研究論文
Group Assessment of Verbal Short-Term Memory and Verbal Working Memory in Sixth-Grade Elementary School Students
2024 年 2024 巻 1 号 p. 165-181
詳細
2024 年 2024 巻 1 号 p. 165-181
SAKUMA Yasuyuki (Fukushima University) TAKAKI Shuichi (Fukushima University) YUZAWA Masamichi (Hiroshima University)
Keywords: verbal short-term memory, verbal working memory, language familiarity effect
Among the multiple components of working memory (WM), the phonological loop and central executive are essential for language acquisition. These functions can be evaluated individually using the word span test (WST) and forward- and backward-order digit span tests (DSTs), which are based on serial recall, in which high-frequency numbers and words are presented verbally in a certain order to elementary school students and recalled orally. In this study, we measured the phonological loop and central executive using a bilingual (Japanese language as the first language and English language as a foreign language) version of the Hiroshima University Computer-based Rating of Working Memory (HUCRoW), a computer-programmed group-based measurement tool. The responses to the test were recorded using serial reconstruction, in which participants selected the text choices based on the spoken information presented, as opposed to the conventional verbal serial recall. This study explored whether the findings from traditional measures (i.e., native language dominance due to the language familiarity effect) are replicated in the bilingual version of HUCRoW. The results show that the language familiarity effect was evident in all three tests. In addition, similar results were obtained for serial reconstruction by HUCRoW and the conventional measure (serial recall) in forward-order DST.
The primary objective of the 4-year curriculum for English as a foreign language, which commences during the middle grades of elementary school, is to familiarize students with fundamental English phonetics. Furthermore, in the upper grades, students are required to develop the ability to transcribe orally received vocabulary into the written form. By the time elementary school children graduate, they are expected to have acquired a vocabulary of 600–700 words in both spoken and written language through the repetitive learning of English vocabulary during
English classes. This repetitive learning process is closely related to the temporary retention and immediate processing of language information. More effective temporal retention and processing lead to the acquisition of words. The cognitive function responsible for temporal retention and processing is known as working memory (WM). WM has several components, including the phonological loop and the central executive system, both of which are essential for the acquisition of first and second languages (Baddeley, 2003; Cowan, 2013; Gathercole & Baddeley, 1993; Juffs & Harrington, 2011; Linck et al., 2014; Wen, 2012, 2014; Williams, 2012). Both of these components exhibit notable differences of the memory spans, even among individuals of the same age (Gathercole & Alloway, 2008). In this regard, understanding the characteristics of WM, a cognitive resource that underpins the cognitive activities related to learning, is essential for the effective implementation of individualized instruction in English classes.
The Hiroshima University Computer-based Rating of Working Memory (HUCRoW) enables the evaluation of individual WM characteristics within a group in the class. Initially developed in Japan, this program serves as a group-based WM assessment tool for elementary and junior high school students. It includes both Japanese and English as language materials. The English version of WM assessment was developed recently.
In the present study, we used HUCRoW to assess WM functions (i.e., phonological loop and central executive) related to two languages (e.g., Japanese as the first language and English as the foreign language) among sixth-grade students who had studied English for four years in elementary school. We administered the Japanese and English versions of the word span test (WST) and forward- and backward-order digit span tests (DSTs), which involve tasks encompassing words and numbers familiar to elementary school students.
The WST and forward- and backward-order DSTs evaluate WM functions (i.e., phonological loop and central executive system), which are crucial elements for language acquisition (Pickering, 2006). In the case of Japanese as a native language, the words and numbers used in these tests are frequently encountered in daily life. Similarly, the English materials included within these tests are frequently encountered in elementary school English classes, albeit as a foreign language. Both test materials are firmly lodged in the long-term memory of sixth-grade children in elementary school and are regularly retrieved. The bilingual versions of the three tests involve the immediate recall of a predetermined sequence of verbal stimulus materials. An important aspect of interlingual differences within this type of task is that of the language familiarity effect, which posits that immediate memory recall is higher for familiar language materials than unfamiliar language materials. For example, the degree of familiarity is higher for one’s native language than for the second language, indicating the pronounced advantage of the native language in this
context.
Multiple studies have supported the language familiarity effect in Indo-European speaking countries (e.g., Chincotta & Hoosain, 1995; Chincotta & Underwood, 1996, 1997; da Costa Pinto, 1991; Thorn & Gathercole, 1999, 2001; Thorn et al., 2002). For example, Thorn and Gathercole (1999) demonstrated that English-speaking children whose first language was English performed better on the DST sequential recall task for English (i.e., native language) numbers than for French (i.e., the second language) numbers.
This effect is also observed in the immediate serial recall task, even when participants are subjected to articulatory suppression, which requires repeated utterances of words unrelated to the task to prevent participants from rehearsing articulation (Thorn & Gathercole, 2001). These effects were also observed in the probed recall task but not in the continuous recognition task (Thorn et al., 2002). These findings imply that the language familiarity effect is not attributable solely to faster articulation and shorter output latency in the familiar language due to faster recall, but is derived primarily from the recall process itself.
In contrast to European countries, few studies have investigated sequential recall using the Japanese and English versions of the DST among Japanese elementary and junior high school students whose native language is Japanese. A study evaluated the 3-year developmental progression of elementary and junior high school students who were taught English in elementary school (Sakuma & Saito, 2012). Another study focused on the impact of four years of learning English in elementary school (Sakuma, 2018). Both studies reported the dominance of the native language due to language familiarity effect.
Phonological loop is characterized by domain-specificity, which depends on the type of verbal information. This feature is one of the subsidiary systems of the central executive, akin to an audio tape recorder with a specific loop length. Language, including phonological and written information, is recorded in an auditory format, following the sequence in which it is perceived. Unless rehearsed by the individual, this recorded information attenuates rapidly or is overwritten by new auditory input. The phonological loop plays a crucial role in language acquisition due to its involvement in verbal short-term memory (STM) and language processing. It contributes significantly to the storage of information that can be expressed verbally, such as numbers, words, and sentences. Individuals with a larger phonological capacity exhibit superior vocabulary and language acquisition abilities compared to those with a smaller capacity. The phonological loop can be conceptualized to have two components: phonological store and articulatory rehearsal system (i.e., articulatory control process) as shown in Figure 1.
Phonological store involves the retention of phonological information during the seconds preceding the loss of the input memory trace (e.g., verbal material). Conversely, articulatory control pertains to a syntactic rehearsal process that continues to be retrieved from storage and restored by articulation (Pickering, 2006). In other words, the articulatory control process is responsible for vocalizing visually coded information and rehearsing phonologically coded
information to reactivate representations that are in danger of being lost from the phonological storehouse.
Phonological Loop (Logie, 1995
The WST and DST sequential units of span assess verbal STM as measured by immediate serial recall of words or numbers. These tests focus on the phonological loop function. The DST entails a series of fixed span units consisting of digits from one to nine. Each sequence is promptly recalled after hearing a fixed span unit sequence. This test evaluates the phonological store capacity for temporarily retaining and recalling a sequence of numbers, each time presented in a different sequence. During recall, familiarity with these numbers is also influenced by existing knowledge within the long-term memory store.
The WST, similar to the forward-order DST, involves the recall of a sequence of phonetically presented words within a fixed span of time. Nevertheless, words possess significantly richer semantic content than numbers. Consequently, performance on the WST is also influenced by existing knowledge within long-term memory stores. In the case of words, the number of syllables, phonological structure, and semantic associations are extremely diverse compared to numbers. Consequently, the WST requires the instantaneous reconstruction of a substantial volume of verbal information inherent in long-term memory stores. In both cases, these tests assess the phonological storage capacity within the phonological loop. Moreover, these assessments also encompass the influence of linguistic knowledge stored within the long-term
memory during processing, retrieval, and recall of verbal stimuli.
The central executive acts as an overarching controller for both information residing within the long-term memory as well as the three other subcomponents (i.e., phonological loop for being capable of holding and rehearsing sound and speech-based information; visuospatial sketchpad for performing non-verbal visual material; and episodic buffer for being a temporary multimodal store that combines information from the phonological loop and visuospatial sketchpad subsystems). This function has domain-generality and is not contingent on stimulus features. Furthermore, the central executive is involved in multiple operations, such as coordinating retrieval and encoding, selecting relevant information, suppressing irrelevant information, associating new information with existing knowledge in long-term memory stores, and updating new long-term memory representations.
The reverse recall component of the DST assesses functions associated with the central executive. In this task, distinct sequences of numbers from one to nine are presented, similar to the forward-order DST, but the numbers are recalled immediately after they are presented in a reverse order to the original chronological presentation of the fixed span unit. Consequently, in this task, individuals are required not only to retain the numerical information but also to recall it in a reverse order. This task involves a much more complex cognitive process than the sequential recall in DST. The backward-order DST entails inhibiting the initial order of the input numerical information, temporarily retaining this information, and then recalling it in a sequence that is the complete opposite of the input order. Therefore, the execution of this task requires selecting the pertinent information while filtering out the unnecessary information.
HUCRoW is a Japanese version of the test that was originally developed for group-based WM assessment of native Japanese-speaking elementary and junior high school students. This tool aligns with a multi-component model of WM comprising three components (Yuzawa et al., 2019). During the development of the Japanese version, we used the Automated Working Memory Assessment (Alloway, 2007), a commercially available WM test for native English-speaking children, and the WM Behavior Scale (Alloway et al., 2008). The validity of the Japanese version of HUCRoW has been confirmed. The Japanese version encompasses eight task configurations, divided into four categories based on Baddeley and Hitch’s (1974) model: verbal STM, which involves phonological loops; visuospatial STM, which involves visuospatial memory; verbal WM, which involves phonological loops, mainly in the central executive; and visuospatial WM, which involves visuospatial memory, mainly in the central executive. Two tasks assess each of these categories. The forward-order DST and the WST evaluate the verbal STM, whereas the listening span test and the backward-order DST evaluate the verbal WM. Visuospatial STM is assessed using the line span test and the figure span test, whereas visuospatial WM is assessed using the comparative line span test and the rotated figure span test. In addition,
four tasks are included in the English adaptations of the Japanese version of verbal STM and verbal WM. As this study focused on WM functions related to language in children who have learned English, we will describe the characteristics of HUCRoW in relation to the abovementioned verbal STM and verbal WM tasks.
In the group WM assessment, each student is seated in front of a computer and performs tasks at their own pace, which enables evaluation of the cognitive characteristics of each individual learner. Conventional language tasks often involve serial recall, wherein spoken language information is reproduced verbally in a specific order. The responses to such tasks are recorded for each individual. However, in a group setting, it becomes difficult to determine the individual responses due to potential interference caused by the simultaneous verbal reproduction of speech among participants. Therefore, to enable the execution of this task within a group setting, the participants were required to respond through a serial reconstruction method, which involves selecting multiple alternatives presented in written form based on a predetermined order when responding to verbal information. This task entails a shift in modality between the input (i.e, spoken language) and output (i.e, written language) stimulus materials. In particular, instead of recalling the spoken language information emanating from the personal computer (PC), a list of multiple language stimuli is presented on the PC screen as written language information. The users respond to questions by clicking on these stimuli in a certain order using the computer mouse.
The serial reconstruction method used to record the responses of the Japanese version of
HUCRoW has been proven to be a valid alternative to the verbal serial recall in spoken language (Yuzawa et al., 2019). However, the potential effects of the response process in the English version of HUCRoW developed in this study need to be explored further.
The purpose of this study was to determine whether data obtained from serial reconstruction using the English version of the three language tasks (i.e., WST and forward- and backward-order DSTs) in HUCRoW demonstrate similar findings to the conventional oral serial recall (i.e., dominance of the native language due to language familiarity effect). If the results of the conventional oral serial recall, based on individual test implementation, can be reproduced using HUCRoW (i.e., serial reconstruction) in a group setting, the validity of test administration in a group setting can be demonstrated, despite the fact that these three tasks are part of the HUCRoW. Therefore, a bilingual version of these tasks was administered to the participants. In particular, we evaluated the reproducibility of forward-order DST by comparing our results with those reported previously by Sakuma (2018), who used a conventional oral recall method among children who had studied English for four years in the same elementary school where the present research was conducted. The research questions (RQs) for this study were as follows:
RQ1: Is there a dominance of native language in forward-order DST and WST, which mainly evaluate the function of the phonological loop?
RQ2: Is there a dominance of native language in backward-order DST, which mainly evaluates the function of the central executive?
Prior approval for this study was obtained from the university’s research ethics committee. In addition, the written consent forms were also obtained from the participants themselves and their parents prior to the present study. In total, 92 sixth-grade students who had been attending a weekly class of English as foreign language in the third and fourth grades and two weekly classes in the fifth and sixth grades participated in the December survey. Of these participants, 77 completed the tasks, including the non-verbal tasks of HUCRoW, that were included in the analysis. In addition, the serial reconstruction method used in HUCRoW was compared with the conventional oral serial recall method used in forward-order DST. Furthermore, we also compared our results with those of Sakuma (2018), who used the conventional oral serial recall method. Sakuma enrolled 90 sixth-grade students who had studied English for four years in the same elementary school as the school where this study was conducted.
The Japanese EFL sixth-grade students participating in this study had been exposed to highly frequent vocabulary items in their English lessons, but were still seen to have less skill in English than in Japanese. Thus, we compared the Japanese and English versions of the serial reconstruction tasks in WST and forward- and backward-order DSTs related to word units (words or numbers) to the STM and WM tasks in HUCRoW to evaluate their WM functions for word processing.
The participants were presented with a series of speech stimuli consisting of two syllables. They were instructed to select the nine words (Figure 2), such as kagi (key), ashi (foot), pan (bread), kutsu (shoes), kago (basket), isu (chair), basu (bus), koma (spinning top), and kasa (umbrella) presented in Japanese hiragana, displayed on the PC screen. The number of words presented increased sequentially, starting with one span and increasing to a maximum of nine spans.
Japanese Version of WST on HUCRoW
Nine English words (i.e., wolf, bear, bird, sheep, cow, bee, frog, pig, and rat) were presented as monosyllable audio stimuli with high-frequency audio and text information in English classes. Similar to the Japanese version, in the English version, the participants were presented with speech and instructed to select the nine English words presented on the PC screen in the same order as the speech. The words were presented in a different alphabetical arrangement at each instance. The numbers of words presented on the PC screen were increased sequentially, starting with one span and increasing to a maximum of nine spans. Prior to administering the English version of WST, all participants had learned the letters, and their ability to hear the sound and read the letters were confirmed.
The participants were presented audibly with a series of Japanese or English language numbers. They were required to select the numbers presented on the PC screen in the same order as that of the presented stimuli. The number of stimulus numbers ranged from one to nine, increasing by one word each span. The sequences of the audio stimuli and questions presented on the screen were similar for both language versions.
Similar to DSTs, the participants were presented with audio stimuli of numbers from one to nine audio in Japanese or English language. They were instructed to select the numbers presented on the PC screen in the backward order of the presented stimuli, which differed between each trial.
The number of words presented on the PC screen were two in the first span, which increased by one in each subsequent span. The sequences of words presented on the PC screen were similar for both language versions.
The experiment was conducted in two sessions of 45 minutes each, within the class, in the school’s computer lab. Breaks were not allowed during the test called “the game” to avoid interruptions. The investigator initiated the HUCRoW on the PC, entered each child’s identification number, date of birth, and sex (Figure 3) on the PC, and wrote each child’s identification number, name, and date of birth on the test progress sheet (see Appendix). Each participant was seated in front of a desktop computer and followed the instructions presented through headphones. The options presented on the PC screen were selected using the mouse. The participants were instructed to raise their hands once the test was completed to allow the completion of the assignment progress checklist. The next test was started once the investigator had provided further instructions. Each test was preceded by a practice session to confirm that the participants had understood the instructions.
Personal Information Entry Screen of HUCRoW
The span scores of the Japanese and English versions of HUCRoW were calculated based on the number of correctly selected words among trials in which all target words were selected in the correct order (Friedman & Miyake, 2005). Each test comprised six trials per span, and the task was terminated when four or more trials had failed. Individuals with three or more successful trials were allowed to proceed to the next span of trials with the addition of another stimulus digit. The
number of successful trials until the termination of the task indicated the score of the respective participant. For example, if a participant had six correct trials for one span, five for two spans, four for three spans, three for four spans, and one for five spans, the test was terminated after five spans and the total score was 19 (i.e., 6 + 5 + 4 + 3 + 1).
Conversely, in the conventional oral serial recall of forward-order DST, participants with a correct response on at least one of two trials for each digit span were allowed to proceed to the next digit span, whereas the test was terminated for those with incorrect responses on both trials. In the final span, if correct responses are provided for both trials, a score of one is recorded, whereas a score of 0.5 is recorded if the correct response is obtained for only one trial. For example, if two trials are correct in the first span, two in the second span, and one in the third span, with two incorrect trials in the fourth span, the test is terminated in the fourth span, and the total score is calculated as 2.5 points (i.e., 1 + 1 + 0.5 + 0).
To validate RQ1 and RQ2, t-tests were performed and effect sizes were calculated for the Japanese and English versions of the three tests. For the forward-order DST, statistical analysis was performed, similar to Sakuma (2018) and the serial reconstruction of HUCRoW. Furthermore, correlation coefficients were calculated to explore the relationship of each test between the two languages. The forward- and backward-order DSTs have identical semantic content in both languages. Although the English and Japanese versions of WST differ in terms of the linguistic form and semantic content, the cognitive task involved in both versions is similar as both involve serial reconstruction from nine words, implying that a relationship exists between the two tests.
Table 1 presents the descriptive analyses of the Japanese and English versions of the conventional verbal serial recall of forward-order DST and the serial reconstruction of the three tests of HUCRoW.
Descriptive Analysis of the Japanese and English Versions of Tests
Tests (HUCRoW/individual) Japanese Version English Version
| N |
|
|
|
|
|
|
|
|---|---|---|---|---|---|---|---|
|
|
|
|
|
[6.38, 6.84] |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
As mentioned in 3.4, due to the differences in calculation methods, the numerical score differs between the two versions of forward-order DSTs: the oral serial recall used by Sakuma (2018) and serial reconstruction of HUCRoW used in the present study.
The mean scores of the Japanese versions on the conventional oral serial recall of DST and the three tests of HUCRoW were higher than those of the English versions.
Table 2 presents a comparison of the results of t-tests and effect sizes between the Japanese and English versions of all tests. Furthermore, Table 3 shows the correlation coefficients for both versions of each test.
Effect Sizes and t-Tests for Japanese and English Versions of Tests
|
|
|
t | df |
|
d |
|
|---|---|---|---|---|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Correlation Coefficients Between Japanese and English Versions of Tests
Note. *p <. 05. ** p <. 01.
The t-test statistics and effect sizes were significantly higher for the Japanese versions than the English versions of all three HUCRoW tests (ps < .001), indicating the superiority of the Japanese version (i.e., native language).
Thus, our findings provide affirmative results for RQ1 and RQ2. Therefore, the native language dominates in complex cognitive tasks, which are mediated by the central executive and the phonological loop in the WM.
Furthermore, the correlation analysis between the Japanese and English versions of the HUCRoW showed positive correlations for all three tests. In particular, the two types of DSTs, the forward- and backward-order, showed moderate correlations for both (r =. 51, r =.49), as shown in Table 2. In particular, with regard to the comparison between Sakuma’s (2018) conventional
verbal serial recall of DST in forward order and the HUCRoW serial reconstruction tasks in forward order, the effect sizes were significantly higher for serial reconstruction with d = 1.85 and d = 2.97, respectively, as shown in Table 2. Similarly, the respective correlations were r =. 45, r =. 51, indicating moderate correlations, as shown in Table 3.
In this study, we evaluated whether dominance of the native language due to the language familiarity effect, a finding observed in conventional oral serial recall, is also present in the three tests administered in the present study.
In all three tests used in this study (i.e., WST and forward- and backward-order DSTs), the dominance of the native language (i.e., Japanese) was observed, indicating a language familiarity effect. Given that this effect was observed in both the forward-order DST in the HUCRoW (i.e., serial reconstruction) and the conventional serial recall, our results imply that similar results were obtained with both tests. The observed correlations between Japanese and English versions of the conventional oral measures and HUCRoW confirm the validity of the latter (Table 2). Furthermore, the effect sizes were larger for serial reconstruction than for the conventional serial recall, indicating that the native language has greater predominance when evaluated by HUCRoW compared to the conventional oral method.
Next, we present the characteristics of serial reconstruction and serial recall. Verbal serial recall requires the verbal reproduction of spoken information immediately after it is presented. Conversely, serial reconstruction requires the participants to retain the presented speech and respond to the questions by matching or selecting options from among multiple choices of textual information presented on the PC screen, thereby converting the retained speech information into textual information. Although oral serial recall involves only the speech modality, serial reconstruction of the HUCRoW involves the selection of the information input by speech from a choice of textual information, necessitating the conversion of speech information into textual information. In both cases, the phonetic information is stored temporarily within the phonological loop, followed by its articulation and restoration by articulatory control in a rehearsal process. Conversely, certain differences may exist in the rehearsal process of restoring articulatory control between oral serial recall and serial reconstruction. In serial recall, the input speech information is directly synthesized and answered, whereas in serial reconstruction, the input speech information is converted into a visual code to select the corresponding characters. Therefore, answering a question is more complicated and not fully automated during serial reconstruction, as compared to conventional oral recall, implying that the former leads to a higher cognitive load compared to the latter (Yuzawa et al., 2019).
In this study, it is revealed that English as a foreign language requires a greater cognitive load
than Japanese as a native language in terms of the language familiarity effect, although English numbers are more familiar stimuli than English words. The load due to foreign languages may be
higher when the response method involves the conversion of the modality from speech to writing, as in the present serial reconstruction of the HUCRoW.
In the present study, using the original serial reconstruction method of HUCRoW, the dominance of the native language was confirmed in verbal STM tasks (i.e., WST and forward-order DST) and the verbal WM task (i.e., backward-order DST), which use simple familiar stimuli (i.e., words and numbers), as well as in conventional oral serial recall tasks. The present study confirmed that the phonological loop and central executive can be evaluated in group settings using HUCRoW.
Differences in processing between the conventional serial recall and the present serial reconstruction are based on the language activities in classrooms, which may have pedagogical implications.
Serial recall involves language activities based on the repetition of phonetic information, such as words and numbers. Conversely, serial reconstruction involves temporarily retaining phonetic information and converting it into textual information to answer a question, which is based on cognitive activity that is commonly used during learning situations. For example, similar cognitive processes to those of serial reconstruction are used when an instructor explains something verbally and the learner transcribes it into a notebook or when the instructor or the learner reads aloud from a textbook (i.e., audio information) and matches it with the textual information in the textbook. In this regard, the STM task and WM task of serial reconstruction more accurately simulate the cognitive processes underlying common learning activities, compared to serial recall.
Finally, two issues need to be addressed by future studies. First, the present study did not directly compare the serial recall and serial reconstruction tasks in terms of WST and backward-order DST. Furthermore, the listening span test, another verbal WM task, was included in the English version of the HUCRoW. In future, it is essential to examine whether similar results can be obtained for tests administered using HUCRoW. Second, with regard to the English version of the STM task and WM task, individuals who demonstrate similar performances on both versions need to be evaluated further, to determine the relationship between cognitive ability in the two languages and English proficiency.
This research was supported by JSPS KAKENHI Grant Numbers JP17H02356 and JP22H00675. In developing the English version of WST and two DSTs, many significant opinions on the validity of the Japanese version were offered by Professor Satoru Saito from
Kyoto University, a member of this research project. In addition, many significant comments on this research were provided by elementary school teachers. Additionally, pupils and educators (a principal and the teacher in charge of the English lessons) also cooperated in this research. We are very grateful for the involvement of the professor, teachers, students, and pupils in this research. Furthermore, this study has been presented as an oral presentation at the 23rd Japan Association for English Education in Elementary Schools (JES) Kinki/Kyoto Conference held in 2023. We thank the participants for their useful comments, which are noted here.
The authors declare no conflicts of interest associated with this manuscript. In addition, the authors have no conflicts of interest directly relevant to the content of this article.
Alloway, T. P. (2007). Automated Working Memory Assessment. Psychological Corporation.
Alloway, T. P., Gathercole, S. E., & Kirkwood, H. J. (2008). Working memory rating scale (WMRS). Pearson.
Baddeley, A. D. (2003). Working memory and language: An overview. Journal of Communication Disorders, 36(3), 189–208. https://doi.org/10.1016/S0021-9924(03)00019-4
Baddeley, A. D., & Hitch, G. J. (1974). Working Memory. In G. A. Bower (Ed.), Recent advances in learning and motivation. Vol. 8, pp. 47-89. New York: Academic Press. http://dx.doi.org/10.1016/s0079-7421(08)60452-1
Chincotta, D., & Hoosain, R. (1995). Reading rate, articulatory suppression and bilingual digit span. European Journal of Cognitive Psychology, 7(2), 201–211. https://doi:10.1080/09541449508403100
Chincotta, D., & Underwood, G. (1996). Mother tongue, language of schooling and bilingual digit span.British Journal of Psychology, 87, 193–208.
https://doi:10.1111/j.2044-8295.1996.tb02585.x
Chincotta, D., & Underwood, G. (1997). Bilingual memory span advantage for Arabic numerals over digit words. British Journal of Psychology, 88, 295–310.
https://doi:10.1111/j.2044-8295.1997.tb02636.x
Cowan, N. (2013). Working memory and attention in language use. In J. Guendouzi, F. Loncke and M. J. Williams (eds.) The Handbook of Psycholinguistic and Cognitive Processes: Perspectives in Communication Disorders (pp. 75–97). Psychology Press.
da Costa Pinto. A. (1991). Reading rates and digit span in bilinguals: The superiority of mother tongue. International Journal of Psychology, 26(4), 471–483. https://doi.org/10.1080/00207599108247135
Friedman, N. P., & Miyake, A.2005Comparison of four scoring methods for the reading span test. Behavior Research Methods, 37(4), 581–590. https://link.springer.com/article/10.3758/BF03192728
Gathercole, S. E., & Alloway, T. P. (2008). Working memory and learning: A practical guide for teachers. Sage Publications
Gathercole, S. E., & Baddeley, A. D.1993Working memory and language. Lawrence Erlbaum Associates.
Juffs, M., & Harrington, M. (2011). Aspects of working memory in second language learning.
Language Teaching, 44(2), 137–166. https://doi.org/10.1017/S0261444810000509
Linck, J. A., Osthus, P., Koeth, J. T., & Bunting, M. F. (2014). Working memory and second language comprehension and production: A meta-analysis. Psychonomic Bulletin & Review, 21(4), 861–883. https://doi.org/10.3758/s13423-013-0565-20
Logie, R. H. (1995). Visuo-spatial working memory. Lawrence Erlbaum Associates. Pickering, S. J. (2006). Working Memory and Education. Burlington, Academic Press.
Sakuma, Y. (2018, August). Sho chu gakusei no working memory kino no sokutei: susho hani kadai (junsho) to hitango hanpuku kadai [Measuring working memory function in elementary and junior high school students: Forward-order digit span task and a nonword repetition task]. In Sakuma, Y (Chair), Working memory kino ni miru sho chu gakusei no ninchiteki tokucho [Cognitive characteristics of elementary and junior high school students in terms of working memory function] The 44th Kyoto Research Conference of the Japan Society of English Language Education, Forum on Research Subjects (2nd year) conducted at the Ryukoku University, Ohmiya campus, Kyoto.
Sakuma, Y., & Saito, S. (2012). The positive influence of English-language activities on English digit-span performance among Japanese elementary school children: A three-year cross-sequential study. Psychologia: An International Journal of Psychological Sciences, 55(4), 257-268. https://doi.org/10.2117/psysoc.2012.257
Thorn, A. E., & Gathercole, S. E. (1999). Language-specific knowledge and short-term memory in bilingual and non-bilingual children. The Quarterly Journal of Experimental Psychology, 52A (2), 303–324. https://doi:10.1080/ 713755823
Thorn, A. S. C., & Gathercole, S. E. (2001). Language differences in verbal short-term memory do not exclusively originate in the process of subvocal rehearsal. Psychonomic Bulletin & Review, 8(2), 357–364. https://doi:10.3758/ BF03196173
Thorn, A. S. C., Gathercole, S. E., & Frankish, C. R. (2002). Language familiarity effects in short-term memory: The role of output delay and long-term knowledge. Quarterly Journal of Experimental Psychology Section A, 55(4), 1363–1383. https://doi:10.1080/02724980244000198
Wen, Z. (2012). Working memory and second language learning. International Journal of Applied Linguistics, 22(1), 1–22. https://doi.org/10.1111/j.1473-4192.2011.00290.x
Wen, Z. (2014). Theorizing and measuring working memory in first and second language research.
Language Teaching, 47(2), 174–190. https://doi.org/10.1017/S0261444813000517
Williams, J. N. (2012). Working memory and SLA. In S. Gass & A. Mackey (Eds.), The Routledge Handbook of Second Language Acquisition. (pp. 427–441). Routledge.
Yuzawa, M., Kuranaga, H., Saito, S., Minakuchi, K., Watanabe, D., & Morita, A. (2019). Jido seito you shudanshiki working memory asesumento testo no sakusei [Development of a Working Memory Assessment Test for Children in a Group]. The Japanese Journal of Developmental Psychology, 30(4)253–265https://doi.org/10.11201/jjdp.30.253
The Test Progress Sheet
フリガナ
児童番号 氏名
生年月日（西暦）ʁ 年 月 日生まれ
＊各ゲームが終了したら，手を挙げて下さい。各欄にチェック（ V）を入れましょう。
|
|
|
||||
|---|---|---|---|---|---|
|
|
|
|
|
|
|
Note. Game No.1: Forward-order DST. Game No. 5: WST. Game No.7: Backward-order DST.
