Ed.), Handbook of research for educational
communications and technology (2nd ed., pp. 605–620). Lawrence Erlbaum
Associates.
Spiro, R. J., Coulson, R. L., Feltovich, P. J., & Anderson, D. K. (1988). Cognitive flexibility
theory: Advanced knowledge acquisition in ill-structured domains (Tech. Rep. No.
441). Urbana-Champaign, IL: University of Illinois, Center for the Study of
Reading.
Spiro, R. J., Vispoel, W. L., Schmitz, J., Samarapungavan, A., & Boerger, A. (1987).
Knowledge acquisition for application: Cognitive flexibility and transfer in
complex content domains. In B. C. Britton & S. Glynn (Eds.), Executive control
processes. Hillsdale, NJ: Lawrence Erlbaum Associates.
Sweller, J. (1988). Cognitive load during problem solving: Effects on learning.
Cognitive Science, 12, 257–285.
Sweller, J. (1994). Cognitive load theory, learning difficulty and instructional design.
Learning and Instruction, 4, 295–312.
150
P.D. Antonenko, D.S. Niederhauser / Computers in Human Behavior 26 (2010) 140–150
Sweller, J. (2008). Human cognitive architecture. In M. Spector et al. (Eds.),
Handbook of research for educational communications and technology (3rd ed.,
pp. 369–383). Lawrence Erlbaum Associates.
Sweller, J., van Merrienboer, J., & Paas, F. (1998). Cognitive architecture and
instructional design. Educational Psychology Review, 10, 251–296.
Vygotsky, L. S. (1978). Mind and society: The development of higher psychological
processes. Cambridge, MA: Harvard University Press.
Xie, B., & Salvendy, G. (2000). Review and reappraisal of modeling and predicting
mental workload in single- and multi-task environments. Work and Stress,
14(1), 74–99.
Zhu, E. (1999). Hypermedia interface design: The effects of number of links and
granularity of nodes. Journal of Educational Multimedia and Hypermedia, 8(3),
331–358.