Content uploaded by Gabriel A Radvansky
Author content
All content in this area was uploaded by Gabriel A Radvansky
Content may be subject to copyright.
The American Journal of Psychology
The American Journal of Psychology
Volume 124 • Number 1 • Spring 2011
Volume 124 • Number 1 • Spring 2011
Founded in 1887 by G. Stanley Hall
Edited by Robert W. Proctor Purdue University
Book Reviews Dominic W. Massaro University of California, Santa Cruz
Obituaries and History of Psychology Alfred H. Fuchs Bowdoin College
0002955692920112115341241010
0002-9556(201121)124:1;1-B
American Journal of Psychology
Spring 2011, Vol. 124, No. 1 pp. 63–73 • © 2011 by the Board of Trustees of the University of Illinois
Note Taking, Review, Memory,
and Comprehension
MARK BOHAY
University of Notre Dame
DANIEL P. BLAKELY
Florida State University
ANDREA K. TAMPLIN and GABRIEL A. RADVANSKY
University of Notre Dame
In previous work assessing memory at various levels of representation, namely the surface form,
textbase, and situation model levels, participants read texts but were otherwise not actively
engaged with the texts. The current study tested the influence of active engagement with the
material via note taking, along with the opportunity to review such notes, and the modality
of presentation (text vs. spoken). The influence of these manipulations was assessed both im-
mediately and 1 week later. In Experiment 1 participants read a text, whereas in Experiment 2
participants watched a video recording of the material being read as a lecture. For each experi-
ment the opportunity to take notes was manipulated within participants, and the opportunity
to review these notes before the test was manipulated between participants. Note taking im-
proved performance at the situation model level in both experiments, although there was also
some suggestion of benefit for the surface form. Thus, active engagement with material, such as
note taking, appears to have the greatest benefit at the deeper levels of understanding.
One of the dominant ideas about language compre-
hension and memory is that people represent infor-
mation at multiple levels, namely the surface form,
textbase, and situation model levels (van Dijk &
Kintsch, 1983). In short, the surface form is a per-
son’s verbatim memory of the words and syntax used.
The textbase is a representation of the abstract idea
units conveyed by language apart from the surface
form. Finally, the situation model is a person’s refer-
ential understanding of the described events. This
serves as a mental simulation and can include inferred
knowledge as well as information that was explicitly
presented (see also Johnson-Laird, 1983; Zwaan &
Radvansky, 1998). These three levels of representa-
tion are important because they dierentially con-
tribute to memory and comprehension (Kintsch,
Welsch, Schmalhofer, & Zimny, 1990; Radvansky,
Zwaan, Curiel, & Copeland, 2001). The aim of the
current study was to assess the inuence of increased
active engagement with a text via note taking, the op-
portunity to review such notes, and the inuence of
modality of presentation.
The bulk of the work on processing at these levels
of representation has focused on the reading of writ-
ten texts. More recent work on cognition has taken
a more functionalist approach, looking at how peo-
ple’s interaction with the environment aects their
cognition. This includes research on embodied and
grounded cognition (Barsalou, 1999, 2007; Glenberg,
1997; Wilson, 2002) that shows that a person’s sen-
sorimotor interactions with the world guide thought,
research on how memory has evolved to favor pro-
cessing of information that is relevant for survival
(Nairne & Pandeirada, 2008; Nairne, Pandeirada,
& Thompson, 2008; Nairne, Thompson, & Pandei-
rada, 2007), and research on how the structure of the
environment itself can alter basic cognitive processes,
such as memory (Copeland, Magliano, & Radvan-
sky, 2006; Radvansky & Copeland, 2006). For com-
prehension and memory, people can engage more
actively with a text by doing more than just reading.
One way to do this is by taking notes. Note taking
increases the degree to which a person attends to the
text, noting which ideas need to be jotted down and
which are better left unnoted. Thus, by more actively
engaging the reader with the material to be learned,
note taking may improve memory over conditions
when no notes are taken.
It is clear that note taking can improve later per-
formance (Annis, 1975; Dyer, Riley, & Yekovich, 1979;
Slotte & Lonka, 1999; Ward & Tatsukawa, 2003; see
Hartley, 1983, and Hartley & Davies, 1978, for a re-
view of this literature), both with text comprehension
and spoken lecture tasks (Nye, Crooks, Powley, &
Tripp, 1984; Slotte & Lonka, 1999). That said, note
taking is often not an idealized representation of the
information, and even successful students do not re-
cord most of the important ideas (Kiewra, 1985). So
the nature of the impact it is having on later perfor-
mance is unclear.
Our aim was to understand how the act of note
taking, as a way to increase engagement has with in-
formation, aects processing at the various levels of
representation. Several general possibilities are con-
sidered. First, it may be that the act of note taking itself
creates a dual task situation that redirects cognitive
resources that would otherwise be devoted to other
memory and comprehension processes and sends
them to the actions of scribbling something down. If
so, then note taking may compromise performance,
particularly at the situation model level, which re-
quires a person to integrate information that is being
presented along with inferences that are generated.
Thus, note taking may orient readers more toward
lower levels of processing because of increased at-
tentional focus at this level (Peper & Mayer, 1978).
This would occur at the cost of the situation model
level (Zwaan, 1994).
Alternatively, it has been suggested that, during
note taking, people elaborate on the material (Einstein,
Morris, & Smith, 1985; Ward & Tatsukawa, 2003), with
greater mental organization (Einstein et al., 1985) lead-
ing to improved memory (Craik & Lockhart, 1972).
If so, then one would expect to see improved perfor-
mance on the situation model level, as this is the level
where such elaboration is often done.
Because notes are often intended for later review,
they are, in some sense, self-generated memory cues.
Therefore, a further role that note taking may play is to
aid later performance (Oakhill & Davies, 1991; Slotte
& Lonka, 1999), even if the notes are not one’s own
(Annis, 1975; Kiewra, 1985), by reminding a person
of various aspects of the information. Additionally,
reviewing notes can serve as another learning trial,
thereby boosting memory performance. However, it
is unclear at what level this assistance would occur. It
could be that note review aids lower levels, such as the
surface form or textbase, because this specic mate-
rial is being exposed to a person once again through
the act of reviewing the notes. Alternatively, it may be
that review provides another opportunity for deeper
processing and elaboration, which would benet the
situation model level.
Briey, the aim of the current study was to assess
the inuence on memory, in terms of which level of rep-
resentation is most aected, when people were more
engaged during comprehension by taking notes. We
also assessed at what level the notes themselves serve as
memory cues to boost performance. Additionally, we
assessed performance from information learned in dif-
ferent media, because theories of comprehension have
been also been applied to video, lm, and interactive
video games (Copeland et al., 2006; Magliano, Rad-
vansky, & Copeland, 2007; Zacks, Speer, & Reynolds,
2009). Finally, to address the long-term impacts of such
manipulations, we tested memory both immediately
after learning and week later.
eXperiMent 1
The aim of Experiment 1 was to assess the inuence
of a more active engagement with a text via note tak-
64 • bohay et al.
ing for various levels of representation. Furthermore,
because people expect to use their notes later, we
also examined the impact of note review by allowing
one group to review their notes. Finally, we assessed
performance both immediately after reading and
week later.
METHOD
Participants
Ninety-seven people (53% female) were recruited
from the participant pool in the Department of Psy-
chology at the University of Notre Dame in exchange
for partial course credit. Forty-eight of these people
were not given the opportunity to review their notes
before testing, whereas the rest were.
Materials
There were three written texts. These texts covered
dierent topics: a summary of Waiting for Godot by
Samuel Beckett, an overview of the biological con-
cept of prions, and an overview of communism and
Russia’s economy from the s to the present. We
generated these texts using information from vari-
ous encyclopedia and Web-based sources (they were
reviewed and edited by two additional people for
coherence and clarity). Although there were three
texts, each participant saw only two of them. These
texts were , , and sentences (,, ,, and
1,392 words) long.
The memory measures were four-alternative,
multiple-choice recognition test questions. For each
topic, there were three types of questions. First, there
were questions that had verbatim answers drawn di-
rectly from the texts. Second were questions with
correct answers that were paraphrases of what was
stated in the text. Finally, the third group of questions
had correct answers that required an inference to be
made about the topic that was not explicitly stated in
the text. There were ve questions for each of these
three types of answers. The question types were ran-
domly ordered for each test. There were two versions
of each test for each topic. One of these versions was
used on the immediate test and the other for the de-
layed memory test. Both versions had questions,
with the same distribution of question types. These
versions were counterbalanced across participants.
Procedure
When participants arrived, they were given consent
forms and a survey for general demographic infor-
mation. They were then given instructions for the
experiment proper. They were told that they would
be reading two texts and that for each one they would
either take notes or not. In the no-notes condition,
people read the text but did not take notes. In the
written-notes condition, people took written notes
during reading.
Participants knew that their memory would be
tested. They were also told that they would be al-
lowed to review their notes before the test. For some
participants, no actual review was done, but they
were told that they would have this opportunity to
encourage them to take note taking seriously. The rest
of the participants were given min to review their
notes for the one text for which notes were taken.
These participants were also given min to review
their notes before the test after the -week delay.
After the notes were collected, participants took
two -question multiple-choice tests, corresponding
to the two texts that they read. The tests were pre-
sented in the same order in which the texts had been
read. A follow-up session week later consisted of a
dierent set of questions, which were also given in
the order in which the topics were read. There was
no time limit.
Analysis
To assess the inuence of these manipulations on
the various levels of representation, we used a varia-
tion of the Schmalhofer and Glavanov (1986) analysis
(see also Fletcher & Chrysler, 1990; Kintsch et al.,
1990; Radvansky, Copeland, & von Hippel, 2010;
Radvansky, Copeland, & Zwaan, 2003; Radvansky et
al., 2001; Zwaan, 1994). First, it should be noted that
this analysis uses four types of items: verbatim items
taken directly from a text; paraphrases of information
that was in the text but is now worded dierently,
inferences of information that were not explicitly
stated but are needed for the topic to be properly
understood, and incorrect items that are inconsis-
tent with a proper understanding of the material. The
primary variation used in the current study was a
four-alternative forced-choice test in which there was
a correct answer (either a verbatim, paraphrase, or
inference item) and three incorrect items.
This method uses a signal detection analysis,
in which performance on dierent items is used to
assess the dierent levels of representation. For the
surface form, positive responses to verbatim answers
are considered hits, and positive responses to para-
phrases are considered false alarms. This is done
not as a measure of how correct people are but as a
note taking and MeMory • 65
means of assessing the degree to which people use
dierent sources of knowledge. In this case, both the
verbatim and paraphrase answers can be answered
using situation model and textbase knowledge, but
only the verbatim answers can be directly boosted
by additional surface form knowledge because those
answers were actually present in the text. Thus, by
comparing performance on these two types of items
we can get a measure of the added benet of surface
form memory. Similarly, a comparison of the para-
phrase and inference answers provides an index of
the textbase because neither of these items was actu-
ally verbatim from the text, but only the paraphrases
capture ideas that were presented in the text, whereas
the inferences do not. Finally, a comparison of posi-
tive responses with inferences and incorrect answers
provides an index of the situation model because nei-
ther of these convey idea units that were actually in
the text, but only the inferences are consistent with
the information in the text and the incorrect answers
are not.
RESULTS AND DISCUSSION
Prior Knowledge
The prior knowledge data showed that participants
had little or no previous familiarity with the material.
A scale ranging from (no knowledge) to 5 (extensive
knowledge) was used. The familiarity ratings were
M = 1.26 for Waiting for Godot, M = 1.30 for prions,
and M = 1.36 for Russia. Analyses indicated that there
was no signicant relationship between prior knowl-
edge and test performance.
Memory
The memory test accuracy data (Table 1) were used to
calculate the A' measures (Table 2), which were submit-
ted to separate 2 (review or not)×2 (delay)× (note
condition) mixed s for each level of representa-
tion, with the rst variable being between participants
and the rest within. For the surface form measure, the
only signicant eect was the main eect of note condi-
tion, F(1, 95) = 5.18, MSE = 0.05, p = .03, with greater
surface form memory when notes were taken (M = .58)
than when they were not (M = .52). Essentially, when
participants did not take notes, their performance
was at chance. Thus, note taking does boost verbatim
memory. The main eects of review and delay were not
signicant, F = 1.04 and F < , respectively, nor were
any of the interactions, all Fs < .
For the textbase level, there were no signicant
main eects of review, F(1, 95) = 2.37, MSE = 0.05,
p = .13; delay, F < 1; or note condition, F < . Although
the two-way interactions were not significant, all
ps > ., the three-way interaction was marginally sig-
nicant, F(1, 95) = 3.38, MSE = 0.05, p = .07. This is
probably due to meaningless variation because none
of these textbase scores were signicantly dierent
from chance, all ps > ., except for the no-notes
condition after a -week delay in the group that was
allowed to review their notes. Even here, performance
was signicantly below chance. Thus, none of these
manipulations had a unique eect on memory at the
textbase level.
Finally, for the situation model level, there was
a main eect of delay, F(1, 95) = 39.10, MSE = 0.04,
p < ., with performance worse after week
(M = .62) than when participants were tested imme-
diately (M = .75). Essentially, participants had more
diculty retrieving a deeper understanding of the ma-
terial a week later. There was also a main eect of hav-
ing taken notes, F(1, 95) = 7.57, MSE = 0.04, p = .007,
with performance being better when participants took
notes (M = .72) than when they did not (M = .66).
The interaction between these two variables was not
signicant, F < . Although the main eect of review
was not signicant, it did interact with having taken
notes or not, F(1, 95) = 4.68, MSE = 0.04, p = .03.
Essentially, as would be expected, for the topics on
which participants had taken notes, being allowed to
review them later (M = .75) served as a retrieval cue,
which improved performance relative to those who
did not review the notes (M = .70).
In summary, interacting with a text by taking notes
improved performance for verbatim information and
led to deeper understanding at the situation model
level. There was no clear benet at the textbase level.
Note that chance performance on this measure does
not mean that participants had no textbase level rep-
resentation, but memory at this level did not contrib-
ute meaningfully to performance on the memory tests
over and above that found at the surface form and
situation model levels. In addition to the inuence of
note taking, the review of these notes appeared to aid
performance only at the situation model level. Finally,
the -week delay in testing aected performance only
at the situation model level. This may be because the
contributions of the surface form and textbase levels
66 • bohay et al.
were so weak to begin with that they could not decline
much, even after a week of forgetting.
eXperiMent 2
Experiment 2 was an extension of Experiment 1. How-
ever, rather than reading texts, participants received the
information in a lecture format (the original texts were
written to be read as lectures anyway). Furthermore, in
addition to written notes, there was a third condition
in which participants also took typed notes. This was
done because although most research has focused on
manual note taking, advances in technology have led to
more students taking notes on laptop computers. Ward
and Tatsukawa (2003) suggested that typing may be
more eective for note taking because it allows search-
ing, editing, sharing, and greater legibility. In addition,
typing may be more ecient, less tiring, and faster than
written note taking, and the faster speed of typing al-
lows people to record more information.
METHOD
Participants
For Experiment 2, 77 people (53% female) were
tested. Thirty-six were not allowed to review their
notes, and the other were. All participants were
drawn from the participant pool in the Department
of Psychology at the University of Notre Dame in
exchange for partial course credit.
Materials and Procedure
The same materials were used as in Experiment 1,
except that readings of the lectures were videotaped.
Each lecture was 9 min long. All three lectures were
presented by the same male professor. There was no
additional information on the videotape (e.g., slides,
blackboard writing).
When participants arrived, they were given con-
sent forms and a survey for general demographic in-
formation. They were then given instructions for the
experiment proper. They were told that they would
view three -min videotaped lectures on the left half
of a computer screen, with the right half open to a
word processing program (for the computer notes
condition), and that for each lecture they would take
either no notes, written notes, or typed notes on a
computer. The no-notes and written-notes condi-
tions were like those in Experiment 1. In the typed-
notes condition, participants took notes on the right
table 1. Accuracy levels by answer type, Experiment 1
Verbatim Paraphrase Inference Wrong
No review
Immediate
No notes .70 .69 .70 .30
Written notes .81 .71 .69 .26
Delay
No notes .60 .61 .57 .41
Written notes .66 .58 .58 .39
Review of notes
Immediate
No notes .71 .66 .63 .33
Written notes .80 .72 .74 .25
Delay
No notes .56 .49 .58 .45
Written notes .69 .58 .63 .37
table 2.
A
' measures (
SE
) for the various levels of representa-
tion, Experiment 1
Situation
Surface form Textbase model
No review
Immediate
No notes .51 (.03) .48 (.03) .76 (.03)*
Written notes .58 (.03)* .52 (.04) .77 (.03)*
Delay
No notes .49 (.03) .53 (.03) .61 (.03)*
Written notes .57 (.04)* .50 (.03) .62 (.04)*
Review
Immediate
No notes .54 (.03) .53 (.03) .69 (.03)*
Written notes .57 (.03)* .48 (.03) .79 (.03)*
Delay
No notes .55 (.03) .42 (.03)* .58 (.04)*
Written notes .58 (.04)* .48 (.03) .68 (.03)*
*
p
< .05.
note taking and MeMory • 67
half of the computer screen. All participants were in
each of the three note-taking conditions, and condi-
tion and lecture type were counterbalanced across
participants. Participants viewed the lectures indi-
vidually on a computer. After participants viewed all
the lectures, the experimenter collected the notes.
After the notes were collected, participants com-
pleted three -question multiple-choice tests corre-
sponding to the three topics. The tests were presented
in the same order in which the topics were viewed. As
in Experiment 1, the follow-up session 1 week later con-
sisted of a dierent set of multiple-choice tests, which
were also given in the order in which participants
viewed the topics. Participants were given an unlimited
amount of time to nish these tests. Again, although
all participants were told that they would be allowed
to review their notes, only some actually were.
RESULTS AND DISCUSSION
Prior Knowledge
The prior knowledge data showed that the partici-
pants had little or no previous knowledge of lecture
material. Again, using a scale from (no knowledge)
to (extensive knowledge), the means were Waiting
for Godot, M = 1.3; prion lecture, M = 1.2; and Russia
lecture, M = 1.2.
Memory
The memory test accuracy data (Table 3) were used to
calculate the A' measures (Table 4), which were sub-
mitted to separate 2 (review or not)×2 (delay)×
(note condition) mixed , with the rst variable
being between participants and the rest within. For
the surface form measure, there were no signicant
eects, although the main eect of review was margin-
ally signicant, F(1, 75) = 3.76, MSE = 0.05, p = .06.
Surprisingly, performance was actually better overall
if people were not given the opportunity to review
their notes (M = .55) than if they were (M = .52). It is
unclear why this was the case. Because of the oddness
of this nding, and the fact that we never replicated it
anywhere else, we assign little weight to it and suggest
that it might be a Type I error.
There was also a marginally signicant eect of
note condition, F(2, 150) = 2.37, MSE = 0.05, p = .10.
Comparisons of the dierent note conditions re-
vealed marginally signicant dierences of the no-
notes condition (M = .51) with the written-notes
(M = .56) and typed-notes (M = .55) conditions, F(,
75) = 3.58, MSE = 0.06, p = .06, and F(1, 75) = 3.05,
MSE = 0.04, p = .09, respectively, whereas the last
two conditions did not differ from one another,
F < . So although there was some benet to taking
notes at the surface form level when it came to an-
swers taken directly from the lecture, this is a weak
eect. The main eect of delay was not signicant,
F(1, 75) = 1.74, MSE = 0.05, p = .19, nor were any of
the interactions, all ps > ..
For the textbase level, there were no signicant
main eects of review, F < 1; delay, F(1, 75) = 1.98,
MSE = 0.06, p = .16; or note condition, F < . None
of the interactions were signicant, F < .
Finally, for the situation model level, there was
a main eect of delay, F(1, 75) = 13.79, MSE = 0.04,
p < ., with performance growing worse over week
(M = .60) compared with immediate testing (M = .67).
Again, people had more diculty retrieving an under-
standing of the content that they had learned a week
earlier. There was also a main eect of taking notes,
F(2, 150) = 6.65, MSE = 0.04, p = .002. Comparisons
Table 3. Accuracy levels by answer type, Experiment 2
Verbatim Paraphrase Inference Wrong
No review
Immediate
No notes .62 .60 .55 .41
Written notes .72 .67 .66 .32
Typed notes .75 .63 .64 .33
Delay
No notes .63 .51 .52 .45
Written notes .58 .48 .52 .47
Typed notes .63 .59 .61 .39
Review
Immediate
No notes .56 .64 .60 .40
Written notes .76 .70 .64 .30
Typed notes .67 .68 .65 .33
Delay
No notes .54 .57 .56 .44
Written notes .65 .55 .61 .40
Typed notes .67 .59 .58 .39
68 • bohay et al.
of the dierent note conditions revealed signicant
dierences of the no-notes condition (M = .59) with
the written-notes (M = .65) and typed-notes (M = .67)
conditions, F(1, 75) = 6.60, MSE = 0.04, p = .01, and
F(1, 75) = 12.17, MSE = 0.04, p = .001, respectively,
whereas the last two conditions did not dier from
one another, F = 1. So, as in Experiment 1, there was
a clear benet to taking notes at the situation model
level. Neither the main eect of review nor any of the
interactions were signicant, all ps > ..
Overall the results of Experiment 2 show that note
taking improved performance. Importantly, this ben-
et had the largest impact at the situation model level.
These results support previous work that suggests
that note taking can improve learning even without
later review (Annis, 1975; Ward & Tatsukawa, 2003).
Moreover, there was no meaningful dierence between
the written- and typed-notes conditions.
Analysis of Note Quantity
One concern that may arise is that people performed
dierently based on how extensive their notes were.
To address this concern, we tallied the amount of
material contained in each participant’s notes and
compared it with test performance in Experiments
and . This was done by rst breaking the lectures
down into the propositional idea units and then
enumerating the number of idea units in each par-
ticipant’s notes (for a similar process, see Hartley &
Cameron, 1967). These measures of note quantity
were then compared across the various conditions
and with performance on the primary task.
The mean number of idea units present in the
written notes from Experiment 1 was M = 40.4. In
Experiment 2, people in the written-notes group had
a mean of M = 36.5 idea units, whereas those in the
typed-notes group had a mean of M = 43.6 idea units.
The mean numbers of idea units in the notes are
reported in Table . The data for two participants
in Experiment 1 were excluded because their notes
were lost.
To assess whether the extensiveness of the notes
was related to performance on the memory tests, the
data were submitted to a correlation analysis. The
relevant note quantity scores for the review and delay
conditions in Experiment 1 and for the review, delay,
and note conditions for Experiment 2 were com-
pared with the surface form, textbase, and situation
table 4.
A
' measures for the various levels of representation,
Experiment 2
Surface form Textbase Situation model
No review
Immediate
No notes .52 .55 .60*
Written notes .55 .50 .70*
Typed notes .60* .49 .70*
Delay
No notes .59* .49 .55
Written notes .56 .48 .54
Typed notes .53 .47 .65*
Review
Immediate
No notes .45 .53 .62*
Written notes .55 .53 .71*
Typed notes .49 .52 .70*
Delay
No notes .48 .51 .59*
Written notes .59* .46 .65*
Typed notes .57 .50 .63*
*
p
< .05.
model level memory scores on the tests correspond-
ing with each of these conditions. For Experiment 1,
there was a signicant correlation in the no-review
condition with immediate situation model memory
performance, r(46) = .47, p = .001. No signicant cor-
relations were found when participants were allowed
to review their notes, all ps ≥ ..
When note quantity was compared with memory
test performance in Experiment 2, no signicant cor-
relations were found. However, note quantity was
Table 5. Mean note quantity (in mean numbers of
idea units recalled)
Experiment 1 Experiment 2
Written notes 40.4 36.5
Typed notes n/a 43.6
Mean 40.4 40.1
note taking and MeMory • 69
marginally correlated with immediate surface form
memory, but in the negative direction, r(41) = –.27,
p = .086, and immediate situation model memory
in the positive direction, r(41) = .28, p = .08, when
participants reviewed their notes. The remainder of
the correlations did not approach signicance, all
ps ≥ .. Thus, the extensiveness of the notes showed
some relation to memory performance, but the cor-
relations were weak and inconsistent.
general disCussion
This study assessed how various levels of processing
are inuenced by a more active interaction with the
information, as with note taking. Moreover, we also
looked at the impact of being able to review such
notes, the retention interval, and the use of both text
and video presentation on later performance.
First, this study conrmed that note taking does
aid performance (see Nye et al., 1984). In both ex-
periments, there was always evidence of some kind
of benet. This was clearly present at the situation
model level, although there was also some small bene-
t at the surface form level. Thus, it appears that note
taking facilitates deeper comprehension, consistent
with previous research that suggests that people who
take notes are more likely to elaborate on the material
(Einstein et al., 1985; Ward & Tatsukawa, 2003).
Second, not surprisingly, the notes themselves
can serve as eective memory cues. However, what
is surprising is that this inuence was not pervasive.
It clearly appeared only at the situation model level
in Experiment 1 when people read a text. With the
spoken lectures in Experiment 2 it manifested itself
only at the surface form level, and even then it was
marginal and showed the reverse pattern. Although
there may be some benet of reviewing notes, it did
not have a profound eect, at least in the context of
the current study.
Third, there was also an impact of delay on per-
formance, with people remembering less a week later
than when tested immediately. However, the inuence
of this delay was clearly observed only at the situation
model level. This may be due in part to the fact that
the inuence of the surface form and textbase levels on
performance was generally low, in many cases at chance
level, so a decline in performance after a -week delay
would be much more dicult to observe.
Finally, the manipulation of whether the informa-
tion was presented as a text or a lecture did reveal some
dierences on later performance, but these were minor.
Looking at the overall pattern of data, it is clear that
note taking had its primary inuence on the situation
model level, with a smaller inuence at the surface
form level. Thus, the patterns of data we observed
here would be attributed to general comprehension
processes, not to reading or listening per se.
Looking at the notes themselves, our analysis
suggests that more note taking provides a later per-
formance benefit. However, the evidence of such
benets was weak and inconsistent. Thus, at least
under the circumstances used here, it seems that to
improve test performance, it is of primary importance
that a person be actively attending to the material by
taking notes, but the amount of information noted is
of secondary importance.
Overall, these results support the idea that when
people are more actively engaged in material that
they are trying to learn, such as by taking notes,
their memory improves, particularly at deeper levels
of comprehension, as with the situation model level.
This is consistent with research showing that note
taking can improve performance (Annis, 1975; Dyer et
al., 1979; Nye et al., 1984; Slotte & Lonka, 1999; Ward
& Tatsukawa, 2003). It should also be noted that if
students are given handouts with parts to be lled
out during comprehension lecture, this can improve
memory even further (Hartley, 1976; Larson, 2009;
Sambrook & Rowley, 2010), perhaps by making the
material better organized and focusing students’ at-
tention on the more important concepts. More gener-
ally, the results reported here are broadly consistent
with the idea that memory can be improved when
people are actively engaged with the material to be
learned, even if that engagement is simply the noting
down of the important ideas.
More specically, there was no support for the
idea that the act of note taking produces a dual task
situation that draws mental resources away from
memory and comprehension processes, thereby
focusing attention at the surface form and textbase
levels. Instead, the fact that the bulk of the memory
improvement we observed was at the situation model
level is consistent with previous research suggesting
that note taking may encourage people to elaborate
on and better structure and organize the material
70 • bohay et al.
being learned (Einstein et al., 1985; Ward & Tat-
sukawa, 2003). This is because the act of situation
model creation requires a person to draw inferences
about elements that are not explicitly created and to
bring together various elements of the information.
More generally, these ndings are broadly consistent
with a wide range of ndings that active interaction
improves memory, as with research on grounded cog-
nition (Barsalou, 1999, 2007; Glenberg, 1997; Wilson,
2002), the benets of survival focus encoding (Nairne
& Pandeirada, 2008; Nairne et al., 2007, 2008), and
event-based mental updating (Copeland et al., 2006;
Radvansky & Copeland, 2006).
One concern about the current research is that all
participants were repeatedly tested. Recent work has
shown that prior testing can improve performance on
later memory tests. This is called the testing eect
(Chan, McDermott, & Roediger, 2006). The testing
eect is well established, and it is present for all the
conditions assessed here, and so it is a constant. If our
data reect some impact of repeated testing (thereby
reducing the forgetting rate to some degree), there is
no clear evidence to date that the testing eect dif-
ferentially inuences the various levels of representa-
tion. Moreover, the testing eect has been found to
be only marginally signicant (p = .09) when, instead
of a recall test, a forced-choice recognition is used
on the initial test (McDaniel, Anderson, Derbish, &
Morisette, 2007), as was used here. Therefore, there
is no reason to believe that the testing eect is driving
the primary results reported here.
In conclusion, this study showed that perfor-
mance improved when people more actively engage
with the information, such as by taking notes. This
improvement was observed primarily at the situa-
tion model level, where deeper understanding oc-
curs. This increased engagement facilitated more
elaborative mental processing of the material, result-
ing in superior situation models that allowed people
to more accurately respond to subsequent memory
test questions.
NOTE
Thanks to Dr. T. Merluzzi for graciously agreeing to be the
lecturer and to Allie Payne for lming and editing the lec-
tures; and to Erica Nason, Alison Dunleavy, Brad Bitterly,
Josephine Kim, Katie Harris, Amanda Caravalho, Ammar
Tahir, and Jessica Harrison for help with the data collection.
Address correspondence about this article to G. A. Rad-
vansky, -C Haggar, Department of Psychology, University
of Notre Dame, Notre Dame, IN (e-mail: gradvans@
nd.edu).
REFERENCES
Annis, L. (1975). The eects of encoding and an external
memory device on note-taking. Journal of Experimental
Education, 44, –.
Barsalou, L. W. (1999). Perceptual symbol systems. Behav-
ioral and Brain Sciences, 22, –.
Barsalou, L. W. (2007). Grounded cognition. Annual Review
of Psychology, 59, –.
Chan, J. C. K., McDermott, K. B., & Roediger, H. L. (2006).
Retrieval-induced facilitation: Initially nontested material
can benet from prior testing of related material. Journal
of Experimental Psychology: General, 135, –.
Copeland, D. E., Magliano, J. P., & Radvansky, G. A. (2006).
Situation models in comprehension, memory, and
augmented cognition. In M. Bernard, J. C. Forsythe, &
T. Goldsmith (Eds.), Human cognitive models in system
design (pp. 37–66). Mahwah, NJ: Erlbaum.
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of process-
ing: A framework for memory research. Journal of Verbal
Learning and Verbal Behavior, 11, –.
Dyer, J. W., Riley, J., & Yekovich, F. R. (1979). An analysis
of three study skills: Note-taking, summarizing and re-
reading. Journal of Educational Research, 73, –.
Einstein, G. O., Morris, J., & Smith, S. (1985). Note-taking,
individual dierences, and memory for lecture informa-
tion. Journal of Educational Psychology, 77, –.
Fletcher, C. R., & Chrysler, S. T. (1990). Surface forms,
textbases, and situation models: Recognition memory for
three types of textual information. Discourse Processes, 13,
–.
Glenberg, A. M. (1997). What memory is for? Behavioral and
Brain Sciences, 20, –.
Hartley, J. (1976). Lecture handouts and student notetaking.
Programmed Learning & Educational Technology, 13(2),
–.
Hartley, J. (1983). Note-taking research: Re-setting the score-
board. Bulletin of the British Psychological Society, 36,
–.
Hartley, J., & Cameron, A. (1967). Some observations of the
eciency of lecturing. Educational Review, 20(1), 30–37.
Hartley, J., & Davies, I. K. (1978). Note-taking: A critical re-
view. Programmed Learning & Educational Technology,
15(3), 207–224.
Johnson-Laird, P. N. (1983). Mental models: Towards a cog-
nitive science of language, inference, and consciousness.
Cambridge, MA: Harvard University Press.
Kiewra, K. A. (1985). Providing the instructor’s notes: An
eective addition to student note-taking. Educational Psy-
chologist, 20, –.
Kintsch, W., Welsch, D., Schmalhofer, F., & Zimny, S. (1990).
note taking and MeMory • 71
Sentence memory: A theoretical analysis. Journal of
Memory and Language, 29, –.
Larson, R. B. (2009). Enhancing the recall of presented mate-
rial. Computers & Education, 53, –.
Magliano, J., Radvansky, G. A., & Copeland, D. E. (2007).
Beyond language comprehension: Situation models as
a form of autobiographical memory. In F. Schmalhofer
& C. Perfetti (Eds.), Higher level language processes in
the brain: Inference and comprehension processes (pp.
379–392). Mahwah, NJ: Erlbaum.
McDaniel, M. A., Anderson, J. L., Derbish, M. H., & Mor-
isette, N. (2007). Testing the testing eect in the class-
room. European Journal of Cognitive Psychology, 19,
–.
Nairne, J. S., & Pandeirada, N. S. (2008). Adaptive memory:
Remembering with a stone-age brain. Current Directions
in Psychological Science, 17, –.
Nairne, J. S., Pandeirada, N. S., & Thompson, S. R. (2008).
Adaptive memory: The comparative value of survival pro-
cessing. Psychological Science, 19, –.
Nairne, J. S., Thompson, S. R., & Pandeirada, N. S. (2007).
Adaptive memory: Survival processing enhances reten-
tion. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 33, –.
Nye, P. A., Crooks, T. J., Powley, M., & Tripp, G. (1984).
Student note-taking related to university examination per-
formance. Higher Education, 13, –.
Oakhill, J., & Davies, A. (1991). The eects of test expectancy
on quality of note-taking and recall of text at dierent
times of day. British Journal of Psychology, 82, –.
Peper, R. J., & Mayer, R. E. (1978). Note taking as a generative
activity. Journal of Educational Psychology, 70, –.
Radvansky, G. A., & Copeland, D. E. (2006). Walking
through doorways causes forgetting. Memory & Cogni-
tion, 34, –.
Radvansky, G. A., Copeland, D. E., & von Hippel, W. (2010).
Stereotype activation, inhibition, and aging. Journal of
Experimental Social Psychology, 46, –.
Radvansky, G. A., Copeland, D. E., & Zwaan, R. A. (2003).
Aging and functional spatial relations in comprehension
and memory. Psychology and Aging, 18, –.
Radvansky, G. A., Zwaan, R. A., Curiel, J. M., & Copeland,
D. E. (2001). Situation models and aging. Psychology and
Aging, 16, –.
Sambrook, S., & Rowley, J. (2010). What’s the use of web-
notes? Student and sta perceptions. Journal of Further
& Higher Education, 34(1), 119–134.
Schmalhofer, F., & Glavanov, D. (1986). Three components
of understanding a programmer’s manual: Verbatim,
propositional, and situational representations. Journal of
Memory and Language, 25, –.
Slotte, V., & Lonka, K. (1999). Review and process eects of
spontaneous note-taking on text comprehension. Contem-
porary Educational Psychology, 24, –.
van Dijk, T. A., & Kintsch, W. (1983). Strategies in discourse
comprehension. New York, NY: Academic Press.
Ward, N., & Tatsukawa, H. (2003). A tool for taking class
notes. International Journal of Human–Computer Stud-
ies, 59, –.
Wilson, M. (2002). Six views of embodied cognition. Psy-
chonomic Bulletin & Review, 9, –.
Zacks, J. M., Speer, N. K., & Reynolds, J. R. (2009). Seg-
mentation in reading and lm comprehension. Journal of
Experimental Psychology: General, 138, –.
Zwaan, R. A. (1994). Eect of genre expectations on text
comprehension. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 20, –.
Zwaan, R. A., & Radvansky, G. A. (1998). Situation models in
language comprehension and memory. Psychological Bul-
letin, 123, –.
72 • bohay et al.
appendiX. saMple teXt and test iteMs For
the waiting For godot leCture
Though dicult and sometimes baing to read or (even) view, Waiting for Godot, Samuel Beckett’s rst play,
was written originally in French in as En Attendant Godot. It premiered at a tiny theater in Paris in .
The play’s uniqueness compelled the audiences to ock to the theaters for a spectacularly continuous four hun-
dred performances. Waiting for Godot is one of the most important works of our time; revolutionizing theater in
the twentieth century and having a profound inuence on generations of succeeding dramatists, including such
renowned contemporary playwrights as Harold Pinter and Tom Stoppard. After the appearance of Waiting for
Godot, theater was opened to possibilities that playwrights and audiences had never before imagined. Waiting
for Godot was a unique outburst on the literary world. It made no claim to have a place in conventional drama;
rather, it carried a “fascination” of its own, authenticated by the undercurrent of resentment in accepting the
illogical and unreasonable norms of the society. This play came to be considered an essential example of what
Martin Esslin later called “Theater of the Absurd,” a term that Beckett disavowed but which remains a handy
description for one of the most important theater movements of the twentieth century. “Absurdist Theater”
discards traditional plot, characters, and action to assault its audience with a disorienting experience. Characters
often engage in seemingly meaningless dialogue or activities, and, as a result, the audience senses what it is like
to live in a universe that doesn’t “make sense.” Beckett and others who adopted this style felt that this disorient-
ed feeling was a more honest response to the post World War II world than the traditional belief in a rationally
ordered universe. Waiting for Godot remains the most famous example of this form of drama. The play opens on
a totally surreal note, with a tramp trying to pull o his boot on a lonely road under a leaess tree. There is no
horizon, no sign of civilization. For a moment, this scene might even be considered comic. Eventually Vladimir
enters and greets Estragon who informs Vladimir that he has spent the night in a ditch where he was beaten.
They are very happy to see each other, having been separated for an unspecied amount of time. Estragon has a
sore foot and is having trouble taking his boot o. . . .
Verbatim test item
This type of theater that discards traditional plot, characters, and action to assault its audience with disorienting
experience is called:
a. Ridiculous
b. Absurdist (correct)
c. Bizarre
d. Strange
Paraphrase test item
_______ spends the night in a ditch where he was beaten.
a. Vladimir
b. Estragon (correct)
c. Pozzo
d. Lucky
Inference test item
Audiences really responded to the play because it ________.
a. Was a unique outburst on the literary world
b. Didn’t accept illogical and unreasonable norms
c. It was not a conventional drama
d. All of the above (correct)
note taking and MeMory • 73
