Computers in Human Behavior 30 (2014) 121–130

Contents lists available at ScienceDirect

Computers in Human Behavior

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

Qualitatively different cognitive processing during online reading

primed by different study activities
Hector R. Ponce a,⇑, Richard E. Mayer b,1
a
Faculty of Management and Economics, University of Santiago of Chile, Av. L.B. O’Higgins, 3363 Santiago, Chile
b
Department of Psychological and Brain Sciences, University of California, Santa Barbara, CA 93106-9660, USA

a r t i c l e i n f o a b s t r a c t

Article history: This article uses eye-tracking technology to examine how study activities such as taking notes or filling in
Available online 29 August 2013 a graphic organizer affect cognitive processing during learning. College students read a computer-pre-
sented passage that compared the characteristics of eastern steamboats (top section) and western steam-
Keywords: boats (bottom section), either by reading it twice (read-only group), typing notes into a textbox on the
Graphic organizer right side of the screen (note-taking group), or typing characteristics of the two types of steamboats into
Note-taking a compare-and-contrast graphic organizer on the right side of the screen (graphic organizer group). Com-
Eye tracking
pared to the note-taking group, the graphic organizer group displayed more eye movements between the
Learning strategy
Reading
top and bottom of the passage (i.e., integrative saccades, d = 1.03), more eye movements between the text
and the type-in window on the right side (i.e., constructive saccades, d = 0.79), fewer constructive sac-
cades during initial reading (d = 0.64), and less time looking to the right side during initial reading
(d = 0.81); and scored higher on a comprehension test given afterwards (d = 1.17), although both study
groups outscored the read-only group. Results suggest that students in the note-taking group (and read-
only group) tended to use a linear learning strategy in which their eyes followed the text in the order pre-
sented whereas students in the graphic organizer group tended to use a generative learning strategy in
which their eyes searched for connections between specific information across the passage required to
make comparisons.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Consider a student who reads an expository passage presented

Being able to learn from expository text (i.e., text that describes
or explains) is a central skill in the development of literacy, and has
received increased scrutiny because of its prominence in new cur-
riculum frameworks such as the Common Core Standards in the US
(Porter, McMaken, Hwang, & Yang, 2011) and in international
assessment programs such as the Programme for International
Student Assessment (Organization for Economic Co-operation
and Development, 2013). The goal of the present study is to com-
pare the cognitive processes that are fostered by two different
study strategies intended to improve learning from expository
text—taking notes and filing in a graphic organizer—in order to
better understand how study activities can improve learning.
Consequently, the primary research question addressed in this
study is whether these study activities—note taking and graphic
organizers—prime qualitatively different cognitive processing
during learning.
⇑ Corresponding author. Tel.: +56 2 27180716.
E-mail addresses: hector.ponce@usach.cl (H.R. Ponce), rich.mayer@psych.ucsb.
edu (R.E. Mayer).
1
Tel.: +1 805 8932472.
on a computer screen, such as the steamboat passage—adapted
from (Meyer & Poon, 2001)—shown in Fig. 1. For example, to
enhance learning, some students could be asked to take notes by
typing into a textbox on the right side of the screen during learn-
ing, such as shown in Fig. 2. In a different attempt to enhance
learning, other students could be asked to complete a compare-
and-contrast graphic organizer that enables them to compare the
two types of steamboats along several key dimensions by typing
in the name of each dimension as well as the corresponding attri-
butes for each type of steamboat, as shown in the right side of
Fig. 3.
Most previous research assesses the impact of different study
strategies by examining learning outcomes, such as measured by
comprehension tests or memory tests (Dunlosky, Rawson, Marsh,
Nathan, & Willingham, 2013; Mason, Pluchino, & Tornatora,
2013; Mayer, 2008). In the present study we sought a more direct
measure of learners’ cognitive processing during learning primed
by different study strategies. Therefore, we used eye-tracking
methodology to examine students’ eye movements as they simply
read a text (read group), took notes by typing into a textbox as they
read a text (note-taking group), or filled in a graphic organizer as
they read a text (graphic organizer group).

0747-5632/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.chb.2013.07.054
122 H.R. Ponce, R.E. Mayer / Computers in Human Behavior 30 (2014) 121–130
Fig. 1. The steamboat passage in the read-only condition.
2. Research on note-taking and graphic organizers
A study activity refers to actions performed by a learner during
learning that are intended to improve learning, whereas a learning
strategy refers to cognitive processing performed by a learner dur-
ing learning which can be primed by a study activity (Mayer, 2008;
Weinstein & Mayer, 1985). The conventional view is that study
activities such as taking notes or filling in graphic organizers cause
students to engage in more useful cognitive processing during
learning and thereby result in better memory for the material than
when students do not engage in study activities (Dunlosky et al.,
2013; Mayer, 2011).
Note-taking and graphic organizers are two well-known exam-
ples of study activities that are intended to support reading com-
prehension (National Institute of Child Health and Human
Development, 2000). Research on note-taking has concentrated
on its process and product functions: the process function (or
encoding function) refers to the impact of taking notes on learning
Fig. 2. The steamboat passage and the note editor.
outcomes, while the product function (or external storage func-
tion) refers to studying the impact of reviewing notes on learning
outcomes (Kiewra, 1985). Although most research has focused on
note-taking during lectures, some general findings are relevant
for this study. The process of note-taking promotes encoding and
storage of information in long-term memory (Barnett, di Vesta, &
Rogozinski, 1981). The process of reviewing notes supports retriev-
ing information from long-term memory, and it facilitates remem-
bering and provides opportunities for deeper elaboration of the
presented material (Bohay, Blakely, Tamplin, & Radvansky, 2011).
The act of note-taking has been demonstrated to improve students’
performance on comprehension tests under appropriate circum-
stances (Kiewra, 1985; Peper & Mayer, 1978). However, research
shows that in general students take incomplete notes and do not
address critical points (Kiewra et al., 1991; Peverly, Brobst,
Graham, & Shaw, 2003).
A graphic organizer, on the other hand, is a spatial structure for
representing material (such as a matrix). Similar to note-taking, re-
search on graphic organizers has concentrated on its process and
product functions. The process function refers to researching the
impact of using graphic organizers as thinking procedures to assist
learners on the steps to carry out a cognitive operation (Beyer,
1997). A Venn diagram, for example, can support the development
of the compare-and-contrast cognitive strategy by guiding stu-
dents to establish the elements to be compared, to indicate their
common characteristics, and to figure out their differences. Asking
students to fill in graphic organizers during reading can improve
students’ learning outcomes (Ponce, Lopez, & Mayer, 2012; Robin-
son, 1998). On the other hand, the product function of graphic
organizers refers to researching the impact of studying filled-in
or teacher-made graphic organizers on learning outcomes (Stull
& Mayer, 2007). Numerous studies have shown that using graphic
organizers and note-taking appropriately employed in learning
contexts are effective text comprehension activities, and that
structured note-taking strategies such as with outline or matrix
format are more beneficial than linear recording of information
(Jairam & Kiewra, 2009; Kiewra et al., 1991; Piolat, Olive, & Kellogg,
2005).
Graphic organizers are tied to basic rhetorical structures (i.e.,
common ways of structuring text), such as compare-and-contrast
(e.g., matrix), sequence (e.g., flow chart), or hierarchy (e.g., tree
 H.R. Ponce, R.E. Mayer / Computers in Human Behavior 30 (2014) 121–130
Fig. 3. The steamboat passage and the comparison graphic organizer.
diagram). Text comprehension performance can be improved
when students receive training in how to outline based on basic
rhetorical structures (Cook & Mayer, 1988) or when signaling tech-
niques such as headings are used to highlight the text’s rhetorical
structure (Meyer & Poon, 2001). In this study, we used a compare-
and-contrast structure as described in the previous section. In par-
ticular, we focused on the steamboat passage as an example of an
expository text with a common rhetorical structure (i.e., compare-
and-contrast), as it is beyond the scope of this study to examine all
major types of expository text.
The present study is inspired by a previous large-scale class-
room experiment, in which students who learned how to build
computer-based graphic organizers for on-screen text passages
showed improvements on reading comprehension tests (Ponce
et al., 2012). To complement that study conducted at the macro-le-
vel, the present study focuses on the micro-level to pinpoint some
of the cognitive processes during learning that could contribute to
the findings. Eye-tracking measures are used as a way of gauging
learners’ cognitive processing during learning.
3. Theoretical framework and predictions
The present study extends the growing body of work on differ-
ences in the amount remembered caused by study activities, by
examining the further proposal that different study activities—
such as taking notes versus filling in a graphic organizer—can fos-
ter qualitatively different learning strategies, that is, qualitatively
different cognitive processes during learning rather than differ-
ences solely in the quantity of learning. According to the cognitive
theory of multimedia learning (Mayer, 2009, 2011), the quality of a
learning outcome depends on the kinds of the cognitive processes
that learners engage in during learning.
Consider the learning strategies that readers could use in read-
ing the steamboat passage shown in Fig. 1. A default approach is to
view the steamboat passage as a list of facts to be remembered, in
which case the learner would use a linear learning strategy of read-
ing the sentences in order while trying to remember each fact in
the order presented in the text. In contrast, a structured approach
is to view the steamboat passage as a compare-and-contrast
123
structure consisting of two types of boats being compared along
several dimensions. In this case, the learner would use a generative
learning strategy of reading the passage in an effort to build a ma-
trix that compares the two steamboats on each dimension, so for
example, when reading that the eastern steamboats operated in
deep rivers the reader might scan down in the passage to see that
western steamboats operated in shallow waters. The goal of this
study is examine techniques that encourage readers to change
from a linear strategy to a generative strategy.
We examine the prediction that the study activity of note-tak-
ing leads to a linear learning strategy whereas the study activity
of using graphic organizers leads to a generative learning strategy.
Table 1 compares a linear learning strategy which involves adding
information to memory as presented versus a generative learning
strategy which involves building a coherent structure in which
the material is reorganized, as summarized in the first line of the
table. The left side of the second line of Table 1 shows that cogni-
tive processes involved in a linear learning strategy are character-
ized in the cognitive theory of multimedia learning (Mayer, 2009)
as essential processing, in which learners select essential elements
from the lesson and mentally organize them in memory as a list
in the order presented in the lesson. The result is the construction
of a surface-level representation, that is, a copy of the main ideas in
the presented material, using the same structure as in the text, as
shown in the left side of the third line of Table 1. We propose that
asking students to engage in standard note-taking can promote
essential processing, in which students simply write down facts
from the lesson in the order presented.
In contrast, the cognitive processes involved in a generative
learning strategy are characterized in the cognitive theory of multi-
media learning as generative processing, in which learners select
relevant material from the lesson, mentally reorganize it to fit into
a coherent structure, and integrate it with a relevant prior schema
activated from long-term memory, as summarized in the right side
of the second line of Table 1. The result is a deep-level representa-
tion based on constructing an integrated structure (which can be
called a mental model) that has internal coherence (i.e., fits together
in a logical way) and external coherence (i.e., relates to existing
schemas), as shown in the right side of the third line of Table 1.
We propose that asking students to engage in filling in a graphic
124 H.R. Ponce, R.E. Mayer / Computers in Human Behavior 30 (2014) 121–130

Table 1
Comparison of linear and generative learning strategies.

Linear learning strategy Generative learning strategy

Description Adding information sequentially in the order presented (such as a Building an organized structure (such as a matrix)
list)
Cognitive processes Essential processing (selecting and organizing information as Generative processing (selecting and reorganizing and integrating
presented) information)
Mental Surface-level representation (e.g., text base) Deep-level representation (i.e., mental model)
representation

organizer can promote generative processing, in which students
use a compare-and-contrast schema primed from long-term mem-
ory and mentally reorganize the material in the lesson to fit within
the compare-and-contrast schema.
The distinction between surface and deep representations is
consistent with Kintsch’s (1988) construction-integration theory
of prose comprehension, and its distinction between text-base
and situation model representation of text. The distinction be-
tween linear and generative learning strategies is consistent with
Wittrock’s (1989; Mayer & Wittrock, 2006) generative theory of
reading comprehension, and its distinction between generative
and non-generative processes during learning.
4. Measures
Table 2 summarizes some key measures used in the study to
evaluate the learning process and learning outcomes primed by
different study activities. We used eye-tracking measures (Johnson
& Mayer, 2012; Rayner, 1998) to gauge the process of learning, by
breaking the screen (such as shown in Figs. 2 and 3) into three
areas of interest (AOIs)—the top portion of the passage consisting
of 9 sentences describing eastern steamboats (area A), the bottom
portion of the passage consisting of 11 sentences describing wes-
tern steamboats (area B), and the right side of the screen contain-
ing a textbox for note-taking or for filling in the graphic organizer
(area C).
First, integrative saccades (or up-down saccades) are the num-
ber of times a learner’s eyes move between areas A and B. We
interpret the total number of integrative saccades as an indication
of a generative learning strategy, because they reflect an attempt to
make connections that go beyond simple linear order.
Second, constructive saccades (or left–right saccades) are the
number of times a learner’s eyes move between area C and com-
bined areas A and B. We interpret the total number of constructive
saccades as an indication of a generative learning strategy, because
they reflect an attempt to make connections to several portions of
the text rather than going in simple linear order. If the graphic
organizer group is more likely to engage in generative learning
strategy and the note-taking group is more likely to engage in lin-
ear processing, then we predict that the graphic organizer group
will produce a greater number of integrative saccades and con-
structive saccades than the note-taking group. We also expect
the read-only group to display a linear learning strategy, indicated
by fewer integrative saccades than the graphic organizer group.
Next consider what happens during the initial reading of the
passage, which we estimate takes about 25 s based on observations
of students in the read-only group. The third line in Table 2 high-
lights the number of early constructive saccades, that is, the num-
ber of constructive saccades that occur during the first 25 s. If a
learner is engaging in linear processing, we would expect construc-
tive saccades to occur from the very start because the learner is
simply reading a bit and entering that information in the textbox,
then reading the next bit and entering that information in the text-
box, and so on. If a learner is engaging in generative processing, we
would expect the learner to read over the entire passage initially as
the basis for figuring out how to organize it, and then go back and
start entering text into the textbox. Similarly, the fourth line of Ta-
ble 2 highlights the number of seconds spent looking at the textbox
(area C) during the first 25 s of reading—that is, early non-text fix-
ation time. We expect learners engaging in linear processing to
show greater early non-text fixation time, and learners who read
over the entire text in order to figure out the overall structure to
show less early non-text fixation time. Based on this analysis, we
predict that the graphic organizer group will score lower on early
constructive saccades and early non-text fixation time than the
note-taking group.
The foregoing process measures are the main new focus of this
study, but we included learning outcome measures to make con-
nections with previous work. The bottom portion of Table 2 sum-
marizes two learning outcome measures based on a summary
test, in which learners write a summary of the passage, and a cloze
test, in which learners fill in blanks corresponding to key facts in
the passage. The summary test score consists of the number of cor-
rect comparisons (i.e., comparable information for both the eastern
and western steamboat on a particular dimension) which indicates
generative processing, and the cloze test score consists of the
number of correctly filled in slots using the verbatim wording of

Table 2
Measures of the process and product of learning.

Name Description
Measures of learning process
Integrative saccades (up- Number of times eyes move from the top portion of the passage (area A) to the bottom portion of the passage (area B) or vice versa.
down) Indicates generative learning strategy
Constructive saccades Number of times eyes move from the passage (area A or B) to the note-taking textbox or graphic organizer (area C) or vice versa. Indicates
(left–right) generative learning strategy
Early constructive Number of constructive saccades in first 25 s (i.e., during initial reading of the passage). Indicates linear learning strategy
saccades
Early non-text fixation Number of seconds spent looking at note-taking textbox or graphic organizer (area C) during first 25 s (i.e., initial reading of the passage).
time Indicates linear learning strategy
Measures of learning product
Summary test score Number of key comparisons recalled. Indicates reading comprehension
Cloze test score Number of key verbatim terms filled in. Indicates rote memory
 H.R. Ponce, R.E. Mayer / Computers in Human Behavior 30 (2014) 121–130
the lesson which may be supported by linear processing but by
generative processing as well. If study activities are effective, we
expect the two study activity groups (note-taking and graphic
organizer) to outperform the read-only group. If the graphic orga-
nizer group is more likely to engage in generative processing then
we predict better performance on the summary test, which is a
measure of comprehension, but not necessarily on the cloze test,
which is a rote memory test.
5. Method
5.1. Participants and design
The participants were 64 students recruited from the Psychol-
ogy Subject Pool at the University of California, Santa Barbara. In
a between-subjects design, 21 participants served in the graphic
organizer group, 22 participants were in the note-taking group,
and 21 participants were in the read-only group. Data from three
participants (1 in the note-taking group and 2 in the graphic orga-
nizer group) were excluded in the eye-tracking measures analysis
due to the low quality of eye-tracking data for these subjects. The
average age of participants was 18.44 years (SD = 1.11), the propor-
tion of females was 0.58, and the proportion of freshmen was 0.77.
The groups did not differ significantly (at p < .05) on mean age, pro-
portion of females and males, or self-rated knowledge of
steamboats.
5.2. Materials and apparatus
The materials consisted of an informed consent sheet, summary
test sheet, cloze test sheet, participant questionnaire, and three
computer-based versions of the steamboat passage. The informed
consent form described the study and human subject protections.
The summary test was computer-based and consisted of a blank
textbox that students had to use to write a summary of the lesson
without access to the original text and notes. The summary was
scored by counting the number of complete comparisons made
based on a blind-scoring process. From the original text, it is pos-
sible to create a maximum of nine comparisons associated with
the following attributes: steamboat type, rivers travelled, river
depth, hull type, cargo storage, number of floors, engine type, en-
gine efficiency, and engine safety. One point was given to each
complete comparison made in the summary (e.g., ‘‘the engines
were low-pressure in eastern boats and high pressure in western
boats’’), either made in one sentence or on separate sentences
within the summary; yielding a total possible of 9 points. Partial
comparisons indicating a characteristic for one type of steamboat
but not for the other (‘‘the engines were low-pressure in eastern
boats’’) were not given any points. The cloze test sheet was pa-
per-based and consisted of the steamboat passage (as shown in
Fig. 1) with the following words eliminated and replaced with
blank spaces: Hudson, deep (three times), below, low-pressure,
shallow, Missouri, Ohio, Mississippi, flat, on, high-pressure. One
point was given for each exact word that was filled in correctly,
yielding a total possible of 13 points. Students were also asked to
complete a paper-based questionnaire that asked for ratings on a
5-point scale for previous knowledge about steamboats, difficulty
of the task, effort exerted on the task, and the degree to which they
liked using the application.
The learning materials consisted of three computer-based ver-
sions of a 123-word expository text, titled ‘‘The Steamboat,’’ which
was originally used by Meyer and Poon (2001). The text summa-
rized the differences between eastern-style and western-style
steamboats, and is presented in Fig. 1. The rhetorical structure of
the passage is compare-and-contrast (Cook & Mayer, 1988; Meyer
125
& Poon, 2001) in which two types of steamboats (i.e., eastern-style
and western-style) are compared along several dimensions (e.g.,
river depth, cargo storage and engine type). We chose this text
because it facilitated to measure the eye-tracking variables used
in this study.
As can be seen in Fig. 1, the top portion of the passage (i.e., first
9 sentences) describes characteristics of eastern-style steamboats,
whereas the bottom portion (i.e., last 11 sentences) describes cor-
responding characteristics of western-style steamboats. There
should be a lot of eye movements between the top and bottom por-
tions of the passage if readers use a generative reading strategy but
not if they use a linear reading strategy.
As shown in Fig. 1, the read-only version presented the text
titled ‘‘The Steamboat’’ as the only window on the screen. The
read-only application was implemented as a PowerPoint presenta-
tion with the steamboat text in similar format to the one presented
for the other treatments. In the read-only version, we asked partic-
ipants to read through the text twice in order to observe a linear
reading pattern and to compare it with those generated in the
note-taking and graphic organizer conditions.
As shown in Fig. 2, the note-taking version presented the text in
a window on the left side and also provided a textbox on the right
side for typing in notes. The note-taking strategy was implemented
as a PowerPoint presentation by integrating a textbox object that
worked as a basic text editor. The note-taking application did not
allow students to copy and paste text into the text box. In the
note-taking condition, we asked participants to read the text and
take notes using the embedded text editor.
As shown in Fig. 3, the graphic organizer version presented the
text in a window on the left and also provided an interactive gra-
phic organizer tool on the right with a compare-and-contrast gra-
phic organizer. The graphic organizer treatment was created by
integrating into a PowerPoint presentation an Adobe Flash object
that implemented a comparison graphic organizer (Almarza,
Ponce, & Lopez, 2011). This interactive graphic organizer applica-
tion allowed students to add and delete text boxes for each com-
parison, and for each comparison to type in the name of the
dimension to be compared and to type in the respective attri-
butes for the two types of boats. An empty organizer was pre-
sented with only one set of boxes. If students wanted to add
attributes, they had to use the respective function in the applica-
tion. The graphic organizer application did not allow students to
copy and paste text into the organizer. In this condition, we asked
participants to read the text and complete the embedded graphic
organizer.
The eye-tracker allows the generation of areas of interest (AOI)
so they can be analyzed as independent elements within the total
area of eye-movement recordings. In our study, we defined three
areas of interest: (1) the top section of the text (i.e., the first 9 sen-
tences, which described eastern-style steamboats) in the window
on the left side of the screen, (2) bottom section of the text (i.e.,
bottom 11 sentences, which described western-style steamboats)
in the window on the left side of the screen, and (3) graphic orga-
nizer section (for the graphic organizer group) or note editor text-
box section (for the note taking group) represented as a window in
the right side of the screen. As indicated previously, the top and
bottom sections of the text correspond to the description of the
eastern-style and western-style steamboats, respectively. In the
read-only condition, only the top and bottom sections of the text
were defined as areas of interest.
If readers focus on building a mental representation based on
the structure of the text (i.e. compare-and-contrast), then they
should produce more saccades between the top and bottom sec-
tions of the text, which we call integrative saccades. For instance,
when contrasting the hulls, it is possible to read the sentences
‘‘. . .the deep hulls of the eastern steamboat’’ and ‘‘. . .Their hulls
126 H.R. Ponce, R.E. Mayer / Computers in Human Behavior 30 (2014) 121–130

Fig. 4. Gaze plot showing a generative learning strategy.

Fig. 5. Gaze plot showing a linear learning strategy.

were flat. . .’’ relationally. In the eye-tracker the relation between
these two sentences should be recorded as a series of fixations
on the first sentence (top section), a saccade to the second sentence
(bottom section) and a series of fixations on the later sentence. We
interpret integrative saccades as an indication that the reader is
using a generative reading strategy. In particular, we are interested
in whether students in the graphic organizer group engage in qual-
itatively different reading strategies than students in the note-tak-
ing group, as indicated by differences in the amount of integrative
saccades.
The apparatus consisted of a Tobii T-60 eye-tracking system and
Dell computer for controlling the eye-tracking system.
5.3. Procedure
Participants were randomly assigned to one of the three groups,
and were tested individually. They were told that they would par-
ticipate in a study whose aim was to understand eye-movement
patterns while reading a text. Participants received a short expla-
nation of the eye-tracker at the beginning of the session and then
 H.R. Ponce, R.E. Mayer / Computers in Human Behavior 30 (2014) 121–130
Table 3
Means and standard deviations on four eye-tracking measures for each group.
Eye-tracking measure Graphic-organizer (N = 20)
M SD
Integrative saccades (up-down) 20.40 7.41
Constructive saccades (left–right) 64.25 18.94
Early constructive saccades (left–right) 2.40 3.45
Early non-text fixation time 1.00 1.37
they were asked to complete a form with some basic demographic
information and to sign a consent form. For the read-only group,
the next step was to calibrate the eye-tracker for each participant,
and then present the steamboat passage. Participants in the read-
only group were asked to read the text twice while the eye-tracker
recorded their eye movements and to tell the researcher when they
had completed the task, which stopped the recording. In the note-
taking group, each participant was introduced to the note-taking
application, using a practice text. For training purposes, partici-
pants in the note-taking group were asked to take some notes on
the practice text so they could familiarize themselves with the edi-
tor; they were told that it was not possible to copy-and-paste from
the original text. After this training, the calibration of the eye-
tracker and the recording of the note-taking session followed. Par-
ticipants were told that they had five minutes to complete the task;
in case they finished before the allocated time, they were asked to
tell the researcher in order to stop the recording, otherwise the re-
searcher told the participants when the time was up. In the graphic
organizer condition, the procedure was the same as for the note-
taking condition, except in the training session participants were
taught how to use the interactive graphic organizer embedded in
the application. The training sessions took about 5 min.
For each condition, immediately after each participant com-
pleted studying the steamboat passage and the recording was fin-
ished, the participant was asked to write a short summary of the
passage by typing directly into a textbox window using the MS-
Word editor; the participant was informed that the maximum time
to complete the summary was 4 min. The participant was not al-
lowed to review notes or the original text to write the summary.
After completing the summary, the paper-based cloze test sheet
was administered with a time limit of 2 min. Finally, the partici-
pant was asked to complete the paper-based questionnaire, which
solicited information about previous knowledge on steamboats,
difficulty and effort during the task, and satisfaction with the
application.
6. Results
6.1. Do different study activities foster qualitatively different learning
strategies?
The primary prediction is that studying with a graphic organizer
primes a generative learning strategy (reflected by a high number
of integrative or up-down saccades see Fig. 4) whereas studying
with note-taking or with no aids primes a linear learning strategy
(reflected by a low number of integrative saccades see Fig. 5). The
top row of Table 3 summarizes the mean number of integrative
saccades (and standard deviations) during studying for each group.
An analysis of variance (ANOVA) showed that the groups differed
significantly, F(2, 58) = 34.71, p < .001, and follow-up post hoc
comparisons using a Tukey test (with p < .05) demonstrated that
the graphic organizer group made significantly more integrative
saccades than the note-taking group (d = 1.03) and read-only group
(d = 2.55), and that the note-taking group also made significantly
more integrative saccades than the read-only group (d = 1.77).
127
Note-taking (N = 20) Read-only (N = 21) GO vs. NT
M SD M SD d
13.75 5.38 5.86 3.35 1.03
49.30 19.07 – – .79
5.45 5.74 – – .64
3.13 3.46 – – .81
A secondary prediction is that studying with a graphic organizer
primes a generative learning strategy that is reflected in a higher
number of constructive saccades (i.e., left–right saccades) whereas
studying with note-taking primes a linear learning strategy that is
reflected in a lower number of constructive saccades. The second
row of Table 3 summarizes the mean number of constructive sac-
cades (and SDs) during studying. Consistent with this prediction, a
t-test found that the graphic organizer group made significantly
more constructive saccades than the note-taking group,
t(38) = 2.49, p = .017, d = .79.
Next, to conduct a somewhat more fine-grained analysis, we fo-
cused on the first 25 s of reading the text, which is the average time
taken by the participants in the read-only group to read completely
through the text the first time (considering this group was asked to
read the text twice). Focusing on the first 25 s gives some insight
into the learning strategies taken by participants in each group,
particularly the degree to which they started taking notes or filling
in the graphic organizer before reading the text completely which
reflects a linear learning strategy. The third row in Table 3 summa-
rizes the mean number of constructive saccades in the first 25 s
(and SDs) for the note-taking and graphic organizer groups, and
shows that the note-taking group made significantly more early
constructive saccades than the graphic organizer group,
t(38) = 2.04, p = .049, yielding a medium effect size (d = .64).
The fourth row of Table 2 summarizes the mean fixation time on
either the graphic organizer or text editor (and SDs) during the first
25 s, and shows that the note-taking group spent significantly
more time looking on the side of the screen for typing in words
(i.e., showed greater early non-text fixation time) than the graphic
organizer group, t(38) = 2.56, p = .015, with a large effect size
(d = .81). Overall, these two pieces of evidence suggest that par-
ticipants in the note-taking group were more likely than those in
the graphic organizer group to begin writing before they had com-
pletely read over the passage, which is consistent with a linear
learning strategy.
Overall, the eye-tracking data summarized in Table 3 are con-
sistent with the idea that the graphic organizer study activity is
more likely to encourage a generative learning strategy whereas
the note-taking study activity is more likely to encourage a linear
learning strategy (which also appears to be the default strategy
used in the read-only group). In short, the eye-tracking data pro-
vides evidence for qualitatively different learning strategies fos-
tered by different study strategies.
6.2. Do different study activities foster different learning outcomes?
The preceding section focused on measures of the process of
learning based on eye tracking data, which is the main focus of this
research, indicating that the graphic organizer group was more
likely than the other groups to use a generative learning strategy
during reading. This section focuses on measures of learning out-
comes based on the summary test and cloze test. First, we examine
the prediction that the graphic organizer group will outperform the
others on the summary test, which is the best measure of reading
comprehension available in this study, because the graphic
128 H.R. Ponce, R.E. Mayer / Computers in Human Behavior 30 (2014) 121–130

organizer group is more likely to use a generative learning strategy.
The first line of Table 4 summarizes the means and standard devi-
ations on the summary test, along with effect sizes (based on Co-
hen’s d) for all three pair-wise comparisons. Consistent with
predictions, an ANOVA showed a significant difference among
the three groups on the summary score, F(2, 61) = 19.59, p < .001,
and a follow-up post hoc analysis using a Tukey test (with alpha
at .05) revealed that the graphic organizer group significantly out-
scored the note-taking group (d = 1.17) and the read-only group
(d = 1.94), whereas the difference between the note-taking and
read-only groups reached marginal significance (p = .061, d = 0.73).
Second, we examine the prediction that active study tech-
niques such as note-taking and using graphic organizers will re-
sult in better performance than the read-only group on the
cloze test, which is a measure of rote memory. If the cloze test
is mainly a measure of rote memory, then we can assume that
it is less sensitive to the effects of using a generative learning
strategy than is the summary test, and therefore it might not
show a superiority of the graphic organizer group over the
note-taking group. Consistent with predictions, an ANOVA
showed a significant difference among the three groups on the
cloze score, F(2, 61) = 14.48, p = .000, and a follow-up post hoc
analysis (with alpha at .05) showed that the graphic organizer
group significantly outscored the read-only group (d = 1.69) and
the note-taking group significantly outscored the read-only group
(d = 1.12), but the graphic organizer group did not significantly
outscore the note-taking group (d = 0.52).
Overall, the learning outcome data show strong effects for add-
ing study techniques to a reading task, and furthermore, the gra-
phic organizer group excels over the note-taking group when a
measure of reading comprehension is used.
6.3. Do different study activities foster different levels of self-reported
effort, difficulty, and motivation?
Asking students to engage in study activities such as filling in a
graphic organizer as they read an expository passage is expected to
require more mental effort during learning (which reflects in-
creased generative processing as compared to the read-only condi-
tion) but lower perceived difficulty (which reflects decreased
extraneous processing as compared to the read-only condition).
The top two lines of Table 5 show the mean ratings and standard
deviations (on a 5-point scale) of each group on difficulty and ef-
fort, respectively. Consistent with predictions, an ANOVA revealed
significant differences among the groups on difficulty, F(2,
61) = 4.77, p = .012, and effort, F(2, 61) = 5.92, p = .004, with pair-
wise Tukey tests (with alpha less than .05) showing that the
read-only group reported significantly higher difficulty and lower
effort as compared to the graphic organizer group and the note-
taking group, which did not differ significantly with each other.
Engaging in study activities such as filling in a graphic organizer
while reading an expository passage is expected to be more moti-
vating than simply reading the passage. The third row of Table 5
shows the mean agreement ratings and standard deviations (on a
5-point scale) for the three groups on the item, ‘‘I would like more
lessons like this one’’, which is intended as a measure of an aspect
of motivation, namely, persistence on a learning task. Consistent
with predictions, an ANOVA revealed a significant difference
among the groups, F(2, 61) = 5.30, p = .008, with pair-wise Tukey
tests (with alpha less that .05) showing only that read-only group
gave a significant lower rating than the graphic organizer group.
The fourth line of Table 5 summarizes the mean ratings and
standard deviations for the groups, on an item aimed at measuring

Table 4
Means and standard deviations on summary test and cloze test for each group.

Test Graphic-organizer (N = 21) Note-taking (N = 22) Read-only (N = 21) GO vs. NT GO vs. RO NT vs. RO
M SD M SD M SD d d d
Summary 4.86 1.82 2.73 1.83 1.43 1.72 1.17 1.94 .73
Cloze 7.81 2.87 6.32 2.82 3.48 2.20 .52 1.69 1.12

Note: GO: Graphic-organizer, NT: Note-taking, RO: read-only.

Table 5
Mean ratings and standard deviations on effort, difficulty, and motivation for three groups.

Item Graphic-organizer Note-taking Read-only GO vs. NT GO vs. RO NT vs. RO

M SD M SD M SD d d d
Difficulty 2.38⁄ .59 2.27⁄ .70 2.95 .97 .17 .71 .81
Effort 3.33⁄ .66 3.27⁄ 1.08 2.52 .75 .07 1.15 .80
Motivation 3.86⁄ .57 3.45 .67 3.19 .75 .66 1.01 .37
Satisfaction 3.81 .81 3.91 .68 3.43 .68 .13 .51 .71

Note: Asterisk indicates significant difference from read-only group, GO: Graphic-organizer, NT: Note-taking, RO: read-only. Higher number indicates higher levels of
difficulty, effort, motivation, or satisfaction, respectively.

Table 6
Means and standard deviations on total fixation time (in s).

Areas of Interest Graphic-organizer (N = 20) Note-taking (N = 20) Read-only (N = 21)

M SD M SD M SD
All areas of interest 199.49 42.41 144.65 37.12 62.94 13.19
Top 64.36 19.28 47.19 12.57 32.00 7.06
Bottom 63.03 18.37 47.54 14.67 30.94 7.66
GO or NT 72.10 25.18 49.92 26.53 – –

Note: GO: Graphic-organizer and NT: Note-taking.

 H.R. Ponce, R.E. Mayer / Computers in Human Behavior 30 (2014) 121–130
overall affect for the lesson (i.e., ‘‘I like this program’’). An ANOVA
shows that the groups did not differ significantly, F(2, 61) = 2.59,
p = .083, indicating that liking is not the same as learning.
6.4. Do the groups differ on basic eye-movement measures?
The foregoing results constitute the main empirical focus of the
study, concerning the idea that different study activities prime
qualitatively different cognitive processing during learning. This
section examines whether the groups differ on basic eye-move-
ment measures. Table 6 shows the total study time for the whole
lesson (all areas of interest) and on each area of interest separately.
The means for total fixation times (in seconds) were significantly
different among the three groups, as revealed by a one-way ANO-
VA, F(2, 58) = 88.18, p = .000, with post hoc Tukey tests (with alpha
at .05) indicating that each of the groups differed significantly from
each other. The read-only group was asked to read the text twice,
whereas the GO and NT groups had up to 5 min to complete their
study activities. However, participants using the graphic organizer
spent significantly more time focused on typing into the graphic
organizer than the note-taking group spent focused on typing into
the textbox, t(38) = 4.35, p < .001. In order to determine whether
study time in each condition had an influence on learning out-
comes, we ran two one-way ANCOVAs with total fixation time as
a covariate and summary score and cloze score as independent
variables, respectively. Similar results were observed as those ob-
tained through one-way ANOVAs for the summary test and cloze
test as reported in the previous section; suggesting that study time
did not have a major influence on the results.
7. Discussion
7.1. Empirical contribution
The main empirical contribution of this study comes from the
eye-tracking analysis, in which the graphic organizer group pro-
duced more integrative saccades (i.e., up-and-down within the
passage) and constructive saccades (i.e., left and right between
the passage and the notes or graphic organizer) but fewer early
constructive saccades and less early non-text fixation time than
the note-taking group. Secondarily, the graphic organizer group
outscored the note-taking group on a test of reading comprehen-
sion (i.e., the summary test) but not on a test of verbatim memory
(i.e., the cloze test), and both groups outperformed the read-only
group.
7.2. Theoretical contribution
The pattern of results supports the idea that the study activities
of taking standard notes and filling in a graphic organizer prime
qualitatively different cognitive processes during learning (i.e.,
qualitatively different learning strategies). Specifically, note-taking
is more likely to prime a linear learning strategy in which learners
add facts to memory in the order presented, whereas graphic orga-
nizers are more likely to prime a generative learning strategy in
which learners mentally reorganize the facts to fit within a struc-
ture (such as a compare-and-contrast matrix) activated from
long-term memory. Rather than viewing study strategies as affect-
ing the amount of cognitive processing during learning, this study
shows that different study strategies can prime differences in the
quality and type of cognitive processing during learning.
In this particular case, we have observed that the use of the
graphic organizer assisted learners in carrying out the compare-
and-contrast cognitive operation required to have a deeper
understanding of the steamboat passage. This is consistent with
129
the multimedia theory of learning (Mayer, 2009), in the sense that
the organization embedded in the graphic organizer primes the
selection of information in the text, resulting in a new spatial rep-
resentation that is integrated in long term memory not as list of
facts but as compare-and-contrast structure. This integration
process was observed through the significantly different ways stu-
dents wrote their summaries in the three experimental conditions.
The read-only and note-taking groups wrote their summaries
mostly as list of facts, remembering aspects of one element but
not necessarily relating it to the other. On the contrary, the graphic
organizer group tended to write sentences that related both ele-
ments through the appropriate categories. These two ways of
approaching the writing of the summaries can be interpreted as
expressions of Kintsch’s (1988) text-base and situation model of
the text.
7.3. Practical and methodological contribution
These findings have instructional implications, favoring the use
of graphic organizers for promoting deep understanding of expos-
itory text. This conclusion is consistent with Beyer’s (1997) recom-
mendation to use graphic organizers that highlight a clear spatial
structure such as comparison, sequence, classification, and cause-
and-effect. He also emphasizes the need to teach students proce-
dures for organizing information based on appropriate graphic
organizers, showing them the cognitive operations involved. From
our study, we can additionally recommend that teachers show stu-
dents how their reading patterns are improved when processing
information through graphic organizers, and that some study
activities can facilitate deeper processing.
This study shows that eye-tracking methodology can be suc-
cessfully extended to research on study skills. If we had used solely
conventional measures of learning outcome that focus on differ-
ences in the amount learned, we would have concluded that
note-taking and graphic organizers are each effective study activi-
ties that improve learning as compared to read-only conditions.
When we conduct eye-tracking analyses, we find otherwise
unavailable evidence that the two study activities prime qualita-
tively different cognitive processing during learning.
7.4. Limitations and future directions
One limitation of the study is that we used only one type of
expository text—a comparison-and-contrast text. The text on
steamboats is well structured which naturally facilitated not only
its processing and comprehension but also the eye-tracking mea-
sures used in this study. Less structured text may require more fix-
ations and saccades and may show other reading patterns. In this
case, a text processing strategy such as graphic organizers may
play a more important role on a deeper analysis of the text than
a less structured strategy such as note-taking. Therefore, one pos-
sible research avenue is to vary the organizational quality of texts.
Another research avenue is to incorporate other types of exposi-
tory text structures (such as process or hierarchy) or other types
of text genres (such as narratives) along with appropriate graphic
organizers. A further option is to incorporate other commonly used
text processing strategies (such as highlighting and questioning)
and observe eye movement behaviors. Additional research is re-
quired as well in terms of analyzing reading patterns and the cog-
nitive operations involved when using more than one strategy to
comprehend a text—for example, using a graphic organizer to-
gether with a highlighting strategy. Finally, under some circum-
stances note taking may serve as a generative activity (Cook &
Mayer, 1988; Peper & Mayer, 1978), such as when learners are
experienced in taking structured notes, so more work is needed
130 H.R. Ponce, R.E. Mayer / Computers in Human Behavior 30 (2014) 121–130
to pinpoint the conditions that determine the learning strategies
primed by note-taking.
References
Almarza, F. A., Ponce, H. R., & Lopez, M. J. (2011). Software component development
based on the mediator pattern design: the interactive graphic organizer case.
IEEE Latin America Transactions, 9(7), 1105–1111. http://dx.doi.org/10.1109/
tla.2011.6129710.
Barnett, J. E., di Vesta, F. J., & Rogozinski, J. T. (1981). What is learned in note taking?
Journal of Educational Psychology, 73(2), 181–192.
Beyer, B. K. (1997). Improving student thinking: a comprehensive approach. Boston,
MA: Allyn and Bacon.
Bohay, M., Blakely, D. P., Tamplin, A. K., & Radvansky, G. A. (2011). Note taking,
review, memory, and comprehension. The American Journal of Psychology,
124(1), 63–73. http://dx.doi.org/10.5406/amerjpsyc.124.1.0063.
Cook, L. K., & Mayer, R. E. (1988). Teaching readers about the structure of scientific
text. Journal of Educational Psychology, 80, 448–456. http://dx.doi.org/10.1037/
0022-0663.80.4.448.
Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013).
Improving students’ learning with effective learning techniques: promising
directions from cognitive and educational psychology. Psychological Science in
the Public Interest, 14(1), 4–58. http://dx.doi.org/10.1177/1529100612453266.
Jairam, D., & Kiewra, K. A. (2009). An investigation of the SOAR study method.
Journal of Advanced Academics, 20(4), 602–629. http://dx.doi.org/10.1177/
1932202x0902000403.
Johnson, C. I., & Mayer, R. E. (2012). An eye movement analysis of the spatial
contiguity effect in multimedia learning. Journal of Experimental Psychology:
Applied, 18(2), 178–191. http://dx.doi.org/10.1037/0033-
2909.124.3.37210.1037/a0026923.
Kiewra, K. A. (1985). Investigating notetaking and review: a depth of processing
alternative. Educational Psychologist, 20(1), 23.
Kiewra, K. A., DuBois, N. F., Christian, D., McShane, A., Meyerhoffer, M., & Roskelley,
D. (1991). Note-taking functions and techniques. Journal of Educational
Psychology, 83(2), 240–245.
Kintsch, W. (1988). The role of knowledge in discourse comprehension: a
construction-integration model. Psychological Review, 95(2), 163–182.
Mason, L., Pluchino, P., & Tornatora, M. C. (2013). Effects of picture labeling on
science text processing and learning: evidence from eye movements. Reading
Research Quarterly, 48(2), 199–214. http://dx.doi.org/10.1002/rrq.41.
Mayer, R. E. (2008). Learning and instruction (2nd ed.). Upper Sadle River, N.J.:
Pearson Merrill Prentice Hall.
Mayer, R. E. (2009). Multimedia learning (2nd ed.). Cambridge, New York: Cambridge
University Press.
Mayer, R. E. (2011). Applying the science of learning. Boston, MA: Pearson/Allyn &
Bacon.
Mayer, R. E., & Wittrock, M. C. (2006). Problem solving. In P. A. Alexander & P. H.
Winne (Eds.), Handbook of educational psychology (pp. 287–303). New York, NY:
Routledge.
Meyer, B. J. F., & Poon, L. W. (2001). Effects of structure strategy training and
signaling on recall of text. Journal of Educational Psychology, 93(1), 141–159.
http://dx.doi.org/10.1037/0022-0663.93.1.141.
National Institute of Child Health and Human Development (2000). Report of the
National Reading Panel. Teaching children to read: an evidence-based
assessment of the scientific research literature on reading and its implications
for reading instruction: Reports of the subgroups (NIH Publication No. 00-
4754). Washington, DC.
Peper, R. J., & Mayer, R. E. (1978). Note taking as a generative activity. Journal of
Educational Psychology, 70(4), 514–522.
Peverly, S. T., Brobst, K. E., Graham, M., & Shaw, R. (2003). College adults are not
good at self-regulation: a study on the relationship of self-regulation, note
taking, and test taking. Journal of Educational Psychology, 95(2), 335–346. http://
dx.doi.org/10.1037/0022-0663.95.2.335.
Piolat, A., Olive, T., & Kellogg, R. T. (2005). Cognitive effort during note taking.
Applied Cognitive Psychology, 19(3), 291–312. http://dx.doi.org/10.1002/
acp.1086.
Ponce, H. R., Lopez, M. J., & Mayer, R. E. (2012). Instructional effectiveness of a
computer-supported program for teaching reading comprehension strategies.
Computers & Education, 59(4), 1170–1183. http://dx.doi.org/10.1016/
j.compedu.2012.05.013.
Porter, A., McMaken, J., Hwang, J., & Yang, R. (2011). Common core standards: the
new US intended curriculum. Educational Researcher, 40(3), 103–116. http://
dx.doi.org/10.3102/0013189x11405038.
Rayner, K. (1998). Eye movements in reading and information processing: 20 years
of research. Psychological Bulletin, 124(3), 372–422. http://dx.doi.org/10.1037/
0033-2909.124.3.372.
Robinson, D. H. (1998). Graphic organizers as aids to text learning. Reading Research
and Instruction, 37(2), 85–105. http://dx.doi.org/10.1080/19388079809558257.
Stull, A. T., & Mayer, R. E. (2007). Learning by doing versus learning by viewing:
three experimental comparisons of learner-generated versus author-provided
graphic organizers. Journal of Educational Psychology, 99(4), 808–820. http://
dx.doi.org/10.1037/0022-0663.80.4.448.
Weinstein, C. E., & Mayer, R. E. (1985). The teaching of learning strategies. In M. C.
Wittrock (Ed.), Handbook of research on teaching (3rd ed., pp. 315–327). New
York: Macmillan.
Wittrock, M. C. (1989). Generative processes of comprehension. Educational
Psychologist, 24(4), 345.