2.1 Influential societal and global factors

English teaching and learning among young learners is situated in larger societal and global contexts. Although each of the focal countries in this review has unique factors influencing their early English language education, two societal factors are worth mentioning because they are influential across East Asia.

First, even though English has been taught as an FL for a number of years in formal schooling (at the secondary school level and beyond), it has never been a major means of communication in the given communities. While it is undeniable that the use of English is on the rise in certain areas of East Asian societies, such as in media and business (Feng Reference Feng2012), most people have limited opportunities to use English in their daily lives. For example, Song (Reference Song2011) described the situation in South Korea as ‘impersonal bilingualism’ and argued that a seeming increase in the use of English among Koreans is ‘for appealing to the public's positive feelings, not for practical communication’ (p. 37). Given this situation, it is generally hard for learners to receive sufficient input to acquire practical communicative competency in English. For learners who want more opportunities to use English, extra money and resources are often required. This condition partially contributes to widening gaps in accessibility and achievement in English by SES and region (which is also usually associated with SES).

Second, in East Asia, exams have been institutionalized in many aspects of society and play a significant role in these countries’ meritocratic policies and attitudes (e.g., Butler & Iino Reference Butler and Iino2005; L. Y. Cheng & Curtis Reference Cheng and Curtis2009; Carless Reference Carless2011). Under a prevailing belief system in which a person's achievement is primarily attributed to effort rather than innate talent (Stevenson & Stigler Reference Stevenson and Stigler1992), exam results are often seen as a sign of effort and diligence and, as such, are greatly valued. Exams are considered a vehicle for obtaining success and status. With the rise of English as an academic subject that can influence future educational and career opportunities, access to early English education has gained tremendous public attention in East Asia.

Such societal factors are embedded in larger global contexts. Since the 1980s, English, as a global language of political and economic power, has frequently been recognized as a critical competence for nations’ survival and success. Around the turn of the twenty-first century in both in Korea and Japan, proposals were made to designate English as an official language along with the national language; these proposals created heated debates in both countries. It is beyond the scope of this article to review different explanations of the spread of English in Asia, and why the English-learning boom—which is often described as ‘English fever’ (Park Reference Park2009: 50) or an ‘obsession’ (J.-J. Song Reference Song2011: 36)—developed. However, it is worth noting that the intensity of the fever has varied. For example, the ‘obsession’ seems to be much less intense in Japan than in Korea (J.-J. Song Reference Song2011).

Individuals and organizations invest tremendous amounts of money in private entities that provide English-learning services. There is no doubt that ‘English proficiency has become a very expensive commodity’ (Hu Reference Hu2009: 50). In their formal education systems, East Asian nations began to actively reform English teaching around the turn of the 21st century. Perhaps the biggest reform was introducing English as a foreign language at the primary school level (referred to as EFLPS hereafter).

2.2 Language-in-education policies at the primary school level

Kaplan & Baldauf (Reference Kaplan, Baldauf and Hinkel2005) proposed a model for language-in-education policy planning and argued that the following interrelated elements influence the success of language policies: curriculum policy; access policy; personnel policy; methodology and material policy; resources policy; community policy; and evaluation policy. Following Kaplan & Baldauf's framework, Table 1 summarizes the EFLPS policies (as of 2013) in the four focus countries. In the following sections, I discuss these major policy issues as they relate to early English education in East Asia.

Table 1

A policy sketch (as of 2013)^a

Notes: ^aThis table is based on Baldauf et al. (Reference Baldauf, Kaplan, Kamwangamaly and Bryant2011, Reference Baldauf, Kaplan, Kamwangamaly and Bryant2012); Butler (Reference Butler2004, Reference Butler2007a); Chen (Reference Chen2011); Hashimoto (Reference Hashimoto2011); Joo (Reference Joo2008); Kang (Reference Kang, Spolsky and Moon2012); Li (Reference Li2011); Shiga (Reference Shiga2008); Wu (Reference Wu, Spolsky and Moon2012); Zhang & Wang (Reference Zhang, Wang and Feng2011).

^bA minimum of 80–90 minutes total; exact durations are not specified and, in practice, vary.

^c‘Homeroom teachers’ are generalists and not specialist teachers of English.

^d‘English villages’ are English-speaking environments in Korea suitable for immersive study.

3.1 Research examining language development

3.2 Research examining affective/attitudinal domains

The research on affective/attitudinal domains, and motivation in particular, among young EFL learners in Asia has received increasing attention. This growing popularity is partially due to the fact that numerous anecdotes from teachers indicate that children gradually lose their motivation for learning English during the primary school years. So far, however, most studies in East Asia have examined whether previous findings among older learners or children in non-Asian contexts such as Europe can be applied to East Asia. Thus, a number of problems that have been identified in the existing motivation literature, such as inconsistent use of constructs for motivation as well as definitions of the term itself (Murphy & Alexander Reference Murphy and Alexander2000), also mar the studies that I reviewed.

Carreira (Reference Carreira2006, Reference Carreira2011, Reference Carreira2012a, Reference Carreira, Muller, Herder, Adamson and Broan2012b) conducted a series of studies on motivation, using different sets of cross-sectional survey data from Japan. Her primary interest was to test if predictions in self-determination theory (SDT) (Deci & Ryan Reference Deci and Ryan1985) are applicable to primary school students in Japan who are learning English. SDT classifies motivations according to the degree of autonomy, ranging from a higher degree of external control (a higher form of extrinsic motivation) to a higher degree of self-regulation (a higher form of intrinsic motivation). SDT concerns how a person's motivation, which may be driven originally by external control, can be made more self-regulated through the use of social supports that meet his or her psychological needs, namely autonomy, competence and relatedness. Carreira (Reference Carreira2006) found that Japanese children's intrinsic and extrinsic motivation declined as the grade level increased. Carreira (Reference Carreira2011) reported that not only Japanese students’ motivation for learning English but also their intrinsic motivation for studying in general declined as the grade level increased. Carreira's (Reference Carreira2012a) exploratory factor analysis showed that two of the adjacent extrinsic motivational constructs predicted by SDT (i.e., introjected and identified regulations) were undifferentiated among her primary school students, but her correlational analysis confirmed the predicted order of constructs in the self-regulation continuum. Carreira (Reference Carreira2012a) found that the three psychological needs (i.e., autonomy, competence, and relatedness) were more strongly associated with intrinsic motivation than extrinsic motivations, also as predicted by SDT. Finally, Carreira (Reference Carreira, Muller, Herder, Adamson and Broan2012b) investigated how instrumental motivation (considered a type of extrinsic motivation) and two other ‘affective factors’ (p. 197)—interests in foreign countries (which is considered an integrative motivation based on Gardner's socio-educational model of motivation, Gardner Reference Gardner1988) and children's perceived parental involvement—were associated with intrinsic motivation. The results showed that although integrative motivation strongly related to intrinsic motivation, children's perceived parental involvement did not.

Some authors tried to identify factors contributing to children's willingness to communicate, defined as ‘the intention to initiate communication, given a choice’ (MacIntyre et al. Reference MacIntyre, Baker, Clément and Conrod2001: 369). Adachi (Reference Adachi2011, Reference Adachi2012) conducted survey analyses, claiming that Japanese children's ‘attitudes towards learning’ positively influenced their willingness to communicate as well as their motivation. In both of her studies, attitudes towards learning was composed of ‘confidence in English, belief in effort, positive attitude towards learning general subjects, positive attitudes towards multiple languages, and understanding of ALT’ (Adachi Reference Adachi2011: 9).Footnote ⁷ Nishida & Yashima (Reference Nishida and Yashima2009) found that Japanese children's perceived classroom atmosphere influenced their perceived competence, which in turn led to their willingness to communicate. A similar finding was reported in Kunimoto (Reference Kunimoto2006) in Japan as well. Nishida (Reference Nishida2012) followed a group of Japanese students for one year and reported that their willingness to communicate decreased over time, as did their motivation, perceived competence, and interests in language. Instead of willingness to communicate, Nakata (Reference Nakata2009) included willingness to continue learning as one of the components for intrinsic motivation along with two other components (enjoyment and meaningfulness). He asked Japanese junior high school students to retrospectively evaluate their motivation in English classes at each grade level during the primary school years; he found complex, ‘intertwined’ (p. 263) relations among the three components.

Jin, Liang, Jiang, Zhang, Yuan & Xie (Reference Jin, Liang, Jiang, Zhang, Yuan and Xie2014) elicited metaphors related to English learning from Chinese children (Grades 1 and 3, aged 7 and 9, respectively) and interpreted their reasoning about what the metaphors entailed. The researchers found that while children in Grade 1 used positive metaphors to describe their English learning and their achievements, some children in Grade 3 started using negative metaphors to describe the difficulties they encountered in English learning.

T.-Y. Kim (Reference Kim2011) and Kim & Seo (Reference Kim and Seo2012) sought to understand Korean children's demotivation. Kim (Reference Kim2011), based on a cross-sectional survey data, found that Korean primary school students increasingly showed dissatisfaction with English learning and lowered their motivation as the grade level increased. In Kim & Seo (Reference Kim and Seo2012), which employed a mixed method (adding teachers’ interviews to the survey), children who experienced private lessons had higher motivation than those who did not. The teachers attributed their students’ demotivation for learning English to negative influences from teachers’ practices and attitudes, excessive societal pressures, and widening achievement gaps among students.

In Taiwan, Chan & Wu (Reference Chan and Wu2004) investigated fifth graders’ anxiety based on surveys (including the Foreign Language Classroom Anxiety Scale [FLCAS], which has been widely used for adults), interviews with select students and teachers, and classroom observations. Types of anxiety identified by the authors included pressure from testing, feelings of incompetence, and fear of speaking. The major sources of these anxieties were lower proficiency levels, fear of receiving a negative evaluation, having an anxious personality, competition created by games in class, and pressure caused by their own and their parents’ expectations.

As we have seen, most of these studies were based on surveys. We need to be cautious when administering surveys to children; the items used in the studies above were often decontextualized and potentially confusing for children. They sometimes contained dichotomous types (items that force children to respond to one answer over the other), which also potentially contain a lot of noise in their responses. Additionally, although it makes sense to give children as few items as possible, some of the studies reviewed above used surveys with only one or two items per construct, which may make the constructs unreliable psychometrically. Yim & Yu (Reference Yim and Yu2011) proposed using the English Learning Anxiety Scale, which is a revised and shorter version of the FLCAS mentioned above, and they conducted a validity check among fourth to sixth grade students in Korea. Conducting such a validation test is important before administering a survey to children. It is also important to question whether the nature of motivation remains the same across age groups and contexts. The ‘same’ response to an item may mean something different depending on a student's age. Indeed, Butler's (Reference Butler2015a) interviews with primary school students in China indicated that the nature of both intrinsic and extrinsic motivation differed across different grade levels and the students’ SES. Given the complex nature of motivation constructs, we may not be able to expect motivation to change linearly in any consistent fashion; linear statistical models need to be employed with care. As has been suggested in the general motivation literature, we need more contextualized approaches to understand motivation. Finally, because declining motivation among primary school students seems to prevail, we need more studies on the mechanism of the demotivation process, as attempted in Kim & Seo (Reference Kim and Seo2012), as well as more research on instructional strategies for preventing children from losing motivation.

5.1 NNS teachers’ qualifications

Among different qualifications considered necessary for a good language teacher, NNS teachers’ English proficiency has been a leading concern across East Asia. However, we have little empirical data about their proficiency or how their proficiency influences children's English learning. Butler (Reference Butler2004), using the Stanford Foreign Language Oral Skills Evaluation Matrix (FLOSEM), asked primary school NNS teachers in Korea, Japan, and Taiwan to self-rate their current English proficiency as well as to describe the minimum level that they thought necessary to teach English under the current policies. The results indicated that 91.1%, 85.3%, and 80.1% of NNS teachers in Korea, Japan, and Taiwan, respectively, believed that their current proficiency levels fell short of the minimum levels required, particularly in oral productive domains. I found no published studies of NNS teachers’ proficiency using objective measures.

In a 2008 case study, Chou used interviews and classroom observations to research three Taiwanese NNS teachers’ knowledge. The study confirmed previous findings that ‘these teachers’ practical knowledge [was] formulated through a process of reshaping their existing knowledge of English teaching and learning from training programs and their classroom practices rather than just having language learning theories, methods, or materials imposed’ (p. 539). In China, the policy encourages teachers to conduct action research as a means of evaluating their performance. Through a survey and interviews with Chinese NNS teachers in an in-service teacher training program, Gao and his colleagues (Gao, Barkhuizen & Chow Reference Gao, Barkhuizen and Chow2010; Gao & Chow Reference Gao and Chow2011) revealed a number of contextual constraints (e.g., lack of funding, appropriate training, and resource to conduct research) that made teachers’ research efforts less effective for their own professional development.

5.2 The role of NS teachers

Since implementing EFLPS, East Asian countries have recruited NS teachers to serve in various capacities; however, their roles in the English education systems are often controversial.

Among the general public in East Asia, there is a widespread belief that native speakers are ideal language teachers. In Japan, Butler (Reference Butler2007b) found that NNS teachers with lower perceived English proficiency, with more negative attitudes toward non-standard forms of English, and with a higher sense of pride in their own language and culture (i.e., Japanese language and culture) tended to support the notion that native speakers are ideal language teachers.

NS teachers often co-teach classes with NNS teachers, and several studies examined challenges that both types of teachers faced in team-teaching, or what Carless (Reference Carless2006a) called ‘intercultural team-teaching’ (p. 341). Luo (Reference Luo2007) and Chen & Cheng (Reference Chen and Cheng2010) are case studies that each focused on three NS teachers in Taiwan, using interviews and classroom observations as well as insights from the NS teachers’ NNS colleagues. In Luo, the NS teachers thought that major obstacles for team-teaching included NNS teachers’ low English proficiency; NNS teachers’ unwillingness to try new teaching approaches; and large class sizes. The NNS teachers saw as major problems the NS teachers’ lack of pedagogical context knowledge and their inflexibility when dealing with local practices. Chen & Cheng also noted that NNS teachers were unhappy about some of their NS colleagues’ non-standard English accents (e.g., South African English accents). Carless's (Reference Carless2006a) case study focused on three NS teachers with ‘good practice’ (one from Korea, one from Japan, and one from Hong Kong) and identified a few features for successful team-teaching with NNS teachers. Such features included being sensitive to interpersonal relationships with local participants; developing relationships inside and outside of class; letting minor obstacles go; being respectful of established local practices; and having a long-term commitment. Carless (Reference Carless2006b) evaluated team-teaching in Hong Kong in light of its three policy objectives: ‘impact on pupils, innovative teaching, and professional development’ (p. 328). While the teachers’ experiences were positive in general, they thought that team-teaching had a positive influence on the first objective but less of a positive influence on the other two objectives.

While the studies described above examined teachers’ perceptions, Oga-Baldwin & Nakata (Reference Oga-Baldwin and Nakata2013) asked fifth-grade students about different combinations of team-teaching among NS teachers, NNS Japanese homeroom teachers, and NNS Japanese teachers of English. The students reported that they received input from NS teacher-led classes most; however, the students reported little difference in their perceived influence by both NS teachers and NNS Japanese teachers of English. The students evaluated the team-teaching led by NNS homeroom teachers as most helpful when it came to speaking. It would be interesting to see if the same result would be obtained after controlling for the NNS teachers’ English proficiency and the frequency of different types of team-teaching that the students experienced.

Butler (Reference Butler2007c) examined how NS and NNS teachers’ accented English influenced the listening comprehension of Korean primary school students. Butler used a matched-guise technique in which a Korean-American individual recorded a story both in American-accented English and Korean-accented English. The students were tested on their comprehension of the story recorded in the different accents and were asked their perceptions of the two accented guises. (The students did not know that both recordings came from the same person.) The study found that the students thought that the American-accented English guise had better pronunciation, was more confident, would focus on fluency (rather than accuracy) in class, and would use less Korean in class; the students showed a preference for the American-accented guise. However, the accents did not influence the students’ listening comprehension, contrary to a general concern about NNS teachers’ accented-English, at least as far as listening goes.

In sum, to better inform policy decisions, we need more empirical studies that can help identify teachers’ knowledge and skills as they relate to teaching EFLPS. What types of practical knowledge and skills are most helpful for teaching young learners? When it comes to English proficiency, we need guidelines that suggest necessary or optimal proficiency levels for teachers depending on their teaching objectives. Such guidelines would provide teachers with sufficient assistance to improve their proficiency and bolster their confidence. Many studies reviewed above tried to identify distinctive features of NS and NNS teachers. Instead, these studies revealed that teachers’ perceptions heavily depended on individual teachers’ specific experiences of and views about language teaching. We probably need to focus more on what kinds of qualifications are necessary regardless of NS or NNS status, and how best to train FL language teachers or to assist their professional development in the specific East Asian educational systems in which they teach.

6.1 Communicative activities and tasks

By conducting a series of classroom observations and interviews with teachers, Carless (Reference Carless2004b, Reference Carless2008) and Deng & Carless (Reference Deng and Carless2009, Reference Deng and Carless2010) revealed that many activities introduced in primary schools in Hong Kong and Guangdong were not ‘communicative’ in nature, and that teachers’ beliefs about CLT as well as various contextual factors made it difficult to implement communicative activities in class. Zhang & Adamson's (Reference Zhang, Adamson and Feng2007) three teachers in Shenzhen also exhibited varying beliefs about TBLT. Major challenges with CLT/TBLT identified by these studies included the students’ excessive use of their L1, difficulties with classroom management, the quality of target language production by the students, and exam pressures. Taiwanese teachers in Su (Reference Su2006) also believed that skill-based components (e.g., drills and exercises) were necessary even in communicative activities. Peng & Zhang (Reference Peng, Zhang and Nikolov2009) reported that Chinese Grade 5 students still heavily depended on L1-based communication strategies (e.g., used L1 when not knowing how to say something in L2) in supposedly CLT classes. Using a sociocultural framework, Butler (Reference Butler2005) focused on concerns with CLT among teachers in Japan, Korea, and Taiwan with respect to ‘(1) creating motives and goals that drive communicative activities; (2) identifying developmentally appropriate mediational means; and (3) situating activities in specific contexts’ (p. 423). By employing multivocal ethnography (Tobin, Wu & Davidson Reference Tobin, Wu and Davidson1989), which allowed the researcher to incorporate multiple people's perspectives in addition to her own in interpreting focus teachers’ classroom activities, Butler found that the teachers’ difficulties with implementing communicative activities were largely due to their lack of understanding of what constitutes ‘teaching for communicative purposes’, the roles that developmental factors play in EFL learning and teaching, and strategies for harmonizing learning/teaching in context’ (p. 423).

Teaching English through the medium of English only (TETE) also has been promoted by curriculum and policies in East Asia, and several studies examined the use of L1 and the target language (English) by teachers and students in class. Carless (Reference Carless2004a), a case study of a primary school teacher in Hong Kong, indicated that the teacher's experience of and beliefs about TETE, as well as her proficiency in English, largely determined the way that she used English in class. The same conclusion was drawn from Kang's (Reference Kang2008) case study of one teacher in Korea. Qian, Tian & Wang (Reference Qian, Tian and Wang2009) performed a longitudinal (for four years) study of two Chinese teachers’ code-switching between their L1 and the target language (English). They found that inter-sentential switching, as opposed to tag or intrasentential switching, was used by the teachers much more frequently, and that the code-switching was intentionally used for clarifying instructions and eliciting responses from students. The researchers also found that the frequency of code-switching by the teachers drastically decreased from the first to the second year and beyond. Comparing adults’ and young learners’ perceptions of TETE, Macaro & Lee (Reference Macaro and Lee2013) found that the sixth graders showed less favorable attitudes toward English-only instruction than the adults. We suspect, however, that their responses were greatly influenced by the instruction and teachers that they had had as well as the goals of instruction.

Based on sociocultural theories, some researchers examined how knowledge was constructed during activities and tasks. Kim & Kellogg (Reference Kim and Kellogg2007) compared Korean primary school students’ language production during role-plays and rule-based games, and discussed the significance of these plays in language development. Based on their observations that role-plays had more complex grammar whereas rule-based games produced more complex discourse, the researchers argued that the latter was a precursor to the former, and that the transition from inter-mental discourse complexity observed in the rule-based games to intra-mental grammar complexity observed in role-plays was precisely what Vygotsky (Reference Vygotsky1978) called a zone of proximal development (ZPD). Yi & Kellogg (Reference Yi and Kellogg2006) examined Korean primary school students’ reported speech in their English writing and saw ‘a progression from inter-sentential cohesive links and simple sentences suggestive of other mediation and spoken dialogue to intra-sentential links that include subordination more typical of self-mediation and written language’ (p. 50). Importantly, according to the authors, this process is not simply a development of abstract knowledge but also indicates the children's greater awareness of audience. Nishida & Yashima (Reference Nishida and Yashima2010) examined Japanese children's interaction during a musical project and reported that peer scaffolding emerged as the frequency of teacher scaffolding decreased over time.

6.2 Specific instructional strategies and interventions

A number of studies examined the effectiveness of different instructional methods and strategies. Shintani & Ellis (Reference Shintani and Ellis2010) compared the effectiveness of comprehension-based and production-based instructions on Japanese children's (6–8-year-olds) incidental acquisition of plural –s. Incidental learning is defined as learning without intention but not necessarily without awareness (i.e., implicit learning). Interestingly, the study failed to find a significant difference between the two types of instructions when it came to student comprehension, but students receiving the comprehension-based instruction outperformed in production. It turned out that the nature of corrective feedback that the students received differed. The students receiving comprehension-based instruction received feedback on their comprehension errors, which appeared to help them acquire the meaning of singular and plural. In contrast, the students in the production-based instruction received recasts (rephrasing without changing meaning) on their production errors, which did not lead to corrections of the errors. Similarly, Shintani (Reference Shintani2011, Reference Shintani2013) compared the effectiveness of comprehension-based instruction and production-based instruction on English vocabulary learning among Japanese children (6–8-year-olds). While both types of instruction were effective in the children's noun production, the comprehension-based instruction was more effective in their adjective production. Shintani, after analyzing the interaction between the teachers and the children during instruction, suggested that the comprehension-based instruction, which offered the children more opportunities for contextualized input, negotiations of meaning, and student-initiated production, is better suited for children to develop knowledge for using adjectives. Rhee (Reference Rhee2012) examined the effect of recasts when they were combined with explicit form-focused instruction on the acquisition of English past-tense skills among Korean fifth graders. Not too surprisingly, in both oral and written production tests, the students who received both recasts and form-focused instruction performed better than the students in the form-focused-only group and in the control group. The effect was larger among students with higher proficiency, suggesting the importance of readiness for intake.

Jeon & Shin (Reference Jeon and Shin2011) and Y.-H. Lee & H.-K. Lee (Reference Lee and Lee2012) examined the effects of different explicit vocabulary instructional strategies among Korean students. In Jeon & Shin, Grade 3 students who had been given a chance to spell out words showed higher retention not only in an immediate test but also in a delayed test (conducted two weeks after the treatment) compared to students who learned the words from the list receptively. Lee & Lee examined the effects of three input enhancement techniques: visual enhancement; semantic enhancement; and input flooding (unfortunately, the details of these treatments were not reported). At an immediate recall, the visual enhancement and input flooding techniques facilitated the children's recall of the forms of the target words, whereas the semantic enhancement facilitated recall of the meanings of the words. However, all of the differentiated effects disappeared within a week.

As a literacy instructional method, storytelling and extensive reading are increasingly popular in EFLPS in East Asia. Cho & Choi (Reference Cho and Choi2008) examined the effect of the combination of listening to stories (read by the teacher) and free reading (self-selected reading) among Korean sixth graders. In addition to listening to stories for 10 to 15 minutes per week for 21 weeks, the students in the treatment group engaged in one hour of free reading per week. The results indicated that the students in the treatment group outperformed in English tests (listening, reading, and writing), showed higher interest and confidence in reading, and had lower anxiety. A positive result was also reported in Tsou (Reference Tsou2011), who examined the effectiveness of Readers Theater, a read-aloud activity, in reading and writing among Taiwanese fifth graders. To identify successful factors for extensive reading implemented in Korean primary schools, Kim (Reference Kim2013) conducted a meta-analysis. Kim found that extensive reading was effective both cognitively and affectively at both lower and upper grade levels. The effect was bigger at private schools than public schools; the biggest effect was found in private schools in economically wealthier areas. This result suggests that external support by teachers and schools is important, in particular for children who are economically less privileged.

In Butler & Okazawa (Reference Butler, Okazawa, Ogura and Kobayashi2009), Japanese teachers who participated in a year-long, extensive reading program for Grade 6 students reported a number of instructional supports that they found effective for their students early in the year. Strategies included giving tips for choosing books, reading aloud to students individually, and asking guiding questions to help them follow story lines. Butler & Okazawa also shed light on how aspects of extensive reading were a social activity in addition to being an individual activity. For example, after trying extensive reading for one semester, the students collectively decided not to switch books for at least ten minutes so that they could better engage in stories. The students also voluntarily created a space where they could talk about stories with their peers.

Content-based instruction (CBI, or content and language integrated learning, CLIL), in which parts of academic subjects are taught in the target language, is increasingly popular at EFLPS. Using an adapted version of Communicative Orientation of Language Teaching (COLT), a classroom observational scheme, in combination with qualitative analyses of classroom observations, Huang (Reference Huang2011) compared classroom interaction in CBI sessions and form-focused EFL sessions (e.g., phonics and spelling instructions) among the same group of students in Taiwan. The result indicated that the students showed more active verbal participation and produced more cognitively complex sentences during the CBI sessions. According to Huang, this was because CBI, a type of meaning-focused instruction, created more opportunities for children to engage in meaning negotiation in class.

6.3 Use of technology in instruction

Various types of technologies have been used to educate young learners of English in East Asia. In Korea, Joo (Reference Joo2008) and Im (Reference Im2009) both surveyed students and teachers, asking them their opinions of a multimedia electronic English textbook (Joo) and CD-ROMs (Im), both of which were developed by the government. Joo found that both students and teachers, particularly those from rural areas, viewed the multimedia electronic textbook favorably. According to Im, the teachers depended on the CD-ROMs heavily when demonstrating conversations and songs in class; however, the students rarely used the CD-ROMs at home, suggesting that the content should be made more attractive for students.

A couple of intervention studies examined the effectiveness of new instructional technology. Based on their preliminary observation of a lack of collaboration among students in a Taiwanese EFLPS class, Lan, Sung & Chang (Reference Lan, Sung and Chang2007) developed ‘a mobile-device-supported peer-assisted learning system’ (MPAL) that was designed to promote peer collaboration through online reading partnerships and peer assessments. The researchers found that MPAL not only facilitated collaboration among students in class but also increased their motivation for reading. Lai, Tsai & Yu (Reference Lai, Tsai and Yu2009) developed a multimedia English learning (MEL) system to help children develop phonemic awareness and pronunciation by providing instant feedback. They tested the MEL system with Taiwanese students (aged 9–10) for 12 weeks and found that it improved the students’ phonemic awareness, spelling, and reading abilities. It was particularly effective among children who had lower phonemic awareness before the intervention. Suh, Kim & Kim (Reference Suh, Kim and Kim2010) examined the effectiveness of a ‘massive multiplayer online role-playing game’ (MMORPG) implemented in five schools in Korea. The students in the treatment group outperformed their peers in English listening, reading, and writing tests but not in a speaking test. The students’ prior knowledge (unfortunately, how it was defined and measured is unclear), motivation for learning, and network speed also influenced the game's effectiveness.

Butler, Someya & Fukuhara (Reference Butler, Someya and Fukuhara2014) examined Japanese children's (aged 4–12) use of online instructional games as part of a learning tool for a set of online assessments. The study identified game features shared by attractive games (games played by the children frequently); such game features included requiring optimal cognitive demand, having elements that can evoke children's curiosity, allowing greater player control and autonomy, and having multiple players to stimulate competition. Different game-playing behaviors were found by age but not by gender. The study also showed varying relational patterns between the children's game scores and their English performance, depending on which games they played and the difficulty level of the assessments.

8.1 Challenges with assessment for learning and age-appropriate assessment

East Asian governments promote the use of assessment for learning in EFLPS, but implementing such assessment does not seem to be easy. According to survey studies among teachers in Korea (Hwang Reference Hwang2006) and Taiwan (Chan Reference Chan2008), although the teachers still depended on traditional paper-and-pencil tests, they also claimed to use various types of classroom-based assessments. However, such assessments were not necessarily administered systematically nor used for formative purposes.

Butler (Reference Butler2009c) compared how English teachers in Korean primary and secondary schools observed and evaluated Grade 6 students’ oral performance in group activities. The author found that there was a substantial variability among teachers both within and across school levels. Major differences in views between primary and secondary school teachers included their conceptualizations of evaluation criteria, evaluations of students’ motivation and confidence, and inclusion/exclusion of students’ potential ability in their evaluations. Such differences in views appeared to be deeply rooted in the teachers’ respective teaching environment and previous assessment experiences.

In line with the recent emphasis on learner-centered language teaching and self-regulated language learning, self-assessment has been widely promoted in EFLPS. Butler & Lee (Reference Butler and Lee2006) questioned the validity of a holistic type of self-assessment that is frequently employed in EFLPS programs (referred to as off-task self-assessment) and compared it with a more contextualized self-assessment that was designed to assess one's performance in a specific task or lesson (referred to as on-task self-assessment) in Korean primary schools. They found that young learners could evaluate their performance more accurately in the on-task self-assessment condition in relation to a proficiency test and their teachers’ evaluation. The sixth graders evaluated themselves more accurately than did the fourth graders. Moreover, for students in both grades, on-task self-assessment was less influenced by personality factors compared with off-task self-assessment. Butler & Lee (Reference Butler and Lee2010) examined the effect of self-assessment among Korean students longitudinally. They found that the students became better at assessing themselves over time, and that the regular administration of self-assessments positively influenced their English learning as well as their confidence. However, the degree of such effects differed depending on the teachers’ beliefs about assessments and their teaching environments.

Several studies looked into classroom-based speaking assessments in which teachers or peers were allowed to mediate during the assessments. Lin's (Reference Lin2010) exploratory study showed that the teachers’ mediation helped children's English learning in a bilingual kindergarten in China from the point of view of their final English production, although the details of the mediation process were not reported.

Butler & Zeng (Reference Butler, Zeng, Tsagari and Csepes2011) compared the interactions between the individual assessment (student–teacher pair) format and the paired-assessment (student–student pair) format among Grade 6 students in China. The researchers found that individual assessment can be beneficial, especially for students with lower English proficiency, in that it enabled teachers to stretch their students’ abilities. However, in the individual assessment format, the teachers’ dominant and authoritative manners tended to result in simpler teacher–student interactions. On the other hand, paired assessments tended to yield a variety of communicative strategies and language use from the students as long as they could collaborate with each other; otherwise, teachers’ careful oversight was necessary. Butler (Reference Butler, Thomas and Reinders2015b), however, showed that there was substantial variability in the way that individual Chinese teachers interacted with their students during both types of assessments (i.e., individual and paired assessments) and during the mediation that they provided to the students during the assessments. As a result, the language that they elicited from the students varied as well. Such variability of interactions among teachers seemed to reflect their beliefs about assessments and the challenges that they saw in implementing assessment for learning in a context where traditional, measurement-driven exams were dominant.

Butler & Zeng (Reference Butler, Zeng, Connor-Linton and Amoroso2014b) focused on a task-based language assessment in the paired-assessment format (paired-TBLA) and investigated developmental differences in interaction between fourth and sixth graders in China. The younger student dyads, compared with the older dyads, worked less collaboratively in general, had a harder time developing mutual topics and taking the partners’ perspectives, and showed more formulaic turn-taking patterns. In order to test the age effect separately from the proficiency levels (because age and proficiency levels are often confounded in EFLPS programs), Butler & Zeng (Reference Butler and Zeng2013) examined the students’ interaction during the paired-TBLA in their L1 as well as in their FL and obtained the same result as above. Taken together, these results suggest that the potential benefits of the paired-assessment format—eliciting wider ranges of language use and interactional strategies—–may have limited applications for younger children.

Yukawa and her colleagues in Japan (Yukawa, Koyama & Takanashi Reference Yukawa, Koyama and Takanashi2010; Koyama & Yukawa Reference Koyama and Yukawa2012) developed Let's Talk, a speaking test with scaffolding in which primary school students, in pairs, converse with a foreign/native English-speaking teacher. The students were allowed to help each other and use their L1 (Japanese) if necessary during the test. The individual students’ speaking abilities were evaluated on six criteria (phonology, vocabulary, syntax, attentiveness, expressiveness, and management) using a four-level point system. Koyama & Yukawa validated Let's Talk in four different ways: intercorrelations validity; convergent validity; discriminant validity; and known-groups validity (p. 28). The results indicated that Let's Talk was reasonably high in validity.

Finally, Bae & Lee (Reference Bae and Lee2010) showed a process for validating two parallel sets of test forms that use pictures as prompts. Pictures are often used as prompts in assessments, especially for young learners, but it was not known how to ensure that any given ‘parallel’ forms are indeed parallel psychometrically. By presenting empirical evidence for the parallel nature of pictorial forms in an assessment for young learners (tested with Korean primary school students), the authors not only made the tests available as an oral assessment tool but also shared a validation method with other researchers.

8.2 Washback effects

Considering the vulnerability of children, the influence of assessments (washback effect) is a critical topic for empirical investigations. Although it is commonly believed that exam pressure prevents teachers from implementing innovative pedagogies and assessments in their classrooms in East Asia, Deng & Carless's (Reference Deng and Carless2010) case study of four primary school teachers in Hong Kong revealed that their perceived influence of exams varied greatly among the teachers, reflecting their beliefs about pedagogy and assessment as well as where the instruction took place.

Deng & Carless's finding notwithstanding, children may feel pressure starting from an early stage of their English learning. Employing group interviews and a picture-drawing technique with lower-grade primary school students in Hong Kong, Carless & Lam (Reference Carless and Lam2012) asked the children how they perceived tests and how their parents reacted to their test performances. Although good test results induced positive feelings in the children, there were ‘considerable negative student perceptions’ (p. 7) of tests, such as anxiety, unhappiness, and fearfulness. It is important to note that the parents and teachers played significant roles in children's feelings about testing—both positive and negative.

A growing number of commercial proficiency tests targeted at young learners of English have become available in East Asia in recent years, and researchers have considered the impact of the availability of these tests. Taiwanese teachers believed that the introduction of EFLPS made parents push their children to take various proficiency tests (Su Reference Su2006). Chik & Besser (Reference Chik and Besser2011) asked stakeholders (including teachers and parents) their perceptions of the use of the Cambridge tests (YLE), a popular commercial proficiency test, in order to understand its social consequence. The authors found that YLE was not used for official placement/admission purposes for secondary schools. However, at the less-privileged school that they observed, only a few chosen students were allowed to take the test because of budgetary limitations, and taking YLE was considered a means of empowering students to be admitted into English-medium secondary schools (more prestigious than Chinese-medium schools). At the more privileged school, students who felt pressure from their parents and feared being left behind by their peers took any possible tests and compared their scores with their peers. The authors argued that, despite the YLE's unofficial status within the formal educational system in Hong Kong, the test gave additional advantages to those who could afford to take it, while ‘systematically’ disadvantaging other groups (p. 88) at an early stage of their schooling.