Introduction
Note taking has been recognized as the most common learning strategy both in first language (L1; Kobayashi, Reference Kobayashi2005) and second language (L2; Siegel, Reference Siegel2021) research. To better understand how note taking contributes to learning, researchers have identified two main functions of note taking: encoding and external storage functions (also known as process and product/review effects; Dunkel, Reference Dunkel1988). The encoding function occurs as notes are being taken. The physical practice of writing (or typing) notes involves processing input beyond verbatim copying such as perceiving and organizing information as well as relating it to existing knowledge (Rickards & Friedman, Reference Rickards and Friedman1978). The external storage function, which occurs after the note-taking process is completed, involves reviewing and/or reorganizing and storing the written input into memory to prevent forgetting as well as help retrieve and relearn information that has been forgotten, serving as the basis for further activities or tests (Armbruster, Reference Armbruster, Flippo and Caverly2000; Kobayashi, Reference Kobayashi2006a). Note taking is generally considered to be an effective learning strategy that allows students to encode information and permits later review to stimulate information recall (Dunkel, Reference Dunkel1988; Siegel, Reference Siegel2022).
L1 studies have maintained a strong interest in exploring the effectiveness of note taking. Several meta-analyses have been conducted to investigate the effect of taking notes (Kobayashi, Reference Kobayashi2005), reviewing notes (Kobayashi, Reference Kobayashi2006a), and note-taking mode (i.e., longhand versus digital note taking; Allen et al., Reference Allen, LeFebvre, LeFebvre and Bourhis2020; Voyer et al., Reference Voyer, Ronis and Byers2022) on L1 learning. Furthermore, Morehead, Dunlosky, Rawson, Blasiman, et al. (Reference Morehead, Dunlosky, Rawson, Blasiman and Hollis2019b) conducted a large-scale survey of university students to investigate their note-taking habits and preferences to inform teachers and researchers of how best to support and enhance students’ note-taking abilities and address their specific needs.
L2 research investigating the effects of note taking is less common. Several studies have compared one type of note taking (e.g., conventional note taking: Jin & Webb, Reference Jin and Webb2021; Cornell note taking: Hayati & Jalilifar, Reference Hayati and Jalilifar2009) with a control condition that did not take notes. Other studies have compared different types of note taking (e.g., conventional note taking versus Cornell note taking: Tsai & Wu, Reference Tsai and Wu2010; outline note taking versus conventional note taking: Song, Reference Song2012). Studies have also looked at note taking with unimodal input such as reading (Mežek, Reference Mežek2013; Najar, Reference Najar1997) or listening (Clark et al., Reference Clark, Wayland, Osthus, Brown, Castle and Ralph2014; Hale & Courtney, Reference Hale and Courtney1994). whereas others looked at note taking through multimodal input such as viewing (Sakurai, Reference Sakurai2018). Additionally, several studies have investigated learning through exposure to L2 input to enhance comprehension skills (Hayati & Jalilifar, Reference Hayati and Jalilifar2009; Moradi et al., Reference Moradi, Ghahari and Abbas Nejad2020), whereas other studies were directed at L2 learning to improve language ability such as vocabulary knowledge (e.g., Chen & Yang, Reference Chen and Yang2013; Jin & Webb, Reference Jin and Webb2021) and writing skills (Alzu’bi, Reference Alzu’bi2019). The influence of note taking on exposure to L2 input and learning the L2 also led to inconsistent results. For example, taking notes has been found to have a positive effect (Hayati & Jalilifar, Reference Hayati and Jalilifar2009), a negative effect (Hale & Courtney, Reference Hale and Courtney1994), and no effect (Clark et al., Reference Clark, Wayland, Osthus, Brown, Castle and Ralph2014) on listening comprehension. Studies have also found note taking to make both positive contributions (Jin & Webb, Reference Jin and Webb2021) and no contribution to L2 vocabulary learning (Kashani & Shafiee, Reference Kashani and Shafiee2016). Therefore, in L2 research the degree to which note taking contributes to learning as well as the variables such as learning targets that influence the effectiveness of note taking remain unclear.
One way to deepen our understanding of the effects of note taking is to conduct a meta-analysis. Because a relatively small number of L2 studies of note taking have been focused on language learning, this meta-analysis was intended to examine the effects of note taking through exposure to L2 input, including both L2 learning and content learning. Individual studies are restricted, for example, by their research contexts, participant population, and methodological features. In contrast, meta-analysis allows the examination of how the effects of note taking vary in relation to variables. Therefore, this study also investigated the extent to which the effect of note taking is moderated by different variables across studies such as learner characteristics (e.g., institutional level, region), treatment features (e.g., mode of input, material type), note-taking features (e.g., provision of note-taking strategy instruction, opportunity to review notes), learning targets (e.g., linguistic forms, listening), and measurement type (e.g., recognition, recall).
Effects of note taking through exposure to L2 input
A relatively large number of L2 note-taking studies have investigated the effects of conventional note taking (Chen & Yang, Reference Chen and Yang2013; Hale & Courtney, Reference Hale and Courtney1994; Jin & Webb, Reference Jin and Webb2021). Conventional note taking is a free-form note-taking behavior in which learners freely use their desired method to take notes. However, learning gains have been somewhat inconsistent. For example, studies have revealed that for learning L2 vocabulary, the percentages of learning gains were 12.5%–23% (Jin & Webb, Reference Jin and Webb2021) and 31.4% (Chen & Yang, Reference Chen and Yang2013); for listening comprehension, the percentage of learning gains ranged from 32.9% (Wilberschied, Reference Wilberschied1998) to 72.5% (Hayati & Jalilifar, Reference Hayati and Jalilifar2009).
Taking notes after strategy instruction is another highly examined area. Note-taking strategies involve using note-taking formats such as taking outline notes (Moradi et al., Reference Moradi, Ghahari and Abbas Nejad2020; Song, Reference Song2012) and Cornell notes (Hayati & Jalilifar, Reference Hayati and Jalilifar2009; Tsai & Wu, Reference Tsai and Wu2010), using vocabulary notebooks (Zarei & Adami, Reference Zarei and Adami2013), and providing learners with teachers’ verbal instruction to employ strategies such as attending to the teacher’s flow of speech (e.g., pause and tone of voice), organizational cues (e.g., numeric identifiers), and use of abbreviations before students take notes in a treatment (Balaban, Reference Balaban2017; Bozorgian & Pillay, Reference Bozorgian and Pillay2013). There is a great deal of variation in the percentage of learning gains in studies investigating learning from taking notes after strategy instruction. For example, Zarei and Adami (Reference Zarei and Adami2013) found that learners who wrote unfamiliar words in their vocabulary notebooks had gains of 67.8% on a recognition test and of 42.7% on a recall test. Walters and Bozkurt (Reference Walters and Bozkurt2009) found that using vocabulary notebooks led to gains of 35.7% on a recall test but gains of only 40.4% on a recognition test. Other types of note-taking instruction that are directed at increasing learners’ comprehension through note taking led to differences in the percentage of learning gains. Studies of reading comprehension after note-taking instruction have reported gains of 63.5% (Moradi et al., Reference Moradi, Ghahari and Abbas Nejad2020) and 84.2% (Najar, Reference Najar1997). There is similar variation in gains in studies of listening comprehension from taking notes after instruction, with results ranging from 44.8% (Aminifard & Aminifard, Reference Aminifard and Aminifard2012) to 83.7% (Hayati & Jalilifar, Reference Hayati and Jalilifar2009). The variation in the size of gains in both L2 learning and content learning is likely owing to the many differences in learner and treatment variables, note-taking features, learning targets, and measurement types between studies.
Review of moderator variables
Context
Learning settings were categorized into both foreign language (FL) and second language (SL) contexts. Learners in FL contexts tend to have minimal exposure to the target language outside of the classroom, whereas learners in SL contexts tend to have plenty of opportunities to encounter target language both inside and outside of the classroom (Webb & Nation, Reference Webb and Nation2017). No experimental studies have singled out these two settings as an independent variable. Also, studying in FL and SL contexts may be quite different, leading to varying effects of using note taking.
Region
The region in which the study took place might also affect learners’ note taking. Siegel and Kusumoto (Reference Siegel and Kusumoto2022) conducted a survey to compare Japanese and Swedish students’ perspectives and habits toward note taking and found cross-cultural differences in note taking. For example, Japanese learners tended to feel that note taking is difficult, whereas fewer Swedish learners reported difficulty in note taking. However, self-reported data can be biased (Wagner, Reference Wagner, Paltridge and Phakiti2015). Aggregating and analyzing data contributed by all empirical research can be more robust.
Orthographic script
Previous L2 research has investigated the effect of orthographic script (Krepel et al., Reference Krepel, de Bree and de Jong2021; Zhang & Zhang, Reference Zhang and Zhang2022). For instance, Spanish and English both have alphabetic scripts, whereas Chinese and English use different scripts. The distance in terms of orthography and representation is much shorter for L1 Spanish than L1 Chinese so it might be easier for Spanish students to learn English because they can depend on their L1 orthographic knowledge as a resource (Pasquarella et al., Reference Pasquarella, Deacon, Chen, Commissaire and Au-Yeung2014). A recent meta-analysis also found that sharing the same orthographic script between L1 and L2 allows learners to take advantage of L1 transfer, which may assist both L2 vocabulary acquisition and comprehension (Zhang & Zhang, Reference Zhang and Zhang2022). Because note taking is likely to involve the transfer of orthographic knowledge, the extent to which learning might be affected by the orthographic script between L1 and L2 would be useful to investigate.
Institutional level
Research has investigated the effects of note taking with participants at different places of study (elementary school, high school, language institute, university level). Kobayashi’s (Reference Kobayashi2005) meta-analysis of the encoding effect of note taking on L1 learning revealed that younger learners in primary and secondary schools benefited more from note taking than university students because note taking can compensate for lack of cognitive abilities and skills. However, a recent meta-analysis of listening-strategy instruction on L2 learners’ listening performance revealed that older learners tended to benefit more from strategy instruction because they were equipped with superior cognitive and metacognitive abilities (Dalman & Plonsky, Reference Dalman and Plonsky2022). Several meta-analyses examining L2 vocabulary learning also found that older learners tended to make greater learning gains than younger learners (de Vos et al., Reference de Vos, Schriefers, Nivard and Lemhöfer2018; Uchihara et al., Reference Uchihara, Webb and Yanagisawa2019). Given that no studies have examined age or place of study as an independent variable, the effects of note taking with L2 learners at different institutional levels remain unclear.
Mode of input
Research has investigated the act of note taking when encountering written input (Najar, Reference Najar1997), aural input (Hale & Courtney, Reference Hale and Courtney1994), and bimodal (Zarei & Adami, Reference Zarei and Adami2013) and multimodal input (Chen & Yang, Reference Chen and Yang2013). Different modes of input may affect the encoding function of note taking. For example, students are likely to sequentially alternate between reading and note taking, whereas note taking during listening involves only one step of simultaneous processing (Kiewra, Reference Kiewra1991). This might lead to higher learning gains from reading than from listening because reading provides opportunities for learners to adjust their reading rate to improve their concentration and organize their notes when they skim and reread specific types of information (Slotte & Lonka, Reference Slotte and Lonka1999). In contrast, note taking during listening involves multiple cognitive processes (Armbruster, Reference Armbruster, Flippo and Caverly2000; Piolat et al., Reference Piolat, Olive and Kellogg2005). Students must pay attention to a lecture, temporarily capture the important information provided by the instructor, hold and organize these ideas in working memory, and simultaneously write them down before they are forgotten.
Material type
Note-taking and L2 studies consist of both academic (Hayati & Jalilifar, Reference Hayati and Jalilifar2009; Zarai & Adami, Reference Zarei and Adami2013) and nonacademic input (Chen & Yang, Reference Chen and Yang2013; Jin & Webb, Reference Jin and Webb2021). Academic input such as academic lectures, discussions, conversations, and dialogues, the topics of which are academic in nature, may be more complex and dense than nonacademic input such as everyday conversations and teacher’s anecdotes (e.g., personal stories and experiences) that might be considered both informative and entertaining (Uysal & Tezel, Reference Uysal and Tezel2020). Thus, academic input may require students to engage in higher level thinking and deeper cognitive processing than nonacademic input (Tsai & Wu, Reference Tsai and Wu2010). Therefore, the complexity of academic input may pose a challenge for note taking. Because little attention was given to the effects of note taking between these two types of input, it is useful to examine material type as a variable to determine whether it affects learning.
Note-taking session
Note-taking studies involve participants taking notes within a single session or across multiple sessions. For example, Najar (Reference Najar1997) allowed students to practice taking notes over nine sessions and found positive encoding effects. In contrast, Hale and Courtney (Reference Hale and Courtney1994) did not find a positive encoding effect within a single session; however, Jin and Webb (Reference Jin and Webb2021) found that taking notes within a single session positively affected vocabulary learning. Examining number of sessions as a moderator may reveal the degree to which it influences the effects of note taking.
Note-taking behavior
Research investigating the use of note taking in the L2 context is typically categorized into two different scenarios: required note taking, which stipulates that students must take notes, and voluntary note taking, which allows students to take notes at their own discretion. Voluntary note taking may represent the most typical learning situation because students take notes as they choose (Hale & Courtney, Reference Hale and Courtney1994). Students who voluntarily choose to take notes are likely to be motivated and engaged, which may contribute to positive learning outcomes (Koren, Reference Koren1997). However, students who are required to take notes may benefit less from the encoding function of note taking because they may be less motivated and engaged with the note-taking process. Hale and Courtney (Reference Hale and Courtney1994) conducted the only study comparing these two types of note-taking behavior. They found that voluntary note taking had little effect on listening comprehension and required note taking negatively affected their learning performance. However, the negative effect of note taking does not mean there is less value of considering this variable, because many L2 primary studies have revealed a positive effect of voluntary note taking (Jin & Webb, Reference Jin and Webb2021; Wilberschied, Reference Wilberschied1998) and required note taking (Hayati & Jalilifar, Reference Hayati and Jalilifar2009; Kang, Reference Kang2010). A meta-analysis of studies with learners who have completed voluntary note taking and required note taking can shed more light on this variable.
Provision of note-taking strategy instruction
Even though most L2 learners acknowledge the value of note taking (Armbruster, Reference Armbruster, Flippo and Caverly2000), they often produce poor notes (e.g., verbatim transcript), which decreases the learning benefits of note taking. One way to make up for shortcomings in students’ personal notes is to employ note-taking strategies (Siegel, Reference Siegel2022) such as taking notes using a specific framework (e.g., Cornell notes) or recording key words and any useful linguistic forms. However, the results of note-taking-strategy instruction have been inconsistent. For example, after providing note-taking-strategy instruction, Kang (Reference Kang2010) found a large positive note-taking effect (g = 1.17) on L2 vocabulary learning. Similarly, Najar (Reference Najar1997) found a large positive note-taking effect (g = 1.11) on reading comprehension. However, Aminifard and Aminifard (Reference Aminifard and Aminifard2012) found little difference between note taking after instruction and no note taking on listening comprehension, whereas Kashani and Shafiee (Reference Kashani and Shafiee2016) found a negative effect of note taking after instruction on L2 vocabulary learning.
Instruction time
As noted by Plonsky and Oswald (Reference Plonsky and Oswald2014), examining the correlation between treatment length and outcomes would be useful for considering the cost/benefit ratio of interventions. Several meta-analyses have investigated length of strategy instruction (SI) in domains such as pragmatics instruction (Plonsky & Zhuang, Reference Plonsky, Zhuang and Taguchi2019) and listening SI (Dalman & Plonsky, Reference Dalman and Plonsky2022). Therefore, the synthetic approach of this study enables us to examine the relationship between the effectiveness of note taking and the overall duration of instruction time.
Opportunity to review notes
In L1 research, participants who had the opportunity to review their notes produced higher learning gains than those who did not have the opportunity to review their notes (Kiewra, Reference Kiewra1991; Kobayashi, Reference Kobayashi2006a). Several studies of L2 note taking have included the opportunity for participants to review notes (Hayati & Jalilifar, Reference Hayati and Jalilifar2009; Najar, Reference Najar1997; Walters and Bozkurt, Reference Walters and Bozkurt2009), whereas other studies have not included review of notes (Jin & Webb, Reference Jin and Webb2021; Kashani & Shafiee, Reference Kashani and Shafiee2016; Moradi et al., Reference Moradi, Ghahari and Abbas Nejad2020). However, no L2 studies have explicitly investigated the effects of reviewing notes on learning. Thus, the degree to which reviewing notes influences learning through exposure to L2 input remains unclear.
Learning target
Research has investigated the effects of note taking on L2 listening comprehension (Hayati & Jalilifar, Reference Hayati and Jalilifar2009; Tsai & Wu, Reference Tsai and Wu2010), reading comprehension (Moradi et al., Reference Moradi, Ghahari and Abbas Nejad2020; Najar, Reference Najar1997), vocabulary learning (Jin & Webb, Reference Jin and Webb2021; Zarei & Adami, Reference Zarei and Adami2013), grammatical structures (Kang, Reference Kang2010), and writing skills (Alzu’bi, Reference Alzu’bi2019). The degree to which note taking may affect different learning targets has yet to be examined.
Measurement type
The effect of note taking may vary depending on the measurement. L2 vocabulary research (Laufer & Goldstein, Reference Laufer and Goldstein2004) indicated that learners tend to score higher on recognition tests (e.g., multiple-choice items) than recall tests (e.g., write the meaning of the given word or write the L2 word that corresponds with a given meaning). Similarly, Walters and Bozkurt (Reference Walters and Bozkurt2009) found that note takers learned more words on recognition tests than on recall tests. However, in L1 research, Kobayashi’s (Reference Kobayashi2005) meta-analysis examining the encoding effects of note taking found that the benefits of note taking were greater when learning was measured by recall tests than by recognition tests. This finding is consistent with those of an earlier L1 study by Weener (Reference Weener1974). Thus, there is value in synthesizing prior note-taking studies to determine the influence of measurement type in the L2 context.
Research questions
This study has two main purposes. One is to better understand the overall effect of note taking on learning through exposure to FL or SL input, and the second is to explain potential moderators of the note-taking effect. To this end, two research questions (RQs) were formulated:
-
1. What is the overall effect of note taking through exposure to L2 input?
-
2. To what extent does the effectiveness of note taking vary across learner variables, treatment variables, note-taking features, learning target, and measurement type?
Method
Literature search
Following guidelines on literature search for meta-analyses (In’nami & Koizumi, Reference In’nami and Koizumi2010; Plonsky, Reference Plonsky2015), we comprehensively searched the following databases: ProQuest (including data subsets such as PsycINFO, ProQuest Dissertations and Theses, Linguistics and Language Behavior Abstracts, Education Resources Information Center), MLA International Bibliography, ScienceDirect, Social Sciences Citation Index. Additionally, Google and Google Scholar were searched. The following keywords were used to search for studies: (a) tak* notes, notetaking, note-taking, (b) learning, comprehension, vocabulary, grammar, and (c) second language, foreign language, and L2. Further, we manually reviewed journals that are widely cited in second language acquisition (SLA) and applied linguistics including Applied Linguistics, Canadian Modern Language Review, Computer Assisted Language Learning, Foreign Language Annals, Language Learning, Language Learning and Technology, Language Teaching, Language Teaching Research, Modern Language Journal, Reading in a Foreign Language, Second Language Research, Studies in Second Language Acquisition, System, and TESOL Quarterly. In addition, the reference sections in reviewed studies were carefully examined, resulting in the addition of two more studies (Clark et al., Reference Clark, Wayland, Osthus, Brown, Castle and Ralph2014; Sakurai, Reference Sakurai2018). We set July 2022 as the completion point for data collection. A PRISMA flow diagram (Page et al., Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann, Mulrow and Moher2021) outlining the study selection process can be found in the Supplementary Materials.
Selection criteria
The following criteria were applied:
-
1. The input must be a second or foreign language to the participants.
-
2. One of the independent variables measured must be note taking. This may include one, or a combination, of the following forms of note taking: taking notes using one’s own method, using a note-taking format such as Cornell notes, or taking notes after instructional interventions, meaning that learners were instructed in a combination of note-taking strategies such as how to identify main ideas and use abbreviations.
-
3. At least one dependent variable measured some aspect of L2 or content learning (e.g., listening comprehension, vocabulary learning).
-
4. When measuring vocabulary learning, the study must have focused on learning single-word items, not multiword items because only one study (Jin & Webb, Reference Jin and Webb2021) examined multiword items.
-
5. The studies must be experimental or quasi-experimental and include a control condition in which participants received the same treatment but did not take notes. A true control group that only completed tests was not considered in our study because none of the primary studies included such a group.
-
6. Only studies adopting between-participants designs were included. Studies that used within-participants designs were excluded due to a limited number of studies using within-participants designs (k = 7; Carrell et al., Reference Carrell, Dunkel and Mollaun2004; Clark et al., 2004). This decision was made based on the suggestion that combining both within- and between-participants design in meta-analyses is likely to produce biased results (Plonsky & Oswald, Reference Plonsky, Oswald and Plonsky2015).
-
7. Studies had to control for participants’ preexisting knowledge. Control could occur through the use of pretests revealing no statistically significant difference between groups on pretest scores (e.g., listening comprehension skill: Bozorgian & Pillay, Reference Bozorgian and Pillay2013; vocabulary knowledge: Jin & Webb, Reference Jin and Webb2021; reading comprehension skill: Najar, Reference Najar1997); or if the pretest was not administered, the participants were randomly assigned (Kang, Reference Kang2010); or their proficiency level had been checked and found to be homogenous (Kashani & Shafiee, Reference Kashani and Shafiee2016).
-
8. The studies must have been written in English.
-
9. The studies reported enough statistical information for an effect size to be calculated (i.e., mean, SD, and the number of participants tested).
-
10. The full text of the article was available.
-
11. The studies did not focus on learners with language learning problems.
We contacted the authors and gratefully received additional information from Uysal and Tezel (Reference Uysal and Tezel2020). After applying the inclusion criteria to the retrieved reports, a total of 21 studies (N = 1992) met the criteria and were included in the analysis. The studies consisted of 19 published journal articles and two doctoral dissertations. As only three studies reported delayed posttest results (Jin & Webb, Reference Jin and Webb2021; Kashani & Shafiee, Reference Kashani and Shafiee2016; Piri & Shirkhani, Reference Piri and Shirkhani2021), our data focused exclusively on the results of immediate posttests.
Coding
All studies were coded as specified by the coding scheme table. The table illustrates 27 variables, including 14 moderator variables. Coded features were divided into seven main categories pertaining to study identification, learner variables, treatment variables, note-taking features, outcome features, methodology, and study quality (see Supplementary Materials).
To assess the reliability of our coding procedure, 10 studies (approximately 47.6% of 21 studies) were randomly selected and independently coded by three raters.Footnote 1 All three raters are experts in one area of second language acquisition and had also carried out meta-analyses. The average agreement using Fleiss’s Kappa between three coders was .90. The S index (Norouzian, Reference Norouzian2021) was calculated at the item level of categorical moderators, and intraclass correlation was used for the continuous variables (see Supplementary Materials for item-level estimates). All discrepancies were discussed and resolved.
Learner variables
Context was coded as FL and SL. Region was also coded. At first it was coded as countries such as China, Iran, Sweden, and USA. Due to a limited number of studies in each country, we combined countries according to larger regions (i.e., Asia, Middle East, North America, Europe). The orthographic script used in learners’ L1 and L2 was coded as (a) same, (b) different, or (c) mixed. Institutional level was coded as elementary school, secondary school, language institute, and university.
Treatment variables
Mode of input consisted of three categories: reading, listening, and mixed. The label mixed included treatments in which bimodal or multimodal input was included. Material type was coded as either (a) academic or (b) nonacademic.
Note-taking features
The variable note-taking behavior was coded based on whether students voluntarily took notes or were required to take notes. The second note-taking feature was provision of note-taking-strategy instruction. We coded this variable as absence for studies that did not provide any note-taking strategies. That is, students used conventional note taking, which is taking notes using their desired method. Conversely, we coded the variable as presence when one or more note-taking strategies were provided. The provided note-taking strategy was further divided into linear learning strategy, generative learning strategy, and unreported. Although previous research (Bui & Myerson, Reference Bui and Myerson2014; Kobayashi, Reference Kobayashi2005; Siegel, Reference Siegel2022) commonly categorized note taking into “verbatim” and “generative” methods, this study employed the terms “linear” and “generative” to describe two distinct types of note-taking strategies. Following Ponce and Mayer (Reference Ponce and Mayer2014), a linear learning strategy involves processing information in the same structure presented in material, whereas a generative learning strategy involves selecting and reorganizing information to build a coherent structure. That said, a linear learning strategy is more focused on selecting and recording information following the flow of the material that is encountered without necessarily attempting to copy written or spoken input word for word and a generative learning strategy allows for more flexibility in how the information is organized. For example, using vocabulary notebooks and encouraging learners to write down key vocabulary or any useful linguistic forms was categorized as a linear learning strategy. Strategies such as encouraging learners to use abbreviations and record the gist of the material, break down long sentences into shorter sentences, or develop a specific note-taking format (e.g., Cornell notes) were categorized as generative learning strategies. Although outline notes are also called outline linear (Chen et al., Reference Chen, Teo and Zhou2017), constructing outline notes in this study was considered as a generative learning strategy because learners were allowed to personalize their outlines by using their own words, abbreviations, and symbols (Moradi et al., Reference Moradi, Ghahari and Abbas Nejad2020).
Opportunity to review notes
This variable was initially coded as review, no review, and unreported. However, we acknowledge that the quality of students’ notes may affect the usefulness of reviewing (Kobayashi, Reference Kobayashi2006a) and the provision of strategy instruction on note taking might enhance their note quality. To account for this, we further subdivided the studies into two groups based on the quality of notes: conventional note taking and note taking after strategy instruction. This allowed us to investigate the potential differences in the effect of reviewing notes between students who received instruction and those who did not. By doing so, we sought to provide a more nuanced understanding of how this variable might moderate note-taking effects on learning in the L2 context.
Outcome variables
For dependent variables, maximum score for the test, mean posttest scores, SDs, total number of participants, and participant number in each group were coded. The learning target was coded as linguistic forms, listening, reading, or writing. Learning linguistic forms referred to vocabulary learning except one study (Kang, Reference Kang2010), which examined the combination of grammar and vocabulary learning. Listening and reading consisted of listening and reading comprehension, respectively. One study (Alzu’bi, Reference Alzu’bi2019) looked at the influence of note taking on improving learners’ writing skills.
Measurement type
The type of measurement was initially categorized into different types of instruments such as meaning recall, form recall, multiple-choice tests, and True/False questions. Following two meta-analyses in L1 note taking (Kobayashi, Reference Kobayashi2005; Voyer et al., Reference Voyer, Ronis and Byers2022), recall (e.g., meaning recall, form recall) and recognition (e.g., multiple-choice test, True/False questions) were coded as two distinct measures. Moreover, one study (Alzu’bi, Reference Alzu’bi2019) investigated students’ improvement of writing skills, which is different from recall or recognition of information; therefore, the measurement type of this study was coded as free writing. Also, the results of Zohrabi and Esfandyari’s (Reference Zohrabi and Esfandyari2014) study did not differentiate the types of measurement, so the label mixed was included. To summarize, the measurement type was finally divided into categories of recall, recognition, writing, mixed, and unreported.
Number of note-taking sessions and length of note-taking instruction
These two moderators were continuous variables. It is important to note that note-taking sessions and instruction time might not cover the whole treatment sessions (e.g., Bozorgian & Pillay, Reference Bozorgian and Pillay2013). Therefore, only treatment related to note taking was considered and the total length of instruction time was coded in minutes.
Calculation of effect size
The first step of calculating the effect size was to use standardized mean difference, which is called Cohen’s d. Then, we converted Cohen’s d to the unbiased effect size Hedges’s g by multiplying a correction factor: J= 1 – [3(4 × df− 1)]. The reason for selecting Hedges’s g (Hedges & Olkin, Reference Hedges and Olkin1985) as the effect size is because it is an unbiased estimate of effect sizes compared with Cohen’s d, especially when samples are small (Borenstein et al., Reference Borenstein, Hedges, Higgins and Rothstein2009). To interpret the values for effect size for this study, we referred to Plonsky and Oswald’s (Reference Plonsky and Oswald2014) benchmarks for defining the magnitude of effect sizes: small = 0.40, medium = 0.70, and large = 1.00. The correlation coefficient was used to examine the role of the two continuously measured moderators (length of note-taking instruction, number of note-taking sessions), and the magnitude of effect sizes was small = .25, medium = .40, and large = .60.
To ensure independence of effects and minimize the presence of sample size inflation, all effect sizes pertaining to the same participants were averaged to form a single effect size for each sample. However, for studies reporting data on one or more treatment groups and a control condition (Hayati & Jalilifar, Reference Hayati and Jalilifar2009), effect sizes were calculated by contrasting each treatment group with the control condition on the immediate posttest (see Norris & Ortega, Reference Norris and Ortega2000, p.446). To answer RQ2, if a study reported multiple measurement types, these effect sizes were kept separate in analyzing the moderating effect of the type of outcome measures on the relationship in question.
Data analysis
We conducted all the analyses using the meta package (Schwarzer, Reference Schwarzer2007) in the R statistical environment (R Core Team, 2020). To address RQ1, we aggregated the effect sizes from the studies that compared note taking with no note taking (i.e., control condition) to produce a weighted mean effect size. When doing so, the effect sizes were weighted by inverse variance so that those with less sampling error contributed more to the meta-analytic mean (Plonsky & Oswald, Reference Plonsky, Oswald and Plonsky2015). To answer RQ2, we conducted subgroup analyses for moderator variables to determine whether the coded learner characteristics, treatment variables, note-taking features, L2 learning target, and measurement type (which all served as categorical moderator variables) as reported by the data set were significant moderators of the effectiveness of note taking. Considering the possibility that our categorization of subgroups might introduce new sampling errors at the subgroup level, random-effects modeling was used for between-subgroup comparisons while controlling for such sampling errors (Harrer et al., Reference Harrer, Cuijpers, Furukawa and Ebert2019; Plonsky & Oswald, Reference Plonsky, Oswald and Plonsky2015; Suzuki et al., Reference Suzuki, Kormos and Uchihara2021) and a between-group Q-statistic was used to examine the influence of moderator variables on effect sizes (ESs). For the two continuous moderators (number of note-taking sessions, length of note-taking instruction), the corresponding data were subjected to meta-regression analyses. Additionally, following Plonsky and Zhuang (Reference Plonsky, Zhuang and Taguchi2019), a correlational approach was adopted to examine their relationship with note-taking effects.
Sensitivity analyses
After aggregating the effect sizes, we detected outliers to ensure the robustness of the results and assessed publication bias that might influence the current data sets. It is important to mention that studies identified as potential outliers do not necessarily mean the study is an outlier that does not reflect normal language learning because each study was independently conducted and included a different group of students and varying learning conditions (Yanagisawa & Webb, Reference Yanagisawa and Webb2021). Therefore, we followed Viechtbauer and Cheung’s (Reference Viechtbauer and Cheung2010) guidance and reran the whole analysis while excluding the studies identified as potential outliers and compared the results with those that were obtained when including all studies (see Supplementary Materials for publication bias and sensitivity analyses results).
Results
RQ1: What is the overall effect of note taking through exposure to L2 input?
To answer the first RQ, 26 aggregated effect sizes were included (Figure 1). The results showed that there was significantly greater learning for note-taking conditions than for non-note-taking control conditions. However, the homogeneity test was statistically significant, Q(24) = 308.94, p < .001, indicating that variability in the true effect across studies as well as sampling error could have created this difference. The overall mean effect size from the posttest results for note taking versus no note taking was 0.56 (p = .002, 95% CI [0.24, 0.88]), a small to medium effect size according to Plonsky and Oswald’s benchmarks (Reference Plonsky and Oswald2014; between-group contrast, 0.4 for small, 0.7 for medium, and 1.0 for large). We then conducted moderator analyses to examine the extent to which different factors could account for this variability (see Table 1 for a detailed description of the subgrouped effect sizes in each category).
Figure 1. Forest plot for the posttest results of the comparison of note taking versus control.
Table 1. Moderator analyses (categorical variables)
Note. k = number of studies; n = number of effect sizes; NT = note taking.
a Any provided note-taking strategies directed at improving students’ note-taking performance such as introducing a specific note-taking format.
b Conventional note taking, meaning that students take notes using their own method without receiving any note-taking strategies.
RQ2: What variables moderate the effect of note taking?
Learner variables
Four factors relating to participant characteristics were examined: (a) context, (b) region, (c) orthographic script, and (d) institutional level. The analyses showed that the mean effect size for studies conducted in FL contexts (g = 0.69) was larger than that for studies in SL contexts (g = 0.18) and note-taking effect was the largest in Europe (g = 1.27), followed by Asia (g = 0.70), the Middle East (g = 0.54), and North America (g = 0.18). In terms of orthographic script, it was found that learners whose L1 has the same script as the L2 yielded a higher effect size in note taking (g = 0.96) than those whose L1 and L2 scripts were different (g = 0.58). Also, the note-taking effect was found to be the lowest in a diverse group of learners including those with the same script as the L2 and those without (g = 0.12). In relation to participants’ institutional level, the effect size of note taking from language institute students (g = 1.13) was the largest, followed by those for secondary school (g = 0.57) and for university students (g = 0.33). Also, note taking among elementary school students negatively affected L2 learning (g = -0.09), indicating that note taking might hinder learning for young learners. However, the effect size of secondary school and elementary school students should be interpreted with caution due to the small sample of each.
Treatment variables
Regarding mode of input, the differences in effect size between note taking through mixed input (g = 0.84) and reading (g = 0.85) were small. Both input modes tended to lead to greater learning than did taking notes when listening (g = 0.27). In addition, studies using nonacademic materials produced a larger effect than studies using academic materials (g = 0.64 and g = 0.51, respectively).
Note-taking features
The results showed that requiring students to take notes tended to produce higher learning gains (g = 0.71) than allowing them to take notes (g = 0.29). The effect of note taking is substantially larger when note-taking instruction was involved (g = 0.84) than without note-taking instruction (g = 0.16). Moreover, the effectiveness of note taking on learning varied depending on the type of note-taking strategy. A linear learning strategy was more effective (g = 1.04) than a generative learning strategy (g = 0.75).
In addition, the opportunity to review notes yielded a higher effect size (g = 0.55) than no review of notes (g = 0.35). It should be noted that after the use of note-taking strategies, there was a substantial difference between those who reviewed (g = 0.88) and those who did not review notes (g = 0.33). However, when participants had the opportunity to review notes without note-taking instruction, the effect size was smaller (g = 0.10) than when participants did not review notes (g = 0.33).
Learning target
Medium to large effects were obtained for treatment groups over control conditions in reading (g = 0.99) and learning linguistic forms (g = 0.73), a small to medium effect was found for writing (g = 0.57), and a more modest effect was found for listening (g = 0.22).
Measurement type
The largest effect was found with recall tests (g = 1.07). Somewhat smaller positive effects were observed among studies investigating other types of measurement such as writing a composition (g = 0.57) and taking recognition tests (g = 0.50).
Number of note-taking sessions and note-taking instruction length
Number of note-taking sessions and total amount of instruction time were two continuous variables (see Table 2 for descriptive statistics). Preserving the continuous nature of these two moderators is intended to build a more precise model of the correlation between the cost-ratio benefit of note taking in class and the effects of note taking. The results showed a medium to large positive correlation between number of note-taking sessions and note-taking effects (r = 0.51). For total amount of instruction time (in minutes), however, the correlation was small (r = 0.21). Meta-regression analyses also revealed that number of note-taking sessions was a significant predictor of note-taking effects (p < .001) but total amount of instruction time was not (p = .336).
Table 2. Descriptive statistics of continuous moderators
Note. k refers to the number of primary studies (among a total of 21 studies) that reported the corresponding information; NT = note taking.
DISCUSSION
The analysis of the overall effects of note taking revealed that learning through note taking was significantly more effective than learning without note taking (g = 0.56). This finding is important because it clarifies the value of note taking with respect to learning through exposure to L2 input. Several meta-analyses have already been conducted concerning the effects of note taking on L1 learning (Kobayashi, Reference Kobayashi2005, Reference Kobayashi2006a, Reference Kobayashi2006b), but a relatively small number of intervention studies from L2 contexts have focused on the effects of note taking (Siegel, Reference Siegel2022). However, the effect size found in this meta-analysis was larger than those that were reported in the earlier L1 meta-analyses. Kobayashi (Reference Kobayashi2005), for example, included 57 studies and found a mean weighted effect size of 0.22 for note taking versus no note taking. Therefore, the larger influence of note taking on learning through exposure to L2 input found in the present study reveals an area that warrants further attention.
However, it should, also be noted that the effect of note taking in this study was on both language learning and content learning. Previous meta-analyses that investigated the effects of other types of strategies were focused solely on L2 learning such as corrective feedback (d = 0.64; Li, Reference Li2010) and spacing (g = 0.58; Kim & Webb, Reference Kim and Webb2022), and these strategies were found to be slightly more effective than note taking. One possible explanation for the modest effect of note taking could be the absence of instruction from teachers on developing learners’ note-taking skills. Almost half of the included studies did not involve instruction in any note-taking strategies. However, this may have ecological validity because Siegel (Reference Siegel2019) reported that a considerable proportion of teachers may spend little time teaching note-taking skills. If students take poor notes, they may not effectively encode information and so may not fully benefit from taking and reviewing notes (Kobayashi, Reference Kobayashi2005).
The homogeneity assumption was found to be violated, Q(21) = 280.53, p < .001, indicating that the observed differences in effect sizes across studies could be due to both variability in true effects and sampling error. Therefore, potential moderating variables were further analyzed, and the effect of note taking was found to vary across a range of moderators pertaining to learner, treatment, note-taking features, L2 learning targets, and measurement type.
Moderator analysis
In relation to learner variables, the analyses revealed greater benefit from note taking in FL contexts (g = 0.68) over SL contexts (g = 0.12). This pattern of results aligns with findings from other meta-analyses that explored different SIs such as listening SI (Dalman & Plonsky, Reference Dalman and Plonsky2022) and reading SI (Taylor, Reference Taylor2014). A possible reason for this is that for FL learners taking notes may be more effective for improving learning (Siegel, Reference Siegel2018b) because note taking provides opportunities to optimize learning from the small amount of L2 input that FL learners typically encounter. However, because SL learners encounter large amounts of input inside and outside of the classroom, the effect of note taking might be mediated by many other positive variables such as frequency of encounters, which is a primary determinant of language acquisition and processing (Ellis, Reference Ellis2002).
The results also showed that learners in Europe (g = 1.27) tended to benefit more from note taking than those in Asia (g = 0.70), the Middle East (g = 0.54), and North America (g = 0.18). For FL learners, there was a substantial difference in note-taking effects obtained in Europe relative to Asia and the Middle East. One possible reason may be due to cross-cultural differences between perceptions of note taking. Siegel and Kusumoto (Reference Siegel and Kusumoto2022) reported a cross-cultural investigation of note taking between Sweden and Japan. The study investigated note taking from various perspectives such as L2 note takers’ views about note taking and education systems. For example, Japanese students tended to feel that note taking was a challenging activity, whereas Swedish students did not. The cross-cultural differences might also be attributed to orthographic transparency between L1 and L2. It is possible for L2 learners to record information in their L1 and/or L2 (Siegel, Reference Siegel2021). As information about language use in notes was limited to the included studies, exploring translanguaging in student notes was beyond the scope of this study. However, the results for the orthographic script moderator variable provide some indication that sharing the same orthographic script (g = 0.96) provided greater learning benefit than did sharing different orthographic scripts (g = 0.58). This finding is consistent with Zhang and Zhang’s (Reference Zhang and Zhang2022) meta-analysis indicating that sharing the same orthographic script could contribute better to learning through exposure to L2 input. The magnitude of effect sizes in this study is larger than the effect sizes from Zhang and Zhang’s (Reference Zhang and Zhang2022) study, perhaps due to additional benefits of note taking.
With respect to the significantly moderating effect of institutional level, the results showed a substantial difference in the effects of note taking from language institutes (g = 1.13) relative to secondary (g = 0.57) and postsecondary schools (g = 0.34). This finding is in line with a previous meta-analysis that examined listening SI on L2 listening comprehension (Dalman & Plonsky, Reference Dalman and Plonsky2022). This result may relate to higher motivation to learn a target language for language institute students who choose to study an L2 compared with secondary school or university students who may be required to study an L2. Interestingly, it was found that note taking negatively affected learning for elementary school students (g = - 0.46). Note taking is a complex and demanding activity (Piolat et al., Reference Piolat, Olive and Kellogg2005), which might present a significant challenge for younger learners. However, because there was only one sample included at the level of elementary school students, further research is warranted to test this interpretation.
Of the treatment moderator variables, the analyses revealed that learning gains were larger for note taking with mixed (g = 0.84) and written input (g = 0.85) than with aural (g = 0.27) input. Smaller effects of note taking on listening might be expected because it is difficult for learners to adjust their focus and organize their notes due to time pressure (Bui & Myerson, Reference Bui and Myerson2014). The analyses also revealed that the effectiveness of note taking is more pronounced when using nonacademic input (g = 0.64) than when using academic input (g = 0.51), suggesting that students can derive greater benefits from their note-taking efforts by incorporating nonacademic materials into their learning process. Although academic materials are often perceived as the primary source for learning through exposure to L2 input, our results indicated that nonacademic materials also play a crucial role in enhancing the learning experience through note taking.
Regarding note-taking features, a medium effect size was found when requiring students to take notes (g = 0.71), whereas a small effect size was found when allowing students to take notes (g = 0.29). This can be expected because students who were required to take notes were more likely to pay attention and engage with the material being presented. However, the findings contrast with the results that were obtained by Hale and Courtney (Reference Hale and Courtney1994) who found little effect of voluntary note taking and a negative effect of required note taking on listening comprehension. Further research investigating how these two note-taking behaviors affect learning would be useful.
The results also revealed a significant positive effect of instruction on note-taking strategies (g = 0.84), suggesting that learners may not know how to effectively take notes. Instruction in note-taking strategies can help students more effectively select, organize, and elaborate on information in their notes (Siegel, Reference Siegel2018a). This is supported by Mayer’s (Reference Mayer1996) SOI (i.e., selection, organization, integration) model of learning, which suggests that meaningful learning occurs when learners are able to select relevant information, organize it into a structure in working memory, and integrate it with prior knowledge from long-term memory. When looking at different types of note-taking strategies, the results showed that a linear learning strategy (g = 1.04) led to greater learning than a generative learning strategy (g = 0.75). This finding might be surprising given the general optimism that surrounds the effectiveness of learning from generative note taking (e.g., Siegel, Reference Siegel2022). The larger effect for linear learning strategies is likely due to the focus on learning in this meta-analysis. Learners who used a linear learning strategy may direct more attention to unfamiliar linguistic forms, whereas in generative learning strategies there is additional focus on making decisions about the importance of content and deciding when and how to take notes (Siegel, Reference Siegel2021). This is supported by input processing theory, which states that learners’ processing of input is largely determined by what they attend to (Han & Peverly, Reference Han and Peverly2007). It should also be noted that both types of note-taking instruction are helpful for enhancing learners’ ability to process input, as the effect sizes for note taking are medium and large in generative and linear learning strategy, respectively.
The results also suggested that reviewing notes may not yield positive results unless learners were provided with note-taking-strategy instruction. After note-taking instruction, students who reviewed notes showed greater learning gains (g = 0.88) than those who did not have opportunities to review notes (g = 0.37). Therefore, providing note-taking instruction and allowing students to review their notes could optimize L2 learning. As no L2 empirical research to date has explicitly examined the effect of students reviewing their own notes after they have completed the process of note taking, this finding may provide some insights into the role of the external storage function of note taking in the L2 context.
In relation to different learning targets, note-taking effects appeared to be larger for reading (g = 0.99) and linguistic forms (g = 0.76), followed by writing (g = 0.57) and listening (g = 0.22). This pattern of results is similar to those of Plonsky (Reference Plonsky2011) that the effectiveness of SI is larger in reading (g = 0.74) and vocabulary (g = 0.64) than in writing (g = 0.42) and listening (g = 0.06). Therefore, this finding builds on previous research by demonstrating the significance of note taking in enhancing specific language skills, particularly in reading comprehension and linguistic forms. However, because there is only one study that investigated note-taking and writing skills, further research in this area is warranted.
For the moderator of measurement type, the analysis found that note-taking effects were more pronounced when learning was measured by recall tests (g = 1.07) than by writing (g = 0.57) and recognition tests (g = 0.50). This is consistent with an earlier meta-analysis of L1 note taking that indicated that the benefits of note taking were greater when measured with recall tests than with recognition tests (Kobayashi, Reference Kobayashi2005). This suggests that note taking can be an efficient strategy for helping learners to successfully recall newly learned knowledge from memory because recall tests tend to be a relatively demanding test format in comparison with recognition test formats (Laufer & Goldstein, Reference Laufer and Goldstein2004). The depth-of-processing hypothesis provides support for the effectiveness of note taking in that greater depth of processing results in superior recall (Craik & Tulving, Reference Craik and Tulving1975) and note taking is found to bring about deeper processing of information (Titsworth & Kiewra, Reference Titsworth and Kiewra2004). Although writing a composition typically involves higher level processing, investigating note-taking effects when measured by writing skills in future research is crucial due to the limited number of studies (k = 1).
Regarding the length of instruction, the results revealed a small but positive correlation (r = 0.21) between duration of instruction and note-taking effects, suggesting that longer instruction is more effective than shorter instruction. However, the small correlation indicated that the benefits of extensive instruction time may be limited. Additionally, a medium to large positive correlation was found between number of note-taking sessions and note-taking effects (r = 0.51), suggesting that the frequency or consistency of note-taking practice might be more important than the length of instruction time. This is supported by skill acquisition theory (e.g., DeKeyser, Reference DeKeyser, VanPatten and Williams2007), which suggests that repeated and focused practice as well as revision of skills is important to develop abilities such as note taking. That said, with limited class time, integrating note-taking practice as an ongoing component of instruction and providing regular opportunities for students to engage with this skill may yield favorable learning outcomes.
It is also important to note that the results of the present study focused on the magnitude of the observed difference instead of the direction (i.e., p values) to provide a more nuanced understanding of the results. Within moderator analyses, comparisons are often underpowered, so the findings need to be interpreted with caution.
Limitations and future directions
This meta-analysis provided preliminary evidence that note taking has a small to medium positive effect on learning through exposure to L2 input. Given that note taking is not confined to L2 classrooms, the findings of this study may shed light on the broader implications of the transferability of note-taking skills across languages, disciplines, and beyond educational context (i.e., work context).
However, several limitations were identified that would be useful to address in future research. First, it would be useful for additional studies to include (1) younger learners such as elementary and secondary school students, (2) comparisons of different note-taking types such as outline versus Cornell note taking, (3) greater focus on other L2 learning targets such as grammar learning, and (4) the use of digital note taking on laptops or tablets. It should be noted that this meta-analysis specifically focused on longhand note taking because no studies to date have compared learning in digital note taking versus a control condition. It is also worth noting that digital note taking has received significant attention in L1 research (e.g., Luo et al., Reference Luo, Kiewra, Flanigan and Peteranetz2018; Morehead, Dunlosky, & Rawson, Reference Morehead, Dunlosky and Rawson2019a; Mueller & Oppenheimer, Reference Mueller and Oppenheimer2014) and should also be given appropriate attention in L2 research.
Moreover, researchers mainly looked at note taking through reading or listening. However, with the development of technological advancements, the use of multimodal input such as viewing for L2 learning has become increasingly common (Feng & Webb, Reference Feng and Webb2020; Webb, Reference Webb, Nunan and Richards2015). Only one study to date (Sakurai, Reference Sakurai2018) has investigated L2 learners’ note taking through viewing.
Also, we would like to encourage researchers to consider reporting information recorded in student notes. It would be useful for teachers and researchers to better understand the role of translanguaging in note taking in the L2 context. The bilingual spelling in alphabetic systems (BAST) model (Tainturier, Reference Tainturier, Olive and Perret2019) proposed that the level of coactivation in bilingual spelling is influenced by the extent of similarity between the orthographic and phonological aspects of the two languages. Moreover, the relative proportion of high and low similarity in word form and meaning between two languages can affect the degree of facilitation or inhibitory effects that occur when encoding information (Iniesta et al., Reference Iniesta, Rossi, Bajo and Paolieri2021). Therefore, future research is recommended to investigate the influence of cross-linguistic similarity on note taking and learning through exposure to L2 input.
Regarding study quality in L2 research, out of 21 studies, 15 (71%) reported pretest scores, 8 (38%) reported instrument reliability, and only 3 (14%) included a delayed posttest. Plonsky and Zhuang (Reference Plonsky, Zhuang and Taguchi2019) found that higher study quality is correlated with larger effect sizes. There is a need for future research to increase and report methodological quality (e.g., instrument reliability, large sample size, pretest and delayed posttest inclusion, random assignment) to make studies stronger and yield larger effects.
Finally, it is recommended that a minimum of three studies is included in each moderator subgroup (Li, Reference Li2016; Vuogan & Li, Reference Vuogan and Li2023). However, Plonsky (Reference Plonsky2011) suggested that the prospective value of a meta-analysis sometimes might be greater than its retrospective value. Therefore, given the limited number of L2 studies that have examined the effects of note taking on learning, we did not exclude categories with less than three ESs. Although these analyses are meaningful because they indicate the trends in the data, further research on note-taking effects is needed to reach firmer conclusions.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0272263123000529.
Data availability statement
The experiment in this article earned Open Data and Open Materials badges for transparent practices. The materials and data are available at http://doi.org/10.48316/p9RfC-llHfQ
Acknowledgments
We would like to express our sincere gratitude to the Handling Editor Luke Plonsky and three anonymous reviewers for their constructive feedback that helped improve our manuscript significantly. We are also grateful to Lizz Huntley and Andrew Dennis for their great support during our manuscript submissions. We also thank the following researchers for providing assistance and feedback for the current meta-analysis: Akifumi Yanagisawa, Takumi Uchihara, Emi Iwaizumi,Yanxue Feng, and Dru Michael Sutton.
Open access
