10.

1177_1747021819879434
research-article2019
QJP0010.1177/1747021819879434The Quarterly Journal of Experimental PsychologyBellinger and DeCaro

Original Article

Quarterly Journal of Experimental

Note-taking format and difficulty Psychology

2019, Vol. 72(12) 2807­–2819
© Experimental Psychology Society 2019
impact learning from Article reuse guidelines:
sagepub.com/journals-permissions
instructor-provided lecture notes DOI: 10.1177/1747021819879434
https://doi.org/10.1177/1747021819879434
qjep.sagepub.com

David B Bellinger and Marci S DeCaro

Abstract
The method students use to take notes impacts how they process lecture information. The current experiment examined
how the format and amount of content included in instructor-provided notes affect learning. Undergraduate students
listened to a brief audio-recorded science lecture that emphasised independent facts, while using one of four note-taking
guides. These guides varied in their format (outline notes, cloze notes) and level of difficulty (less-difficult, more-difficult).
Outline notes included a partially complete organisational framework, promoting knowledge of relationships among
concepts. Cloze notes included all lecture content with select words missing, encouraging processing of specific details.
Metacognitive ratings and an objective cognitive load measure confirmed that outline note-taking was the most difficult
method. However, outline notes led to higher performance than cloze notes on free recall and inference questions, and
equal performance on verbatim questions. These benefits were greatest in the more-difficult outline notes condition,
when less information was provided. These findings are consistent with the material-appropriate difficulty framework.
Increasing note-taking difficulty was desirable, but only when the activity elicited semantic processing that complemented
the type of processing afforded by the learning material.

Keywords
Learning; memory; note-taking; desirable difficulty; metacognition; material-appropriate difficulty; cognitive load

Received: 27 June 2018; revised: 15 May 2019; accepted: 8 June 2019

Lecturing is the most frequently used teaching method in encourage them to actively make sense of the information
higher education (Chen, 2002; Svinicki & McKeachie, by engaging in the necessary cognitive processes (i.e.,
2011). Many students choose to take notes during lectures select the appropriate information, organise it by making
(Van Meter, Yokoi, & Pressley, 1994), even without being connections between the ideas presented and the material’s
instructed to do so (Williams & Eggert, 2002), because they underlying structure, and integrate it with their prior
believe it will help them learn the information (Dunkel & knowledge) to bolster learning.
Davy, 1989). However, students’ notes are highly variable To counter these concerns, some instructors have adopted
and typically inaccurate or incomplete (e.g., Kiewra, a modified approach: provide students with guided, but
Dubois, Christensen, Kim, & Lindberg, 1989). incomplete, notes. These notes typically take one of two
For these reasons, some instructors provide lecture formats: cloze notes and skeletal outlines (Austin et al.,
notes, which is associated with better exam performance 2002; Boyle, 2012; Katayama & Robinson, 2000; Morgan,
(e.g., Armbruster, 2009; Kiewra, 1985b). By freeing stu- Lilley, & Boreham, 1988; Neef, McCord, & Ferreri, 2006;
dents from recording the lecture content, students can Williams, Weil, & Porter, 2012). Cloze notes include the
engage in more semantic processing as well as ask and majority of the lecture content, but essential words are
answer questions during lecture (Austin, Lee, Thibeault,
Carr, & Bailey, 2002). The downside of complete instruc- Department of Psychological and Brain Sciences, University of
tor-provided notes, however, is that students may be less Louisville, Louisville, KY, USA
likely to attend lectures (e.g., Cornelius, Owen-DeSchryver,
Corresponding author:
2008). Also, based on Fiorella and Mayer’s (2016) defini- Marci S DeCaro, Department of Psychological and Brain Sciences,
tion of generative learning, students may receive lim- University of Louisville, 2301 S. 3rd Street, Louisville, KY 40292, USA.
ited benefits from complete notes because they do not Email: [email protected]
2808 Quarterly Journal of Experimental Psychology 72(12)

replaced with a blank space and require students to fill in Paas, 1998). Providing lecture notes to students as a learn-
the missing words as they listen to the lecture. Skeletal out- ing aid may reduce the load on working memory, and thus
lines (hereafter, “outlines”) provide students with an organ- set the conditions to allow for learning to occur. In contrast,
isational framework for the lecture, requiring students to a related concept of desirable difficulty emphasises the
fill in the main and/or supporting ideas as they listen to the learning benefits of generating information for oneself,
lecture. The current experiment examined how the format despite introducing greater difficulty for the learner (Bjork
and completeness of instructor-provided notes impacts & Bjork, 2011; Richland, Bjork, Finley, & Linn, 2005).
memory for lecture material. Considering both cognitive load theory and the desirable
difficulty framework, instructors are encouraged to max-
Lecture note-taking (sometimes) impacts imise encoding and retrieval processes that provide a pro-
ductive level of difficulty without overwhelming cognitive
memory resources (e.g., Van Merrienboer & Ayres, 2005). However,
Note-taking can facilitate learning at two time points: numerous studies demonstrate that not all difficulty during
while initially taking the notes (encoding benefit) and learning is desirable—sometimes having students generate
while reviewing the notes at a later time (external storage information themselves benefits learning, and sometimes
benefit; DiVesta & Gray, 1972). The current work focused it does not (McDaniel & Butler, 2011). Furthermore, mem-
on the encoding benefits of note taking. Strategies for ory performance can only be explained by the underlying
improving the external storage benefit of notes (e.g., memory processes and neither cognitive load theory nor
spaced retrieval practice) are well documented (e.g., the desirable difficulty concept specify the memory mech-
Karpicke & Roediger, 2010). Given the ubiquitous use of anisms involved (Mitchell & Hunt, 1989). To reach this
note-taking as a learning strategy, it is surprising that sys- level of understanding, careful attention must be given to
tematic reviews reveal only a slight encoding benefit of both the amount and type of semantic processing that stu-
note-taking relative to no note-taking (d = .26; Kobayashi, dents engage in while learning.
2005), or a lack of consensus for the encoding benefit of Hunt and colleagues (Einstein & Hunt, 1980; Hunt,
note-taking (Kiewra, 1985a). For example, Kiewra (1985a) 2003, 2013; Hunt & Einstein, 1981) provide a framework
found an encoding benefit in 33 out of 56 studies (59%), for understanding the qualitative aspect of information by
meaning that a sizable number of studies found no differ- distinguishing between two types of semantic processing:
ences (21 studies; 37%) or a detrimental effect (2 studies; item-specific and relational processing. When students
4%) of note-taking. These findings suggest that any encod- engage in item-specific processing, they focus on distinc-
ing benefit of note-taking may depend on important mod- tive features of the information such as key terms or indi-
erating factors. Drawing on the material-appropriate vidual ideas. In contrast, when students engage in relational
difficulty framework (McDaniel & Butler, 2011; McDaniel processing, they focus on the organisation and connections
& Einstein, 1989, 2005; McDaniel, Einstein, & Lollis, among ideas. The type of processing used determines
1988), we investigate two such factors: the level of diffi- which aspect of the information students encode in mem-
culty induced by note-taking, and note-taking format. We ory. Different types of learning materials, encoding tasks,
examine how these factors interact to impact performance and memory tests elicit or depend on item-specific or rela-
on different types of memory tests. tional processing to varying extents.
McDaniel and colleagues (McDaniel & Butler, 2011;
Understanding when difficulty during note- McDaniel & Einstein, 1989, 2005; McDaniel et al., 1988)
proposed the material-appropriate difficulty framework as
taking is desirable a guide for determining whether difficulty during learning
One factor that may be important for determining the ben- will be beneficial. Specifically, they extend Jenkins’ (1979)
efits of instructor-provided notes is the level of support tetrahedral model of memory to argue that the benefits of
provided, which determines how easy or difficult the note- difficulty will depend on interactions between the process-
taking task is as well as the amount of generative process- ing induced by four clusters of variables: the learning
ing students engage in. Note-taking requires students to material, the encoding task, the memory test, and individ-
maintain information in their focus of attention while ual differences. In our study, we investigate the first three
recording other information, which depends highly on factors and control for individual differences.
working memory resources (Bui & Myerson, 2014; Bui, According to the material-appropriate difficulty frame-
Myerson, & Hale, 2013; Engle, Tuholski, Laughlin, & work, memory is enhanced when students complete an
Conway, 1999; Piolat, Olive, & Kellogg, 2005). encoding task that elicits semantic processing which com-
According to cognitive load theory, learning is enhanced plements (i.e., is not redundant with) the type of process-
if instructional activities avoid surpassing students’ cogni- ing afforded by the learning material. For example,
tive limits, such as providing students with a partial solu- McDaniel, Einstein, Dunay, and Cobb (1986) found that
tion to a problem (Paas, 1992; Sweller, van Merrienboer, & students learned better from a text passage that elicited
Bellinger and DeCaro 2809

item-specific processing if the encoding task prompted et al., 1986). In contrast, outline notes explicitly organised
relational processing, as opposed to item-specific process- the prose into a hierarchy of main and supporting ideas,
ing. Students could better consider both the general rela- and required students to fill in the missing information
tions between ideas (from the encoding task) as well as denoted by blanks. Previous research demonstrated that
individuating information (e.g., specific facts or words, using outlines induces relational processing (Einstein,
from the text passage), which together improves encoding McDaniel, Owen, & Cote, 1990). Because the lecture
and retrieval from memory (McDaniel & Einstein, 1989, afforded item-specific processing, we therefore expected
2005). Achieving material-appropriate difficulty means material-appropriate difficulty to be achieved when using
that both item-specific and relational processing will outline notes.
occur, thus supporting performance on a range of memory We also manipulated the level of difficulty of the note-
tests—including tests that assess item-specific details (i.e., taking task. As students listened to the lecture, they were
facts directly stated in the material) as well as tests that required to fill in different amounts of missing content in
require integration and synthesis of ideas. the cloze or outline notes. More-difficult conditions
required students to complete more of the notes them-
selves, creating more cognitive load compared to less-
Current experiment
difficult conditions. Students were randomly assigned to a
The current experiment extended the research on material- note-taking condition in a 2 (note-taking format: Outline,
appropriate difficulty to address how different note-taking Cloze) × 2 (level of difficulty: More-Difficult, Less-
formats and difficulty levels impact learning from a brief Difficult) design.
audio-recorded lecture. We reasoned that note-taking serves Memory was assessed using free recall, verbatim short
as an encoding task that can induce qualitatively different answer questions, and inference short answer questions.
types of semantic processing. We collected several process- Free recall relies on both item-specific and relational pro-
level measures (i.e., online cognitive load, metacognitive cessing, and accuracy is enhanced when both types of pro-
ratings of the note-taking experience, note completeness) to cessing occur during encoding (e.g., Einstein & Hunt,
further explain the mechanisms by which note-taking for- 1980; Einstein et al., 1984; Hunt & Einstein, 1981).
mat and difficulty impact learning. Although individual dif- Verbatim and inference short answer questions require
ferences (e.g., cognitive ability) can also impact the amount item-specific and relational processing, respectively (Blunt
of cognitive load experienced during learning, it is difficult & Karpicke, 2014; Karpicke & Blunt, 2011). When using
for instructors to tailor classroom materials to account for cloze notes, students should be engaged in only item-
these differences. We controlled for individual differences specific processing (from both the descriptive lecture and
in prior knowledge and working memory, to examine the encoding task). In contrast, when using outline notes, stu-
interaction between note-taking format and level of diffi- dents should be engaged in both item-specific (due to the
culty for students on average, beyond these individual dif- descriptive lecture) and relational processing (due to the
ference factors. encoding task). Thus, we expected that cloze and outline
Students listened to a descriptive lecture on a science notes would lead to equal performance on verbatim short
topic (i.e., the components and functions of human answer questions, which require item-specific processing
blood). Descriptive prose (e.g., expository text or lecture) (Blunt & Karpicke, 2014; Karpicke & Blunt, 2011).
typically presents a series of independent facts with less However, we expected outline notes to increase memory
emphasis on the underlying structure or connections performance on free recall and inference questions relative
among the ideas. Therefore, when learning from descrip- to cloze notes, due to the need for relational processing on
tive prose, students tend to engage in item-specific pro- these items.
cessing (i.e., treat the information as a list of independent Compared to the less-difficult notes, the more-difficult
facts; Mayer, 1985, 1987) and are less likely to process notes require students to generate more information,
the relational aspects of the information (i.e., the organi- increasing their active involvement in comprehending the
sational structure; Cook & Mayer, 1988; McDaniel et al., lecture while also imposing greater overall cognitive load.
1986). We used the same content from previous text- Drawing from previous research that tested the material-
based studies which found that these descriptive materi- appropriate difficulty framework, there is mixed evidence
als led to item-specific processing (Blunt & Karpicke, regarding the benefits of increasing the difficulty of the
2014; Karpicke & Blunt, 2011). encoding task. In conditions that were set up to achieve
While listening to the lecture, students completed either material appropriate difficulty, increasing the difficulty
cloze or outline notes. Cloze notes included an exact tran- of the encoding task benefitted learning of word lists
scription of the lecture, but with key words missing. This (McDaniel et al., 1988) but hindered memory perfor-
task is similar to tasks used in previous studies shown to mance when learning from a text passage (Einstein et al.,
induce item-specific processing (e.g., letter-insertion task; 1990). One interesting note about these studies is that the
Einstein, McDaniel, Bowers, & Stevens, 1984; McDaniel learning material itself was manipulated by presenting it
2810 Quarterly Journal of Experimental Psychology 72(12)

in a random order or withholding some information from the components or functions of human blood on the prior
the student. In our experiment, however, the lecture was knowledge question (n = 47). Third, students were excluded
intact and only the learning aid was manipulated to induce for reporting that English was not their first language, and
different types of semantic processing. We assumed that thus may not fluently understand the verbal or written
this additive approach (i.e., note-taking is supplementary materials (n = 5).
to the lecture) could accommodate additional difficulty
and thus we predicted that increasing difficulty will ben-
Experimental design
efit learning—but only in the outline notes condition.
In the More-Difficult Outline condition, students Students were randomly assigned to a note-taking condi-
engage in both item-specific and relational processing, tion in a 2 (note format: Cloze, Outline) × 2 (level of
while also generating more information for themselves. In difficulty: Less-Difficult, More-Difficult) between-sub-
contrast, by decreasing difficulty, Less-Difficult Outline jects factorial design (Less-Difficult Cloze n = 28, More-
notes may decrease active processing, reducing learning Difficult Cloze n = 33, Less-Difficult Outline n = 32,
compared to the More-Difficult Outline condition (e.g., More-Difficult Outline n = 30).
Morgan et al., 1988). Importantly, both the degree and type
of processing should matter—despite increasing encoding Materials
difficulty, the More-Difficult Cloze notes should lead to
only item-specific processing. Thus, both Cloze notes con- Prior knowledge assessment. Students completed a brief
ditions should lead to lower learning than the More- cued-recall question assessing their prior knowledge. Spe-
Difficult Outline notes condition. This prediction is cifically, students were asked to list each component of
consistent with findings when learning from word lists as human blood and its corresponding function. Students
well as text passages (McDaniel & Butler, 2011; McDaniel were given the option to write “I don’t know.” Below the
& Einstein, 1989, 2005; McDaniel et al., 1988). prompt were eight lines, labelled, “Component 1, Function
By examining the combined impact of note-taking for- 1, etc.” for each of four component/function pairs. One
mat and level of difficulty, this study may reveal important point was given for each correctly listed component (i.e.,
factors moderating the benefits of instructor-provided plasma, red blood cells/erythrocyte, white blood cells/leu-
notes. In addition, this study applies the material-appropri- kocyte, platelets) and function (eight points total).
ate difficulty framework to a common learning situation
(i.e., lecture note-taking) by providing the first empirical Lecture.  Students listened to a 2-min science lecture about
test to examine the interactions between different types of the components and functions of human blood (adapted
semantic processing (i.e., induced by the encoding task from text-based studies), which affords item-specific pro-
and learning material) and levels of difficulty in an encod- cessing (Blunt & Karpicke, 2014; Karpicke & Blunt,
ing task that is not self-paced. 2011). The 245-word lecture was presented at an average
rate of 117 words per minute and included 33 individual
idea units (i.e., a small group of words that represent a sin-
Method gle idea or fact). Idea units were used to assess note-taking
and free recall performance.
Participants
Participants were undergraduate students enrolled in psy- Online cognitive load.  Students were asked to wear head-
chology courses (N = 123, M age = 20.24 years, SD = 3.36, phones and complete a reaction time measure of online
63.4% female). The majority of students identified them- cognitive load (Piolat et al., 2005) by pressing the space-
selves as White (81%), with the remaining individuals bar as quickly as possible, using their non-writing hand,
identifying themselves as Black (10%), Asian (5%), once they heard a tone. The 2-min task presented six audi-
Hispanic or Latino (1%), or other (4%). Additional stu- tory tones at predetermined random intervals ranging
dents were tested, but excluded from analyses, for three from 15 to 30 s. Students first completed this task as a
reasons. First, students were excluded for not following single-task baseline measure, then as a secondary task
experiment instructions: (a) completing less than 30% of during the lecture. The tones never overlapped with words
the notes handout, indicating that they were not suffi- in the lecture.
ciently exposed to the processing manipulation of the note- Reaction times were calculated by subtracting the onset
taking format (n = 2), (b) committing 20% or more errors time for each tone from the time at which the space bar
on the automated reading span task (n = 4; Conway et al., was pressed. If no response time to a tone was recorded,
2005), or (c) missing more than two responses across both then the missing response time was replaced with the max-
reaction time tasks (n = 7). Second, students were excluded imum time allotted to respond to the corresponding tone.
for having a high level of prior knowledge, and therefore Then, the onset time for the tone was subtracted from the
less to learn from the lecture, by producing at least 50% of replaced response time to calculate the reaction time.
Bellinger and DeCaro 2811

Median interference in reaction time (IRT) scores were The post-experiment questionnaire requested demo-
calculated by (a) subtracting the baseline reaction time graphic information and students’ note-taking preferences
from the secondary-task reaction time and then (b) calcu- (i.e., a 9-option multiple choice question asking which
lating the median value of the six reaction time differences. approach to note-taking they use most often during lec-
A positive median IRT indicates an increase in reaction tures in science courses) and experience with instructor-
time (i.e., slower response) during the secondary task rela- provided notes during their post-secondary education (i.e.,
tive to the baseline task and may be interpreted as an two 7-option multiple choice questions asking for how
increase in cognitive load induced by lecture note-taking many courses their college instructors have provided them
(Piolat et al., 2005). with partial and complete lecture notes and one 3-option
multiple choice question asking when they could first
Note-taking.  Figure 1 illustrates each type of note-taking access the instructor-provided notes).
handout (cloze, outline) for each level of difficulty (less,
more). The handouts were designed to simulate two types Memory tests. To assess learning from the lecture, two
of instructor-provided partial notes. Cloze notes handouts types of memory tests were administered on a computer.
provide a transcription of the lecture but with key word(s) The free recall test (Karpicke & Blunt, 2011) required stu-
deleted from each idea unit. Students were asked to fill in dents to document everything they could remember from
the missing words as they listened to the lecture (e.g., the lecture, within a 7 min time limit. The short answer test
Einstein et al., 1984; McDaniel et al., 1986). Outline (adapted from Blunt & Karpicke, 2014; Karpicke & Blunt,
notes handouts explicitly organised the prose into a hier- 2011) included 10 verbatim and 4 inference questions. The
archy of main and supporting ideas. Students were asked verbatim questions (Cronbach’s α = .52) asked for specific
to fill in the missing information denoted by blanks facts (typically a single idea unit) stated directly in the lec-
(Einstein et al., 1990). ture and thus assessed recall of item-specific information
The difficulty of the note-taking task was manipulated (Blunt & Karpicke, 2014). For example, the question
by requiring students to fill in different amounts of missing “What percentage of plasma is water?” corresponded to
content. In the Less-Difficult Cloze and Less-Difficult the idea unit “Plasma is about 90% water.” In contrast, the
Outline handouts, approximately half (i.e., 17) of the idea inference questions (Cronbach’s α = .48) assessed recall of
units were missing information (the same 17 words were relational information, because they required students to
missing in both handouts). In the More-Difficult Cloze and connect multiple idea units from the lecture to reason
More-Difficult Outline handouts, all 33 idea units were beyond the information provided (Blunt & Karpicke,
incomplete. 2014). For example, the question “What would happen to
the blood flow from a wound if the body had no fibrin?”
Working memory capacity. The Automated Reading Span referred to the following idea units: (a) “The fibrin forms a
Task (Redick et al., 2012) served as a distractor task meshwork of microscopic fibres”; (b) “These fibres trap
between the lecture and memory tests. Scores were also blood cells”; (c) “and create a clot”; (d) “The clot closes
used as a covariate to estimate effects of condition inde- off the cut or wound”; (e) “so that bleeding stops.”
pendent of individual differences in working memory Students were required to spend a minimum of 15 s
capacity. Students were asked to judge whether a sentence answering each short answer question (adapted from
made sense and then presented with a letter for recall at the Karpicke & Blunt, 2011). No maximum time was given.
end of the set. After three to seven sentence-letter combi- Two raters scored 20% of all memory tests and notes, dem-
nations, students were presented with a recall grid and onstrating high consistency between raters (Cohen’s kap-
asked to select the letters they saw during the trial in the pas = .91 to 1.00). Remaining memory tests and notes were
correct order. Students viewed three administrations for scored by only one rater.
each set size (i.e., 75 total sentence-storage pairs). The
total score was calculated by summing the total number of
correct responses out of 75 (Conway et al., 2005).
Procedure
Students completed the study individually in separate
Questionnaires. The post-lecture questionnaire included rooms. After providing informed consent, students were
four questions assessing students’ metacognition. Using a asked to complete the prior knowledge assessment. Then,
7-point Likert-type scale, we assessed students’ perceptions students were informed that they would listen to a lecture
of how (a) difficult (1 = very easy, 7 = very difficult; adapted about the human body, and that they would not be able to
from DeLeeuw & Mayer, 2008), (b) helpful (1 = not at all rewind, fast-forward, or pause the recording. They were
helpful, 7 = very helpful), and (c) enjoyable (1 = not at all instructed to take notes using the handout provided to them
enjoyable, 7 = very enjoyable) the note-taking task was, in order to help them learn the information for a memory
and (d) their perceived comprehension of the lecture test at the end of the experiment. Students were then asked
(1 = did not comprehend it very well, 7 = comprehended it to wear headphones and complete the baseline reaction
very well; adapted from Einstein et al., 1990). time task.
2812 Quarterly Journal of Experimental Psychology 72(12)

Figure 1.  Examples of note-taking handouts: (a) More Difficult Outline notes, (b) More Difficult Cloze notes, (c) Less Difficult
Outline notes, and (d) Less Difficult Cloze notes.
Bellinger and DeCaro 2813

Table 1.  Students’ self-reported experiences with instructor- partially complete copy of the lecture information (e.g.,
provided notes. key terms or definitions deleted). Thus, in this sample, stu-
Complete notes Partial notes dents were less likely to have had experience with partially
completed notes. Students reported that it is more common
0 courses 10.6% 47.2% for them to first access instructor-provided notes before
1–2 courses 22.8% 33.3% lecture (n = 113; 463 courses indicated) relative to after
3–4 courses 31.7% 13.0% lecture (n = 63; 207 courses indicated).
5–6 courses 9.8% 4.1%
7–8 courses 7.3% 1.6%
9–10 courses 3.3% − Preliminary analyses
11 + courses 14.6% 0.8%
Separate 2 (note format: cloze, outline) × 2 (level of dif-
ficulty: less-difficult, more-difficult) between-subjects
factorial analyses of variance (ANOVAs) revealed that
Immediately before the lecture began, students received
neither working memory capacity nor prior knowledge
an example of the type of handout they would use to take
differed based on note format, working memory capac-
notes, to familiarise them with the format. Students were
ity, F < 1; prior knowledge, F(1, 119) = 1.61, p = .207,
shown both a blank and completed example handout, on a
ηp2 = .01; or level of difficulty, working memory capac-
different topic (i.e., the human ear). Students were directed
ity, F(1, 119) = 2.99, p = .086, η p2 = .03, prior knowl-
to focus on the format of the notes rather than the content,
edge, F(1, 119) = 2.41, p = .123, ηp2 = .02. No interactions
and were provided with 60 s to review the example hand-
were found, Fs < 1, indicating that the four note-taking
out before they were given the handout to be used during
conditions were statistically equivalent in working
the lecture.
memory capacity and prior knowledge.
Students then performed the reaction time task during
Subsequent analyses controlled for working memory
the lecture while taking notes, as a measure of online cog-
capacity (M = 57.12, SD = 9.42) and prior knowledge
nitive load. After the lecture, the experimenter collected
(M = 0.12, SD = 0.15), to demonstrate condition effects
the notes, and students completed the post-lecture ques-
while accounting for these individual differences.1 Table 2
tionnaire, followed by the working memory capacity task.
presents the main effects of working memory capacity and
Next, students completed the post-experiment question-
prior knowledge for each model reported below.
naire, followed by the memory tests for the lecture. Finally,
students were debriefed. The experiment lasted approxi-
mately 50 min. Primary analyses
Separate 2 (note format: cloze, outline) × 2 (level of diffi-
Results culty: less-difficult, more-difficult) between-subjects facto-
rial ANCOVAs were used to analyse memory performance,
Note-taking habits and experiences cognitive load, note completeness, and metacognitive rat-
In order to gauge experience with note-taking, we first ings for the lecture. The descriptive nature of the lecture is
examined students’ self-reported note-taking in science thought to encourage item-specific processing (Blunt &
courses. Nearly all students (97.6%) reported taking Karpicke, 2014; Karpicke & Blunt, 2011). Thus, we hypoth-
notes during lectures. The majority reported creating out- esised that taking outline notes, which is thought to encour-
lines (26.0%) or writing a list of bullet points (39.8%). age relational processing (Einstein et al., 1990), would
The remaining students reported that they try to write facilitate material-appropriate processing and benefit mem-
down everything the instructor says (21.1%), use a copy ory performance on free recall and short answer inference
of the instructor’s PowerPoint slides to guide their note- questions relative to cloze notes. Moreover, based on the
taking (4.9%), or responded that they use a method other material-appropriate difficulty account (McDaniel & Butler,
than any of the above (5.7%). These findings suggest that 2011), we predicted an interaction, with the highest perfor-
students are generally familiar with outlining and value mance on these items in the More-Difficult Outline condi-
having notes that are as complete as possible. Additionally, tion. We hypothesised that there would be no difference
no students reported using the Cornell note-taking between conditions on short answer verbatim questions.
method (e.g., Quintus, Borr, Duffield, Napoleon, &
Welch, 2012), graphic organisers (e.g., Ponce & Mayer,
Memory tests
2014), or matrix notes (e.g., Kiewra, Benton, Kim, Risch,
& Christensen, 1995). Free recall.  Level of difficulty did not significantly impact
As shown in Table 1, students indicated that professors free recall, F(1, 117) = 1.65, p = .201, ηp2 = .01, d = .11. As
were more likely to provide a complete copy of the lec- predicted, a main effect of note format indicated that
ture information (e.g., PowerPoint slides) compared to a outline notes (M = 0.25, SE = 0.01) led to superior free
2814 Quarterly Journal of Experimental Psychology 72(12)

Table 2.  Main effects of working memory capacity and prior knowledge for each analysis.

Working memory capacity Prior knowledge

  F ηp2 F ηp2
Free recall 2.01 .02 16.35*** .12
Verbatim (short answer) 1.99 .02 11.07** .09
Inference (short answer) 0.07 .00 10.00** .08
Online cognitive load 0.13 .00 3.02 .03
Note completeness 1.45 .01 3.77 .03
Difficult 4.27* .04 0.39 .00
Enjoyable 0.51 .00 0.17 .00
Comprehend 0.06 .00 8.11** .07
Helpful 3.42 .03 0.44 .00

*p < .05; **p < .01; ***p < .001.

Figure 2.  Mean proportion correct on free recall and short answer tests (verbatim, inference) as a function of note format and
level of difficulty. Error bars represent ± 1 standard error of the mean.

recall compared to cloze notes (M = 0.22, SE = 0.01), On inference questions, there was no main effect of level
F(1, 117) = 4.77, p = .031, ηp2 = .04, d = .32. This effect was of difficulty, F < 1, but there was a main effect of note for-
qualified by a significant interaction, F(1, 117) = 19.95, mat, F(1, 117) = 5.14, p = .025, ηp2 = .04, d = .35. Outline
p < .001, ηp2 = .15. As shown in Figure 2 and Table 3, notes (M = 0.80, SE = 0.03) led to more accurate inferences
students who completed More-Difficult Outline notes compared to cloze notes (M = 0.70, SE = 0.03). This effect
produced more total idea units during free recall com- was qualified by a significant interaction, F(1, 117) = 12.31,
pared to students who completed More-Difficult Cloze, p = .001, ηp2 = .10. As shown in Figure 2 and Table 3, com-
F(1, 59) = 20.44, p < .001, ηp2 = .26, d = 1.10; Less-Difficult pleting More-Difficult Outline notes led to more accurate
Cloze, F(1, 54) = 6.06, p = .017, ηp2 = .10, d = .50; and Less- inferences compared to all of the other guided notes, More-
Difficult Outline notes, F(1, 58) = 17.68, p < .001, ηp2 = .23, Difficult Cloze, F(1, 59) =  18.85, p < .001, ηp2 = .24,
d = .94. These findings suggest that completing more- d = 1.10; Less-Difficult Cloze, F(1, 54) = 5.03, p = .029,
difficult outline notes promotes better learning than the ηp2 = .09, d = .45; Less-Difficult Outline, F(1, 58) = 6.92,
other note-taking conditions. p = .011, ηp2 = .11, d = .62.

Short answer questions.  As predicted, on verbatim questions,

Online cognitive load
there were no main effects of note format, F(1, 117) = 1.59,
p = .210, ηp2 = .01, d = .21; level of difficulty, F(1, 117) = 1.52, Completing outline notes (M = 240.19, SE = 16.45) increased
p = .220, ηp2 = .01, d = .32; or an interaction, F < 1 (Figure 2). secondary task reaction time compared to cloze notes
 Bellinger and DeCaro 2815

(M = 126.71, SE = 16.63), F(1, 117) = 23.31, p < .001,

212.09 (140.36)
Table 3.  Mean memory test scores (proportion correct), metacognitive ratings (out of 7), note completeness (proportion correct), and online cognitive load (median interference

3.44 (1.45)

0.75 (0.23)
0.75 (0.24)

3.56 (1.67)
3.92 (1.45)
4.37 (1.46)
0.24 (0.10)
0.47 (0.18)
More difficult
ηp2 = .17, d = .87, indicating greater cognitive load in the
outline notes conditions. In addition, notes with more

(n = 63)
difficulty (M = 212.53, SE = 16.44) increased reaction time
relative to notes with less difficulty (M = 154.37, SE = 16.88),
F(1, 117) = 5.96, p = .016, ηp2 = .05, d = .39, indicating
Level of difficulty

156.37 (143.09)
increased cognitive load in the More-Difficult conditions.

0.90 (0.13)
2.70 (1.55)
4.77 (1.36)
5.05 (1.51)
4.35 (1.45)
0.23 (0.09)
0.53 (0.19)
0.74 (0.28)
No interaction was found, F < 1.
Less difficult
(n = 60)

Note completeness
Note completeness was measured as the proportion of cor-

241.96 (141.08)
rect words written in the correct blanks. Cloze notes
Outline notes

4.00 (1.37)

0.67 (0.19)
4.13 (1.41)
4.19 (1.60)
4.50 (1.49)
0.25 (0.09)
0.52 (0.19)
0.79 (0.27)

(M = 0.97, SE = 0.01) were more complete than outline notes

(M = 0.67, SE = 0.01), F(1, 117) = 355.53, p < .001, ηp2 = .75,
(n = 62)

d = 2.18. Less-difficult notes (M = 0.90, SE = 0.01) were

more complete than more-difficult notes (M = 0.74,
SE = 0.01), F(1, 117) = 97.95, p < .001, ηp2 = .46, d = .80.
126.92 (122.74)

These effects were qualified by an interaction, F(1,

0.97 (0.04)
4.90 (1.54)
3.77 (1.65)
2.13 (1.26)
4.48 (1.31)
0.22 (0.10)
0.48 (0.19)
0.70 (0.25)
Note format

117) = 58.93, p < .001, ηp2 = .34. Even though More-Difficult

Cloze notes

Outline notes led to higher memory performance, these

(n = 61)
in reaction time—ms) as a function of note format and level of difficulty. Standard deviations are in parentheses.

notes (M = 0.53, SE = 0.02) were less complete than in each

other condition, More-Difficult Cloze (M = 0.95, SE = 0.02),
F(1, 59) = 369.43, p < .001, ηp2 = .86, d = 5.02; Less-Difficult
269.82 (134.76)

Cloze (M = 0.99, SE = 0.02), F(1, 54)= 397.71, p < .001,

3.63 (1.43)

0.52 (0.12)
0.87 (0.17)

4.73 (1.26)
3.50 (1.41)
4.13 (1.43)
0.29 (0.08)
0.49 (0.19)
More difficult

ηp2 = .88, d  = 5.44; Less-Difficult Outline (M = 0.81,

SE = 0.02), F(1, 58) = 81.75, p < .001, ηp2 = .59, d = 2.32.
(n = 30)

Overall, there was a negative partial correlation between

note completeness and free recall, rp(119) = -.21, p = .019,
Values in the table are raw scores, and do not account for covariates used in the primary analyses.

and inference memory scores, rp(119) = -.19, p = .038. No

215.84 (143.94)

relationship was found between note completeness and ver-

Outline notes

0.81 (0.13)
0.71 (0.32)

3.56 (1.32)
4.84 (1.51)
4.84 (1.48)
4.34 (1.23)
0.21 (0.09)
0.54 (0.19)
Less difficult

batim scores, rp(119) = .10, p = .272. These findings mirror

the overall results, in that outline notes were less complete
(n = 32)

and also led to higher free recall and inference scores.

Metacognitive ratings
159.61 (125.49)
0.96 (0.04)
4.58 (1.48)
3.27 (1.46)
0.64 (0.24)

2.49 (1.23)
4.30 (1.40)
0.19 (0.10)
0.45 (0.18)
More difficult

Difficult.  Outline notes (M = 4.14, SE = 0.16) were rated

as more difficult than cloze notes (M = 2.11, SE = 0.16),
(n = 33)

F(1, 117) = 81.00, p < .001, ηp2 = .41, d = 1.50. In addition,

notes taken in the More-Difficult conditions (M = 3.65,
SE = 0.16) were rated as more difficult than those taken
88.39 (109.41)

in the Less-Difficult conditions (M = 2.59, SE = 0.16),

4.36 (1.68)

1.00 (0.02)
0.78 (0.23)

1.71 (1.18)
4.68 (1.19)
5.29 (1.54)
0.25 (0.08)
0.52 (0.19)
Less difficult
Cloze notes

F(1, 117) = 21.67, p < .001, ηp2 = .16, d = .53. No interaction

was found (F < 1; Figure 3; Table 3).
(n = 28)

Enjoyable.  Students enjoyed using the Less-Difficult notes

(M = 4.74, SE  = 0.18) compared to the More-Difficult
Metacognitive ratings

  Online cognitive load

notes (M = 3.92, SE = 0.18), F(1, 117) = 10.02, p = .002,

  Note completeness
Process measures

ηp2 = .08, d = .60. There was no main effect of note format,

F(1, 117) = 1.59, p = .210, ηp2 = .01, d = .20, or interaction,
Memory tests

 Comprehend
  Free recall

F(1, 117) = 3.75, p = .055, ηp2 = .03 (Figure 3; Table 3).

 Enjoyable
 Inference
 Verbatim

 Difficult

 Helpful

Comprehend. Students who completed Less-Difficult

notes (M = 5.01, SE = 0.19) thought they comprehended the
 
2816 Quarterly Journal of Experimental Psychology 72(12)

Figure 3.  Mean self-report ratings of metacognitive factors as a function of note format and level of difficulty. Error bars
represent ± 1 standard error of the mean.

lecture better than students who completed More-Difficult This study also included several process measures, add-
notes (M = 4.41, SE = 0.18), F(1, 117) = 4.95, p = .028, ing metacognitive and objective cognitive load measures
ηp2 = .04, d = .46. There was no main effect of note format, to the material-appropriate difficulty literature. Despite
F(1, 117) = 1.87, p = .174, ηp2 = .02, d = .26, or interaction, overall better memory, outline notes were rated as more
F < 1 (Figure 3; Table 3). difficult and led to greater online cognitive load. In addi-
tion, notes in the more-difficult conditions were rated as
Helpful. Students perceived using Less-Difficult notes more difficult, less enjoyable, less helpful, and as leading
(M = 4.30, SE = 0.19) as more helpful than More-Difficult to lower comprehension than notes in the less-difficult
notes (M = 3.50, SE = 0.18), F(1, 117) = 8.92, p = .003, conditions. More-Difficult Outline notes were also the
ηp2 = .07, d = .63. There was no main effect of note format least complete, but they improved free recall and inference
or interaction, Fs < 1 (Figure 3; Table 3). accuracy compared to the other three note-taking formats.
These results mirror other findings demonstrating that
what students find easier, more helpful, and more enjoya-
Discussion ble do not always match what most benefits their learning
The current experiment addressed a theoretically driven and (Rohrer & Pashler, 2010).
practical question about learning and memory: How do note- These results suggest that outline notes and greater dif-
taking format and level of difficulty impact learning from ficulty posed the greatest challenge for students. Yet, these
instructor-provided lecture notes? Our findings provide addi- conditions increased “desirable difficulty,” or generative
tional evidence in support of the material-appropriate diffi- processes that support learning (Bjork & Bjork, 2011).
culty framework (McDaniel & Butler, 2011; McDaniel & Even though material-appropriate difficulty was intended
Einstein, 1989, 2005; McDaniel et al., 1988). Increasing to be achieved in both More-Difficult Outline and Less-
note-taking difficulty (by including more blanks) led to bet- Difficult Outline conditions, increasing the difficulty of
ter memory outcomes (free recall and inference)—but only outline notes required students to be more actively involved
in the outline notes condition. When giving a descriptive, in making sense of the information, which led to better
fact-based science lecture, outline notes enhanced the ability learning. Moreover, difficulty alone was not sufficient to
to infer relations among the ideas presented. Verbatim short increase learning—learning was not improved in the
answer accuracy, which assesses memory for specific facts, More-Difficult Cloze notes condition. Material-appropriate
was equivalent between outline and cloze note formats. processing was required in order for difficulty to be desir-
Thus, by engaging in both item-specific and relational pro- able (McDaniel & Butler, 2011).
cessing, students given outline notes were able to retain both By showing better learning in the More-Difficult
types of information. In contrast, cloze notes appear to have Outline notes condition, these findings do not align with
induced only item-specific processing, supporting memory cognitive load theory (Paas, 1992; Sweller et al., 1998).
for specific facts alone. Providing instructor notes is often intended to reduce
Bellinger and DeCaro 2817

students’ mental effort during lecture, to increase their processing entirely. More research is needed to test these
ability to process the information more deeply and ask possibilities.
questions (Austin et al., 2002). Yet, we found that increas-
ing difficulty was more beneficial, and modifying diffi-
Conclusion
culty only mattered with one of the two note-taking
activities. It remains likely that further increasing the dif- The current experiment provides strong evidence that both
ficulty of outline notes would increase cognitive load to note format and level of difficulty are important variables
such an extent that material-appropriate processing would that impact the efficacy of guided notes. These findings
no longer have a benefit (Einstein et al., 1990). It is also also replicate and extend the material-appropriate diffi-
possible that increasing the difficulty of cloze notes would culty framework from self-paced (text-based) to instruc-
lead to even lower learning. tor-paced (audio-based) learning materials. These findings
support a recommendation for instructors and students:
when learning from descriptive lectures, students should
Limitations take notes using outlines that require a greater degree of
The conditions created for the current experiments were generative processing. More generally, instructor-provided
not intended to fully replicate the conditions found in the notes should require generative processing, even if stu-
classroom, and some limitations must be noted. For exam- dents experience difficulty. This point is particularly rele-
ple, the lecture was brief, informationally dense, and vant considering our samples’ reports that instructors were
delivered at a relatively quick pace, which may not be rep- much more likely to provide complete than partial notes.
resentative of many lectures. Furthermore, students were Furthermore, only a third of students stated that they elect
tested within an hour of listening to the lecture, so these to create outlines during science lectures, suggesting that
results cannot speak to whether the effects extend over many students do not choose the optimal note format. By
longer retention intervals. This study specifically exam- adopting a “less is more” approach to instructor-provided
ined the encoding benefit of note-taking, and therefore outline notes, students will be encouraged to assume a
any impact of reviewing the notes is unknown. more optimal method of processing descriptive lecture
In addition, we used a descriptive lecture, and thus our information. Given the ubiquitous use of lectures and the
results cannot speak to other lecture types (i.e., a narra- importance of note-taking for capitalising on this learning
tive lecture). Indeed, the nature of a lecture (i.e., descrip- opportunity, the development of outline notes for descrip-
tive versus narrative) is not always likely to be tive lectures is a promising educational intervention.
straightforward. The same lecture could be delivered in
multiple ways—as a story that emphasises the relational Acknowledgements
aspects of the information, or without making the con- The authors thank Cara H. Cashon, Keith B. Lyle, John R. Pani,
nections between ideas explicit and thus emphasising the and Kate E. Snyder for helpful feedback on this research, and
item-specific aspects of the information. Furthermore, Mark McDaniel for feedback on the manuscript.
the same lecture could switch between emphasising the
two types of semantic processing. Cook and Mayer Declaration of conflicting interests
(1988) note that, without extensive tailored training, stu- The author(s) declared no potential conflicts of interest with
dents often have difficulty identifying organisational respect to the research, authorship, and/or publication of this
structures (i.e., relational information) of descriptive article.
prose, even though the prose contains both item-specific
and relational information. Funding
Although we suggest that increasing difficulty in the The author(s) disclosed receipt of the following financial support
More-Difficult Outline notes condition increased genera- for the research, authorship, and/or publication of this article:
tive processing, an alternative possibility is that the Less- This article is based on a PhD dissertation submitted to the
Difficult Outline notes no longer induced relational University of Louisville. The research was supported by a
processing. Specifically, students may not have processed Dissertation Research Award from the Graduate Network in Arts
the relational information because they did not organise the and Sciences at the University of Louisville.
lecture content on their own. Instead, they simply filled in
missing words. Thus, Less-Difficult Outline notes may ORCID iD
have induced processing that is more similar to the cloze Marci S DeCaro https://orcid.org/0000-0001-6753-0725
notes (i.e., item-specific processing) than to the More-
Difficult Outline notes. Again, this possibility demonstrates Note
the importance of considering the level of difficulty in 1. The study was not designed to examine interactions with
encoding tasks. Reducing difficulty may either decrease these individual differences, thus the sample size was insuf-
generative processing or alter the nature of semantic ficient to test additional interactions.
2818 Quarterly Journal of Experimental Psychology 72(12)

References Einstein, G. O., McDaniel, M. A., Owen, P. D., & Cote, N. C.

(1990). Encoding and recall of texts: The importance of
Armbruster, B. B. (2009). Taking notes from lectures. In R. F.
material appropriate processing. Journal of Memory and
Flippo & D. C. Caverly (Eds.), Handbook of college read-
Language, 29, 566–581.
ing and study strategy research (2nd ed., pp. 220–248).
Engle, R. W., Tuholski, S. W., Laughlin, J. E., & Conway, A. R.
New York, NY: Routledge.
A. (1999). Working memory, short-term memory, and gen-
Austin, J. L., Lee, M. G., Thibeault, M. D., Carr, J. E., & Bailey,
eral fluid intelligence: A latent-variable approach. Journal
J. S. (2002). Effects of guided notes on university students’
of Experimental Psychology: General, 128, 309–331.
responding and recall of information. Journal of Behavioral
Fiorella, L., & Mayer, R. E. (2016). Eight ways to promote
Education, 11, 243–254.
generative learning. Educational Psychology Review, 28,
Bjork, E. L., & Bjork, R. A. (2011). Making things hard on your-
717–741.
self, but in a good way: Creating desirable difficulties to
Hunt, R. R. (2003). Two contributions of distinctive processing
enhance learning. In M. A. Gernsbacher, R. W. Pew, L. M.
to accurate memory. Journal of Memory and Language, 48,
Hough & J. R. Pomerantz (Eds.), Psychology and the real
811–825.
world: Essays illustrating fundamental contributions to
Hunt, R. R. (2013). Precision in memory through distinctive pro-
society (pp. 56–64). New York, NY: Worth Publishers.
cessing. Current Directions in Psychological Science, 22,
Blunt, J. R., & Karpicke, J. D. (2014). Learning with retrieval-
10–15.
based concept mapping. Journal of Educational Psychology,
Hunt, R. R., & Einstein, G. O. (1981). Relational and item-spe-
106, 849–858.
cific information in memory. Journal of Verbal Learning
Boyle, J. R. (2012). Note-taking and secondary students with
and Verbal Behavior, 20, 497–514.
learning disabilities: Challenges and solutions. Learning
Jenkins, J. J. (1979). Four points to remember: A tetrahedral
Disabilities Research & Practice, 27, 90–101.
model of memory experiments. In L. S. Cermak & F. I. M.
Bui, D. C., & Myerson, J. (2014). The role of working memory Craik (Eds.), Levels of processing in human memory (pp.
abilities in lecture note-taking. Learning and Individual 429–446). Hillsdale, NJ: Lawrence Erlbaum.
Differences, 33, 12–22. Karpicke, J. D., & Blunt, J. R. (2011). Retrieval practice pro-
Bui, D. C., Myerson, J., & Hale, S. (2013). Note-taking with duces more learning than elaborative studying with concept
computers: Exploring alternative strategies for improved mapping. Science, 331, 772–775.
recall. Journal of Educational Psychology, 105, 299–309. Karpicke, J. D., & Roediger, H. L., III. (2010). Is expanding
Chen, X. (2002). Teaching undergraduates in U.S. postsecond- retrieval a superior method for learning text materials.
ary institutions: Fall 1998. Washington, DC: Government Memory & Cognition, 38, 116–124.
Printing Office (NCES Publication No. 2002-209). Katayama, A. D., & Robinson, D. H. (2000). Getting students
Retrieved from http://nces.ed.gov/pubs2002/2002209.pdf “partially” involved in note-taking using graphic organizers.
Conway, A. R. A., Kane, M. J., Bunting, M. F., Hambrick, D. The Journal of Experimental Education, 68, 119–133.
Z., Wilhelm, O., & Engle, R. W. (2005). Working mem- Kiewra, K. A. (1985a). Investigating notetaking and review: A
ory span tasks: A methodological review and user’s guide. depth of processing alternative. Educational Psychologist,
Psychonomic Bulletin & Review, 12, 769–786. 20, 23–32.
Cook, L. K., & Mayer, R. E. (1988). Teaching readers about Kiewra, K. A. (1985b). Providing the instructor’s notes: An
the structure of scientific text. Journal of Educational effective addition to student notetaking. Educational
Psychology, 80, 448–456. Psychologist, 20, 33–39.
Cornelius, T. L., & Owen-DeSchryver, J. (2008). Differential Kiewra, K. A., Benton, S. L., Kim, S., Risch, N., & Christensen,
effects of full and partial notes on learning outcomes and M. (1995). Effects of note-taking format and study tech-
attendance. Teaching of Psychology, 35, 6–12. nique on recall and relational performance. Contemporary
DeLeeuw, K. E., & Mayer, R. E. (2008). A comparison of three Educational Psychology, 20, 172–187.
measures of cognitive load: Evidence for separable meas- Kiewra, K. A., Dubois, N. F., Christensen, M., Kim, S., &
ures of intrinsic, extraneous, and germane load. Journal of Lindberg, N. (1989). A more equitable account of the note-
Educational Psychology, 100, 223–234. taking functions in learning from lecture and from text.
DiVesta, F. J., & Gray, G. S. (1972). Listening and note taking. Instructional Science, 18, 217–232.
Journal of Educational Psychology, 63, 8–14. Kobayashi, K. (2005). What limits the encoding effect of note-
Dunkel, P., & Davy, S. (1989). The heuristic of lecture note- taking? A meta-analytic examination. Contemporary
taking: Perceptions of American & international students Educational Psychology, 30, 242–262.
regarding the value & practice of notetaking. English for Mayer, R. E. (1985). Structural analysis of science prose: Can
Specific Purposes, 8, 33–50. we increase problem-solving performance?. In B. K. Britton
Einstein, G. O., & Hunt, R. R. (1980). Levels of processing and & J. B. Black (Eds.), Understanding expository prose (pp.
organization: Additive effects of individual-item and rela- 65–87). Hillsdale, NJ: Lawrence Erlbaum.
tional processing. Journal of Experimental Psychology: Mayer, R. E. (1987). Instructional variables that influence cogni-
Human Learning and Memory, 6, 588–598. tive processing during reading. In B. K. Britton & S. Glynn
Einstein, G. O., McDaniel, M. A., Bowers, C. A., & Stevens, D. (Eds.), Executive control processes in reading (pp. 201–216).
T. (1984). Memory for prose: The influence of relational and Hillsdale, NJ: Lawrence Erlbaum.
proposition-specific processing. Journal of Experimental McDaniel, M. A., & Butler, A. C. (2011). A contextual frame-
Psychology: Learning, Memory, and Cognition, 10, 133–143. work for understanding when difficulties are desirable. In
Bellinger and DeCaro 2819

A. S. Benjamin (Ed.), Successful remembering and success- Quintus, L., Borr, M., Duffield, S., Napoleon, L., & Welch, A.
ful forgetting: A festschrift in honor of Robert A. Bjork (2012). The impact of the Cornell note-taking method on
(pp. 175–198). New York, NY: Psychology Press. students’ performance in a high school family and consumer
McDaniel, M. A., & Einstein, G. O. (1989). Material appropriate sciences class. Journal of Family & Consumer Sciences
processing: A contextualistic approach to reading and study- Education, 30, 27–38.
ing strategies. Educational Psychology Review, 1, 113–145. Redick, T. S., Broadway, J. M., Meier, M. E., Kuriakose, P.
McDaniel, M. A., & Einstein, G. O. (2005). Material appropri- S., Unsworth, N., Kane, M. J., & Engle, R. W. (2012).
ate difficulty: A framework for determining when difficulty Measuring working memory capacity with automated
is desirable for improving learning. In A. F. Healy (Ed.), complex span tasks. European Journal of Psychological
Experimental cognitive psychology and its applications Assessment, 28, 164–171.
(pp. 73–85). Washington, DC: American Psychological Richland, L. E., Bjork, R. A., Finley, J. R., & Linn, M. C. (2005).
Association. Linking cognitive science to education: Generation and
McDaniel, M. A., Einstein, G. O., Dunay, P. K., & Cobb, R. E. interleaving effects. Proceedings of the Cognitive Science
(1986). Encoding difficulty and memory: Toward a unify- Society, 27, 1850–1855.
ing theory. Journal of Memory and Language, 25, 545–656. Rohrer, D., & Pashler, H. (2010). Recent research on human
McDaniel, M. A., Einstein, G. O., & Lollis, T. (1988). Qualitative learning challenges conventional instructional strategies.
and quantitative considerations in encoding difficulty Educational Researcher, 39, 406–412.
effects. Memory & Cognition, 16, 8–14. Svinicki, M., & McKeachie, W. J. (2011). How to make lectures
Mitchell, D. B., & Hunt, R. R. (1989). How much “effort” should more effective. In W. J. McKeachie (Ed.), McKeachie’s
be devoted to memory? Memory & Cognition, 17, 337–348. teaching tips: Strategies, research, and theory for college and
Morgan, C. H., Lilley, J. D., & Boreham, N. C. (1988). Learning university teachers (pp. 55–71). Belmont, CA: Wadsworth.
from lectures: The effect of varying the detail in lecture Sweller, J., van Merrienboer, J. J. G., & Paas, F. G. W. C. (1998).
handouts on note-taking and recall. Applied Cognitive Cognitive architecture and instructional design. Educational
Psychology, 2, 115–122. Psychology Review, 10, 251–296.
Neef, N. A., McCord, B. E., & Ferreri, S. J. (2006). Effects of Van Merrienboer, J. J. G., & Ayres, P. (2005). Research on cog-
guided notes versus completed notes during lectures on nitive load theory and its design implications for e-learning.
college students’ quiz performance. Journal of Applied Educational Technology Research and Development, 53,
Behavior Analysis, 39, 123–130. 5–13.
Paas, F. G. W. C. (1992). Training strategies for attaining trans- Van Meter, P., Yokoi, L., & Pressley, M. (1994). College students’
fer of problem-solving skill in statistics: A cognitive-load theory of note-taking derived from their perceptions of note-
approach. Journal of Educational Psychology, 84, 429–434. taking. Journal of Educational Psychology, 86, 323–338.
Piolat, A., Olive, T., & Kellogg, R. T. (2005). Cognitive effort Williams, R. L., & Eggert, A. C. (2002). Notetaking in college
during note taking. Applied Cognitive Psychology, 19, classes: Student patterns and instructional strategies. The
291–312. Journal of General Education, 51, 173–199.
Ponce, H. R., & Mayer, R. E. (2014). An eye movement analysis Williams, W. L., Weil, T. M., & Porter, J. C. K. (2012). The
of highlighting and graphic organizer study aids for learn- relative effects of traditional lectures and guided notes lec-
ing from expository text. Computers in Human Behavior, tures on university student test scores. The Behavior Analyst
41, 21–32. Today, 13, 12–16.