Language Policy Documents

Many language policy documents explicitly state the advantages of early language learning. The Commission of the European Communities (2003) defined them in a broad sense (e.g., better L1 skills, and favorable attitudes to other languages, people, and cultures) and listed conditions under which they can be beneficial (e.g., trained teachers, small classes, enough time devoted to languages). As for which language to teach, the European standpoint is clear. The Council of Europe (2007) language policy document listed several reasons why it does not matter what language YLs study. A recent resolution passed by the European Parliament (2009) stated that “Europe's linguistic diversity constitutes a major cultural asset and it would be wrong for the European Union to restrict itself to a single main language” (n.p.).

The Position Statement on Teaching English as a Foreign or Additional Language to Young Learners (TESOL, 2009), for instance, emphasizes that “age alone does not determine success in learning a foreign language” and “there is no single best way to implement” (p. 1) such programs. The strengthening of YLs’ own cultural values and identities is also often stressed, as in some contexts early English is often perceived not only as a blessing, but also as a threat.

The particular age range considered young varies to a great extent. In the European context, a working group of the European Union member states agreed that pre-school children between the ages of three to six are called very young learners, whereas primary-school pupils between seven to twelve are young learners, although in certain contexts even 14-year-olds are included in the YLs’ group. Practices regarding when early language learning starts in national curricula also vary along a similar continuum.

Types of Early Programs

Much variation characterizes early language learning programs according to when they start, how much time they allocate to early language learning, what type of curriculum they apply, who the teachers are, and how the programs are implemented. A review of European countries found that time devoted to ELL ranges between about one hour (in awareness-raising and language-focused programs) and several hours per week (in content and language integrated learning [CLIL] programs; for details, see Edelenbos, Johnstone, & Kubanek, Reference Edelenbos, Johnstone and Kubanek2007; Dalton-Puffer, this volume).

Johnstone (Reference Johnstone, Enever, Moon and Raman2009a) describes four models of early language learning. In the first, some general topics are used for teaching the foreign language (e.g., colors, parts of the body), whereas in the second model the topics are borrowed from other curricular areas (e.g., animals and their habitat). In both models about one hour per week is devoted to language learning. The third approach seeks to sensitize children to languages, often those in the local community. Finally, the fourth model is immersion, in which a significant part of the curriculum is taught through the medium of the additional language “with a correspondingly large increase in the time made available over the other three models” (p. 35). The first two models are the most widely applied and their common characteristics include (a) a fairly limited amount of time is dedicated to instruction per week; (b) teachers are homeroom teachers, that is, generalists, rather than language specialists; and (c) teachers’ proficiency in the foreign language (FL) often falls short of native speaker level. In many studies reviewed in this paper, the time devoted to ELL ranges between one to three sessions per week in the first two models. However, no comparative study was found on how weekly exposure impacts on outcomes.

Research has been conducted on all of these four models, although most typically the first two are referred to as ELL. Little research is available on awareness raising programs (type three, but see Horst, White, & Bell, Reference Horst, White and Bell2010) and especially on how children benefit from them over time. Our understanding of the main issues is that many findings may be relevant across these models. As Harris (Reference Harris and Nikolov2009) argues, “we are focusing so much on the ‘foreign’ in language learning that we overlook language contexts that actually have quite rich comparative potential” (p. 351). Based on findings of the Irish language primary school program, he claims that FL learning is not really that different from minority, heritage, or regional language learning. Similarly, a lot of the methodological issues are in common between the first two and the fourth model where curricular content is delivered in the FL.

The specific type of program has important implications for general aims and achievement targets in the L2. Overall, the achievements tend to be modest: YLs are not expected to achieve native levels of proficiency (e.g., Curtain, Reference Curtain and Nikolov2009; Haenni Hoti, Heintzmann, & Müller, Reference Haenni Hoti, Heinzmann, Müller and Nikolov2009; Inbar-Lourie & Shohamy, Reference Inbar-Lourie, Shohamy and Nikolov2009; see also the section on testing in this paper) and achievement targets reflect developmental stages in early language learning as children gradually shift from prefabricated utterances to more analyzed language use (Johnstone, Reference Johnstone, Enever, Moon and Raman2009a). In most contexts, programs also aim to develop favorable attitudes towards languages and language learning, and YLs’ own culture and identity (e.g., Prabhu, Reference Prahbu, Enever, Moon and Raman2009; Vickov, Reference Vickov, Nikolov, Mihaljević Djigunović, Mattheoudakis, Lundberg and Flanagan2007). The concern worded by Prabhu (Reference Prahbu, Enever, Moon and Raman2009) resonates in many educational contexts in relation to English as a lingua franca: “we would like this to lead to an enlargement of the linguistic repertoire and cultural horizons of many different communities, not a substitution of English for all or many other languages accompanied by a loss of different cultures” (p. 43).

Therefore, it is not by coincidence that results of many studies enquiring into how children benefit from different types of programs are evaluated as “no loss to L1” (e.g., Johnstone, Reference Johnstone2009b). The worry about the potential negative impact on first language development has led to several studies looking into how YLs’ languages interact in the early language learning process, and many of these document that exposure to a new language may favorably impact L1 development – a most desirable and fascinating outcome. Besides linguistic and socio-affective aims, metacognition and learning strategies are also often among the explicit aims of early language learning programs and how they interact with the learners’ L1 in complex ways. How awareness towards learners’ own culture and their identity are shaped by early language learning is another area, besides the ones above, where more longitudinal research is necessary in all types of programs.

Large-Scale Projects

A national survey conducted in the U.S. (Rhodes & Pufahl, Reference Rhodes and Pufahl2008) revealed both positive and negative trends in enrollment between 1997 and 2008. Most importantly, a significant decrease was found in the percentage of elementary and middle schools offering FL instruction. The decline occurred mainly in public elementary schools, whereas the percentage of private elementary schools teaching languages remained about the same.

Survey results revealed issues of unequal access to FL instruction. Students attending rural or low socioeconomic status (SES) schools were less likely to learn a FL. A disparity was found between public and private elementary schools: the latter offered languages at much higher rates. As for the target languages offered, the most popular language was Spanish; the teaching of French, German, Japanese, and Russian decreased, whereas Arabic and Chinese increased. More than one quarter of all elementary school FL teachers were not certified at all, and schools found it difficult to find qualified teachers.

A working group of the EU member states has conducted a survey on early language learning at a very early age in a project called Early Language Learning at Pre-primary Level in Europe: Current Situation and Future Perspectives. Data were collected with the help of a questionnaire and case studies of good practice from national ministries. The findings are expected to be published in a policy guidebook in 2011 (Hartung, personal communication, July 7, 2010). The main challenges were that the type and amount of data available vary a lot in different countries and ministries are not aware of what goes on in the private sector, a major issue in equity. One finding is clear: few teachers and staff are trained to work with this particular age group in most educational contexts; thus, teacher education is the main stumbling block of meeting parental needs.

A global project funded by the British Council (Investigating Global Practices in Teaching English to Young Learners) using a questionnaire and case studies is also in progress. So far over 4,000 teachers have filled in questionnaires in over 130 countries providing insights into classroom practices in local contexts (Garton, personal communication, July 6, 2010).

An innovative EU-sponsored longitudinal project called Early Language Learning in Europe (ELLiE, http://ellieresearch.eu) was launched in 2007. The main aim was to find out what European YLs can achieve by learning a FL in state schools, where the FL is just one of many school subjects and where a limited amount of time is devoted to FL instruction. During the three years of this seven-country project the same investigations were carried out in England, Italy, the Netherlands, Poland, Spain, Sweden, and Croatia (sponsored by the British Council) using a mixed-method approach. The progress of around 1,200 children was followed along three specific foci: policy implementation in each of the seven contexts, factors that contribute to the effectiveness of early language learning, and learning outcomes (both linguistic and non-linguistic). The role of the teacher and the impact of digital media on early language learning were two special investigation strands within the project. Some findings are discussed in the sections below.

Classroom Observation Studies

Investigations within the ELLiE project included classroom observation of learners’ linguistic and non-linguistic behavior during FL lessons focusing on learners’ attention, participation, relationship to teacher and classmates, language comprehension, and production. Findings of the Croatian cohort learning English in Grades 1 and 2 (Mihaljević Djigunović, Reference Mihaljević Djigunović2009b) showed how students’ engagement in pair work increased significantly and how interest, attention and engagement, and shyness and anxiety interacted in complex ways. Lundberg and Lindgren (Reference Lundberg and Lindgren2008) reported that YLs can pick up on formal aspects of language when focus is on meaning, and that spontaneous production emerges very slowly in the first two years.

Besides observing classroom processes and how children behave, some studies focus on how YLs perform certain tasks. These studies provide valuable insights into what actually happens, how children comprehend and complete tasks, which language they use for what purposes, and how they benefit from participating in them. In an in-depth observation study on classroom tasks implemented in early English classrooms in China (Peng & Zhang, Reference Peng, Zhang and Nikolov2009a, Peng & Zhang, Reference Peng, Zhang and Nikolov2009b), students responded to teachers’ fast questions in English exclusively. Practice was done in drill-like sequences in a lockstep fashion in huge classes, and the drills seemed to provide most of the learners with a feeling of success and allowed them to practice what they experienced as English.

Ten-year-olds’ peer–peer interactions on spot-the-difference tasks were analyzed by Pinter (Reference Pinter2007a, Reference Pinter, den Branden, Van Gorp and Verhelstb). After the task, the students were invited to watch their first and last performances and comment on the changes they noticed. Findings clearly indicated that peer–peer interactions at a very low level of competence brought various benefits. When asked what they thought about task repetition, children could clearly see the advantages: they learned from one another, noticed gaps in their own performances, and their anxiety also decreased.

Interaction of Learners’ Languages

An important development in language pedagogy concerns the revival of studies looking into how learners’ languages interact with one another as they develop multi-competence (e.g., Cook, Reference Cook1991). Recent studies tend to focus not only on how L1 impacts on L2 but also on interrelationships between languages shifting from transfer to access (Walter, Reference Walter2007). This means a more general approach to transfer: it is bidirectional and draws on all kinds of resources (skills, knowledge, and strategies) available to learners.

Another emerging trend is marked by the number of YLs’ languages: several studies involve bilingual children developing in three languages. For example, learners of English as a FL were bilingual in Catalan and Spanish (Muñoz, Reference Muñoz2006), Basque and Spanish (García Mayo & García Lecumberri, Reference García Mayo and García Lecumberri2003), Dutch and various other languages (Goorhuis-Brouwer & de Bot, Reference Goorhuis-Brouwer and de Bot2010), and Hungarian and Ukrainian (Huszti, Fábián, & Bárányné Komári, Reference Huszti, Fábián, Bárányné Komári and Nikolov2009). These studies document that the learners’ three languages develop in harmony, and there is no loss of L1 due to early exposure to a new language.

Standardized tests were used in an innovative way in a study in the Netherlands involving four cohorts of young (age 4+) Dutch and non-Dutch learners of English. Goorhuis-Brouwer and de Bot (Reference Goorhuis-Brouwer and de Bot2010) tested participants’ development over two years in English and Dutch with the help of Dutch and English versions of the Reynell test for language development. Their findings refuted arguments against early language learning assuming a negative impact on learners’ L1 and L2 development. Lower language ability children benefited more from early language learning than their peers with average skills.

One of the focal points, transfer, was similar in a small-scale correlational study conducted by Larson-Hall (Reference Larson-Hall2008). Japanese college students who started studying English early were examined on a phonemic discrimination and a grammaticality judgment task. Modest relationships were found and age effects seemed to emerge after a substantial amount of input had been gained, a finding in line with what Muñoz (Reference Muñoz2006) has repeatedly pointed out. YLs tend to develop at a slow rate compared to their peers who start at a later point in time.

Chinese EFL learners in an immersion program from kindergarten to Grade 3 were compared with YLs learning English for fewer hours (Knell et al., Reference Knell, Haiyan, Miao, Yanping, Siegel, Lin and Wei2007). In the experimental group, YLs studied half of the curriculum in English in groups of 50. They were assessed on a number of measures in Chinese and English. Although the immersion students outperformed their non-immersion peers on most measures, a special finding emerged: pinyin contributed to better results in both Chinese and English.

Early and late EFL beginners’ achievements were compared at age 14 in listening, reading, and writing in Croatian (L1) and English (L2) in a cross sectional study (Mihaljević Djigunović, Reference Mihaljević Djigunović2010a). The degree of interdependence between the two languages varied among the three skills: reading in L1 and L2 were the most strongly connected; the relationship between writing in L1 and L2 was the weakest. Correlations across L1 and L2 skills indicated stronger relationships in the case of earlier beginners than for their later starting peers. The emerging patterns of multi-competence developed by the age of 14 indicated different interactions between YLs’ L1 and L2.

Reading comprehension in L1 and L2 was compared in a complex study on sixth and eighth graders (Nikolov & Csapó, Reference Nikolov and Csapó2010) to explore the relationships between YLs’ reading comprehension in English and Hungarian. Moderate significant relationships were found between students’ performances in the two languages in both grades; variance explained by a subtest of verbal analogies was similar (sixth grade) or higher (eighth grade) than in the case of L1 reading as an independent variable. The underlying aptitude construct impacted reading comprehension in the two languages in different ways, and correlations were stronger between L2 scores.

An international comparative study examined the levels of achievement of 8th grade learners of English in listening, reading, writing, and pragmatics in Croatia and Hungary (Mihaljević Djigunović, Nikolov, & Ottó, Reference Mihaljević Djigunović, Nikolov and Ottó2008). Earlier starters were slightly, but significantly better on all measures; the correlations between skills were stronger in Hungary. Croatian learners outperformed their Hungarian peers despite the fact that they tended to start English later, in larger groups, and for fewer hours – a thought provoking result in light of what is generally expected in more intensive programs. A comparison of the two contexts suggested that out of class exposure to English and quality of EFL teaching were more favorable in Croatia.

Individual Differences

The misconception that children are so similar to one another that there is no real need to focus on individual differences in the field of early language learning is widespread (Mihaljević Djigunović, Reference Mihaljević Djigunović and Nikolov2009a). We argue for the importance not only of attitudes, motivation, learning strategies, and aptitude, but also of language anxiety, something that has only recently been recognized as worth researching in YLs. What is, however, highlighted is the need for investigating interactions of individual differences. Instead of single individual differences variable, it is their interactions that can meaningfully inform early language learning processes. Another valuable perspective is an analysis of interactions between individual differences variables and contextual factors (Mihaljević Djigunović & Szpotowicz, Reference Mihaljević Djigunović and Szpotowicz2008).

Affective Development

Students’ motivation is an extremely popular research area, and fortunately a number of recent longitudinal studies have allowed researchers to examine how attitudes and motivation change over time. The time periods studied range from one to several years, and some studies compare and contrast different groups; a few inquiries compare what students claim happens in their classrooms versus what actually does and to what extent they enjoy them.

A national sample of Hungarian eighth graders (age 14) was involved in a longitudinal questionnaire study conducted by Dörnyei, Csizér, and Németh (Reference Dörnyei, Csizér and Németh2006) in 1993, 1999 and 2004. The results document the larger picture, namely, how students’ language learning attitudes and motivation changed in a dynamic way reflecting globalization processes, but do not explore the classroom level.

An exceptionally long period of time was covered in a study conducted in the U. S. Heining-Boyton and Haitema (Reference Heining-Boynton and Haitema2007) combined a large-scale longitudinal survey and retrospective interviews with a few participants after ten years on attitudes towards early language learning. The inquiries were unique because the students were followed not only in elementary school, but also for years after their foreign language in the elementary school (FLES) experiences. In corroboration of the quantitative data, a qualitative analysis revealed that, for a majority of the students, FL study was viewed positively, as were FL speakers and their cultures.

Japanese elementary school pupils’ intrinsic and extrinsic motivation for learning English was the focus of a study by Matsuzaki Carreira (Reference Matsuzaki Carreira2006). The author investigated how third and sixth graders’ motivation changed with age. A steady developmental decline was found in both types of motivation for learning English. The author attributed it to a general fall in children's motivation in all subjects in Japanese elementary school pupils and found considerable room for improvement in early language learning in Japan.

Declining language learning motivation, however, is not limited to the above contexts. A comparative analysis of data on the motivation of YLs in the ELLiE project involving learners from six countries showed that initial motivation was very high and that girls were more motivated than boys (Szpotowicz, Mihaljević Djigunović, & Enever, Reference Szpotowicz and Szulc-Kurpaska2009). An inspection of what participants found most inspiring showed that it was learning new words. This suggested that, at the start, YLs perceived language learning as vocabulary building. Analysis of motivation in the subsequent years is still under way but preliminary findings indicated a slow decrease in motivation (Lopriore, Reference Lopriore2009; Mihaljević Djigunović, Reference Mihaljević Djigunović2010b). Another study of 36 Croatian ELLiE participants (Mihaljević Djigunović & Letica Krevelj, Reference Mihaljević Djigunović and Letica Krevelj2010) confirmed that YLs started their L2 learning with positive attitudes toward classroom activities. However, under less than ideal conditions (large groups, two lessons per week, unqualified teachers), which was the current reality in many European contexts, some YLs soon started to develop negative attitudes to L2 learning. Also, children started L2 learning with overly favorable perceptions of their abilities and achievements. However, as their early language learning experience developed they soon started building their self-perceptions more realistically, and metacognition developed quite quickly with some young L2 learners.

In a different comparative study involving Croatian and Italian YLs (N = 91), Lopriore and Mihaljević Djigunović (in press) conducted a mixed methods investigation of their initial attitudes to learning English and attitudinal development over two years. In both cohorts, initial attitudes of YLs to EFL learning were mostly positive. In the second year, some YLs’ attitudes continued to be positive, but many of them thought English was more difficult. Others liked it even more because they felt successful in classes.

The popular belief that YLs are anxiety-free is the reason that there have been few studies on language anxiety in early language learning. However, anxiety has been observed in YLs as well. In his comparative study of monolingual and bilingual YLs of English, Legac (2007) found that bilingual (Croatian-Albanian, Croatian-Czech, and Croatian-Italian) learners experienced lower levels of both listening anxiety and overall language anxiety than their monolingual peers. In another study, Mihaljević Djigunović and Legac (Reference Mihaljević Djigunović, Nikolov and Ottó2008) compared mono- and bilingual seventh graders’ anxiety during the three stages of the EFL learning process (input, processing and output). They found that bilinguals had lower anxiety in all three stages. The authors proposed that bilinguals, thanks to their more extensive experience in using two languages in everyday life, developed higher linguistic self-confidence, which contributed to lower language anxiety.

Academic Achievement and Cognitive Development

The relationship between general academic achievement and learning French and Spanish in FLES groups was explored in an impressive large-scale longitudinal study by Taylor and Lafayette (Reference Taylor and Lafayette2010). Remarkable findings emerged from the examination of continued learning of FLs over a period of three years from Grade 3 to Grade 5 in Louisiana public schools. Participants’ achievements were tested with the help of standardized tests in a variety of domains (English language arts, mathematics, science, and social studies). Students learning a FL significantly outperformed their peers who did not study a new language both in English as well as the other subjects. This result provides a very strong argument for early exposure to a FLES: the benefits transferred to general school achievements and the learners’ L1.

Other studies have researched language learning aptitude, an under-researched area in the case of YLs, exploring how cognitive skills impact early language learning in two countries characterized by a tradition of streaming learners according to their learning abilities – a reason why placement is an issue from an early age. Greek learners were examined in a study involving five- to nine-year-old learners’ cognitive abilities and their learning of English vocabulary in Greece. Significant relationships were found between YLs’ aptitude and their vocabulary development in EFL (Alexiu, Reference Alexiou and Nikolov2009).

Three independent studies explored how YLs’ aptitude contributed to their achievements in FLL. In Hungary, where streaming has been typical for many decades and English tends to be provided for the more able learners, whereas learners of German lag behind their peers, one study (Kiss & Nikolov, Reference Kiss and Nikolov2005) involved over 400 sixth graders. Results showed that 22 % of the variation in English performances was explained by participants’ aptitude. An even younger age group was involved in a similar project piloting an aptitude test for 8-year-olds (Kiss, Reference Kiss and Nikolov2009). The test was found to discriminate well among learners.

The role of inductive reasoning was explored in a large-scale longitudinal study on representative samples of Hungarian learners of English and German (ages 12, 14, 16, and 18). Learners’ scores on an inductive reasoning test predicted a large portion of the variance in their performances in reading (e.g., 14% in English at age 12). In the younger age groups inductive reasoning explained larger portions of the variance than in the older cohorts. The relationships between the cognitive variable, L1 reading, and other factors were stronger than in the case of older learners; relationships between L2 skills weakened over time (Csapó & Nikolov, Reference Csapó and Nikolov2009).

How Young Learners Use Strategies

YLs’ learning and communication strategies were examined by either focusing on how they perform certain tasks or by examining their classroom behavior in larger groups. Intervention studies are the exception in researching YLs’ strategies. In tasks based on interaction between peers on spot-the-difference tasks Pinter (Reference Pinter2006b) analyzed what strategies 10-year-olds used and how, and compared them to adults on similarly low levels of English working on the same tasks. She found that YLs used similar strategies to adults: they were able to respond to each other, clarify messages, keep tally of the differences, and they interacted confidently. However, the adult learners used the strategies more often.

A study carried out by Tragant and Victori (Reference Tragant, Victori and Muñoz2006) found that in the early stages of learning children used predominantly memorization strategies and generally relied on external resources. Focusing on the use of vocabulary, pronunciation, spelling, reading, and writing strategies by YLs of differing start ages (8, 11, and adults) who had the same amount of instruction, Tragant and Victori found evidence for age-dependent developmental stages in strategy use that were, however, independent of proficiency level or stage of learning. In contrast to older learners, YLs very rarely used varied and complex strategies at the beginning of learning.

Two studies compared reading strategies of more and less successful EFL learners. Šamo (Reference Šamo, Kormos and Kontra2006) found that more successful learners used 11 reading strategies with a significantly higher frequency than their less successful peers. Quite a few of the strategies were metacognitive; thus, they involved monitoring the process of reading. Nikolov (Reference Nikolov2006) compared the use of reading and writing test-taking strategies of young top achievers and low achievers. Her study indicated variable frequency and types of strategies used by YLs of both groups; thus the author challenged the widely accepted assumption that good language performance necessarily implies using more and varied strategies.

Greek sixth graders’ writing strategies were examined with the help of questionnaires, think-aloud protocols, and interviews. Griva, Tsakiridou, and Nihoritou (Reference Griva, Tsakiridou, Nihoritou and Nikolov2009) found both similarities and differences between boys and girls as well as between more and less successful learners. Generally, all learners considered writing a difficult skill, and sometimes revised and proofread their drafts, translated, and used bilingual dictionaries. Differences between more and less successful writers were observed in the perception and use of metacognitive strategies: the former used more strategies and employed higher-level processing during writing.

An intervention study was implemented by Macaro and Erler (Reference Macaro and Erler2008) aiming to raise the achievement of young-beginner readers of French through strategy instruction in England. A sample of 62 learners aged 11–12 participated in a program of reading strategy instruction lasting 14 months. Measures were taken of French reading comprehension, reading strategy use, and attitudes towards French before and after the intervention, and the findings were compared with a control group. Strategy instruction improved comprehension of both simple and more elaborate texts, brought about changes in strategy use, and improved attitudes towards reading.

A classroom observation study examined Chinese learners of EFL. Peng and Zhang (Reference Peng, Zhang and Nikolov2009b) analyzed the classroom discourse data of fifth graders and found that although the most frequently applied communication strategies were L1-based ones when students asked for clarifications in Chinese and tried to cooperate with their peers, there were few examples of strategy use. According to their explanation, in large classes of 50–60 learners, teachers insisted on repetition mostly and thus prevented learners from applying strategies creatively.

Social Background and Equity

Socioeconomic status (SES) is a widely accepted variable contributing to school success in educational research; however, it is rarely addressed explicitly in studies on early language learning. In educational contexts where the private sector offers a range of early programs, socio-economically advantaged children's parents are more able to afford early English or other languages. Thus, issues of equity and diversity are often entangled in early language learning programs, as, for example Butler (Reference Butler2007) argues in the case of Japan, or Feng (Reference Feng2009) when discussing convergence and divergence in policy and practice in China, or Moon (Reference Moon and Nikolov2009) in Vietnam. Studies are not limited to Asia, however. SES was found to be an important and worrying factor in the U.S. survey cited above (Rhodes & Pufahl, Reference Rhodes and Pufahl2008), but it is also often mentioned in many other contexts (see e.g., national case studies in Enever, Moon, & Raman, Reference Enever, Moon and Raman2009 and European contexts below). As Johnstone (Reference Johnstone2009b) noted, learners’ SES may impact on the generalizability and sustainability of national policy initiatives, as was documented in his discussion of an early French program in Scotland.

SES was addressed in several studies involving nationally representative samples of Hungarian learners studying English and German (Csapó & Nikolov, Reference Csapó and Nikolov2009; Nikolov, Reference Nikolov and Nikolov2009d). The findings showed the larger picture: significant relationships were found between learners’ SES and their language proficiency, language learning plans, and motivation, and SES was also related to language choice: more children with higher SES backgrounds tended to study English than their peers, who studied German.

Greek learners participated in an innovative study examining the relationships between YLs’ SES and language development in English in a particular town in Greece (Mattheoudakis & Alexiou, Reference Mattheoudakis, Alexiou and Nikolov2009). Interestingly, all learners, irrespective of their SES, received private tutoring in addition to learning English at school. Although their families’ SES significantly impacted the participants’ level and progress in English, no such relationship was found for their motivation, a finding in contrast with the previously discussed studies.

A different approach was used by Carmel (Reference Carmel and Nikolov2009). She applied critical discourse analysis to English for YLs. She found that English for young language learnersit was seen as a valuable asset and a widely spread socio-cultural phenomenon in Israel, and documented it as a social, educational and economic divider in parents’, stakeholders’ and pupils’ discourses she analyzed. She found three main discourses: Americanization and internationalism, good parenting, and consumers. Similar analyses could cast light on how early language learning is perceived in other contexts.

Learning Difficulties in Early Language Learning

Due to the growing number of learners and teachers involved in early language learning, a variety of problems has emerged. Challenges caused by individual differences and special needs may make teachers’ jobs demanding. Many teachers find it hard to work with children at risk or suffering from attention deficit (e.g., Ferrari & Palladino, Reference Ferrari and Palladino2007; Morrissette, Reference Morrissette2009). A full volume edited by Kormos and Kontra (Reference Kormos and Kontra2008b) focused on learners with special educational needs, most importantly dyslexic YLs. Ndlovu and Geva (Reference Ndlovu, Geva, Kormos and Kontra2008) stressed the importance of an early diagnosis of learning disability; Sparks, Ganschow, and Patton (Reference Sparks, Ganschow, Patton, Kormos and Kontra2008) claimed that L2 learners with learning disabilities need support in both L1 and L2. Smith (2008) pointed out the lack of adequate language teacher preparation for working with such learners, whereas Helland (Reference Helland, Kormos and Kontra2008) provided an insight into characteristic patterns of spelling errors made by dyslexics, and showed clearly the importance of an interdisciplinary approach to problems.

Classroom Observation Studies on Teachers

Teachers’ roles vary to a great extent in the YL classrooms. A study in Japan (Aline & Hosoda, Reference Aline and Hosoda2006) explored team teaching participation patterns of homeroom teachers in interaction with native speaker teaching assistants in English activities classes in Japanese public elementary schools. Four patterns were observed in homeroom teachers’ behavior: they acted as (a) a bystander, (b) a translator, (c) a co-learner of English, or (d) a co-teacher. The participation patterns affected the classroom interaction in distinct ways.

In a very different EFL context, a study combining classroom observations and follow-up interviews documented that many Hungarian teachers of YLs felt that the role of a caretaker was demanding and keeping children on task required a lot of energy and special pedagogical skills (Nikolov, Reference Nikolov2008). Many of them wished that they could teach older learners, insisted on streaming YLs into ability groups, and preferred teaching more able ones. They lacked not only proficiency and pedagogical skills, but also a desire to improve their practice. Many teachers perceived games and storytelling as a waste of time, and looked forward to what they perceived to be proper teaching in later years.

In a study on pre-service teachers’ teaching practice, very young children in kindergartens were found unpredictable and presented a challenge to novice English teachers in Poland (Szulc-Kurpaska, Reference Szulc-Kurpaska, Nikolov, Mihaljević Djigunović, Mattheoudakis, Lundberg and Flanagan2007). They found it difficult to evoke YLs’ interest, involve them in carefully designed tasks, and manage their activities in English.

Teachers’ proficiency in the target language and its impact on children's development is usually seen as a delicate issue. Inbar-Lourie (Reference Inbar-Lourie2010) explored the instructional linguistic choices of EFL teachers in the Israeli context. She examined how teachers adhered to English or incorporated it with the learners’ L1. The study involved six teachers and 285 young EFL learners aged 6–8 whose L1 was either Hebrew or Arabic. Data were collected with observations and interviews to explore teachers’ classroom practice, beliefs about their learners’ needs, and how their use of L1 contributed to development in English. Teachers were conscious about their L1 use and their decisions concerning when and how to apply L1 and L2 were based on principles they worded in terms of learners’ needs.

Looking into YLs’ exposure to L2, Mihaljević Djigunović (Reference Mihaljević Djigunović2010b) found that teachers of YLs perceived their L1 use to be lower than observation data showed. However, when learner achievement was entered into the equation, it turned out that learners in teachers’ classes who used the least L1 in their English classes did not score the highest on receptive and productive tasks. An analysis of classroom teaching that the highest-scoring group was exposed to suggested that it was not the amount of L1 versus L2 use that was decisive. Rather, it was when and why the L1 and L2 were used that was most important.

Data were collected from classrooms and interviews by Nagy (Reference Nagy2009) focusing on the amount and function of mother tongue use (Hungarian) by teachers working in early and later start fourth grade EFL groups. According to the teachers, their practice was constrained by various factors: children's aptitude, proficiency in English, and the teachers’ preference for using grammar, translation, and memorization to prepare the children for examinations.

Interview and Survey Studies on Teachers

How teachers think about their practices, YLs’, and their own needs is an area where important insights have been gained. As teaching children tends to be a relatively new challenge in many contexts, the introduction of early language learning often happens overnight without appropriate teacher education. Problems are related not only to teachers’ proficiency and new methodologies reflecting how children learn an L2, but also to what the use of a communicative methodology involves.

Butler (Reference Butler2005) explored teachers’ views on communicative methodology in Korea, Japan, and Taiwan and found that they were concerned about using communicative activities for a number of good reasons. In a different study, Butler (Reference Butler2004) asked early language learning teachers to reflect on their own language proficiency and learned that they felt their productive skills lagged behind their receptive skills and they needed to improve their English to be able to implement programs they were involved in (Butler, Reference Butler2004). A similar picture emerged from Moon's (Reference Moon and Nikolov2009) study on Vietnamese teachers. Her case studies cast light on teachers’ needs and how such needs might be met. She also demonstrated how training in communicative language teaching, course materials and local traditions interact in teachers’ beliefs and daily classroom practices, which generally reflected a behaviorist model of language learning and teaching.

The contradictions revealed between teachers’ stated beliefs and their actual classroom practices are typical in other countries as well. Lugossy (Reference Lugossy, Nikolov, Mihaljević Djigunović, Mattheoudakis, Lundberg and Flanagan2007) explored how in-service teachers implemented new ideas in a picture book project. She asked them to use storybooks with their YLs and to reflect on their own practice. She compared what teachers knew and believed, and what they actually did with their YLs of English. She found that very few teachers changed their practice over time.

The findings of a large-scale questionnaire survey in China are more positive. Wang (Reference Wang and Nikolov2009) surveyed 1,000 primary teachers of English in urban and rural schools and found teachers’ beliefs with regard to early English and learner-centered teaching to be supportive and in line with the central curriculum. Teachers’ self-reported classroom practices reflected that they implemented learner-centered teaching, with the exception of frequent drills. The overall findings showed teachers’ enthusiastic support for implementing the new curriculum involving not only teaching YLs, but also documented a major shift towards learner-centered teaching – a relatively new concept in the Chinese context. It is to be seen, however, how teachers implement meaning-focused tasks in learner-centered ways on a daily basis.

In-depth semi-structured interviews with seven Croatian EFL teachers of YLs (Mihaljević Djigunović, Reference Mihaljević Djigunović2010b) offered evidence of a range of teachers’ beliefs. All believed that children learn FLs through games, songs, and role-plays; some thought that YLs need a leader and the teacher needs to take this role. Interestingly, none of the teachers believed that four weekly hours instead of the two prescribed by the national curriculum would contribute essentially to success in early language learning. As for the use of the L1 in the L2 classroom, all shared the belief that L1 was necessary to make sure that YLs could understand what went on and what they were learning.

The role of awareness-raising in the YL classroom is rarely focused on; thus, an exploratory research study conducted by Horst, White, and Bell (Reference Horst, White and Bell2010) in Canada is an exception. They examined how instruction could be designed so that learners build on their L1 knowledge in acquiring a new language. They involved two teachers, one of French (L1) and another of English (L2), and their 9–10-year-old francophone learners of English in cross-linguistic awareness raising activities to explore how certain themes in their L1 syllabus could be exploited in their L2 classes. The results showed how differently the two teachers applied the special pilot activities designed to draw learners’ attention to a variety of specific L1 features.

The Construct: What Children Can Do

The dynamic spread of early language learning has brought about increased interest in testing YLs’ proficiency and how they progress in their new language over time (McKay, Reference McKay2006). The main issue in testing YLs concerns the construct. As Inbar-Lourie and Shohamy (Reference Inbar-Lourie, Shohamy and Nikolov2009) framed it, early language learning programs vary to a great extent from awareness raising to language focus programs, and from content-based curricula to immersion. They argued that testing must be in line with “the assessment constructs along with the language-content program continuum, particularly within the current integrated or embedded models in YL classrooms given its current status and use in diverse contexts” (p. 93). They demonstrated how the reflective process they applied to various tests and exams illustrated “the dynamics of testing construction and validation which need to be constantly re-thought and revised” (p. 94).

Challenges in testing YLs are manifold. Low proficiency levels need to be defined and described along a continuum in small steps so that children's relatively slow development can be documented. This is important for their teachers, parents, and YLs themselves so that they feel that they are making progress. Also, the earlier instruction starts, the more probable it is that YLs are at the early stages of literacy development; what is typical for older learners cannot be automatically applied to YLs in the four skills. Teachers and test developers must bear in mind that YLs are different from older ones not only in their background knowledge of the world and other factors discussed in previous sections, but assessment may impact attitudes, motivation and anxiety in more important ways than in the case of older learners. Therefore, besides considering assessment of learning, a focus on assessment for learning is of crucial importance.

Assessment Frameworks and Exams

The need to quantify what children can do in terms of proficiency outcomes has resulted in various documents. In the U.S., the aim was to provide various types of programs with a profession-wide scale. For this purpose, several proficiency-based language assessment instruments were developed for early language learning. Curtain (Reference Curtain and Nikolov2009) provided a detailed comparative analysis of various proficiency standards, performance guidelines as well as assessment instruments reflecting what YLs are expected to be able to do in various target languages. Work in Europe has focused on adapting Common European Framework of Reference for Languages descriptors (CEFR, Council of Europe, 2001; Figueras & Noijons, Reference Figueras and Noijons2009; Jones, & Saville, Reference Jones and Saville2009; Little, Reference Little2007) to YLs’ needs and examinations (e.g., Hasselgreen, Reference Hasselgreen2005; Papp & Salamoura, Reference Papp and Salamoura2009; Pižorn, Reference Pižorn, Figueras and Noijons2009).

Presently, to our knowledge, four international exams are available to provide YLs and their parents with a certificate on children's proficiency in English – for money (Pearson Test of English Young Learners, www.pearsonpte.com/PTEYoungLearners; Cambridge Young Learners English Tests, www.cambridgeesol.org/exams/young-learners; Integrated Skills in English, www.trinitycollege.co.uk; City & Guilds ESOL for Young Learners, www.cityandguilds.com). A detailed analysis is beyond our scope, but there appear to be some factors in common. In all four exams the construct is more towards the middle of the language–content continuum with a focus on some typical topics which YLs can be realistically expected to be familiar with, and the levels cover A1 and A2 in the CEFR (Council of Europe, 2001). No research is available on how taking exams impacts on children’ motivation, anxiety and later progress. Several in-house publications are available on how Cambridge Young Learners English Tests were developed and validated (e.g., see special issue 28, 2007 at http://www.cambridgeesol.org/rs_notes/rs_nts28.pdf), but no such information is available for the other three exams.

As early language learning programs become part of national curricula, more and more countries develop and publish their achievement targets and introduce national exams for an increasingly younger age group. In Switzerland, for example, a three-year longitudinal study was implemented on YLs development (Haenni Hoti, Heinzmann, & Müller, Reference Haenni Hoti, Heinzmann, Müller and Nikolov2009). In Slovenia, research into what YLs can do resulted in validated national tests for sixth graders (Pižorn, Reference Pižorn, Figueras and Noijons2009).

Assessment For Learning

Recent research into educational assessment has emphasized how dynamic testing can boost learning potential (Sternberg & Grigorenko, Reference Sternberg and Grigorenko2002) and a shift away from assessment of learning to assessment for learning has been observed. The emphasis is on how children benefit from classroom testing and feedback from teachers (Assessment Reform Group, 2002; Black & Wiliam, Reference Black and Wiliam1998; Davison & Leung, Reference Davison and Leung2009; Leung & Scott, Reference Leung and Scott2009; McKay, Reference McKay2006; Teasdale & Leung, Reference Teasdale and Leung2000). As McNamara and Roever (Reference McNamara and Roever2006, p. 251–252) argued, assessment should be sensitive to the issue of readiness to develop and there should be more research on learners’ ability to profit from instruction and different kinds of interaction. In most of the cases in early language learning classrooms, interaction takes place between teachers and YLs, as well as between peers. Some studies discussed earlier provided insights into how task repetition and peer interaction contributed to better oral performance (e.g., Peng & Zhang, Reference Peng, Zhang and Nikolov2009a, Reference Peng, Zhang and Nikolov2009b; Pinter, Reference Pinter2007a), but a lot more similar studies are needed to find out more about the way feedback scaffolds learning.

How Teachers Assess Young Learners

There is not enough research on how teachers assess YLs. Butler (Reference Butler2009a) provided an overview of studies on teacher-based assessment in three Asian countries. She found that Korean and Taiwanese teachers faced a number of difficulties. They had to cope with large classes and lacked knowledge about teacher-based assessment, sufficient time to administer tests, and information on how best to use information gained through teacher-based assessment for formative and summative purposes. In Japan, teachers were not required to assess their YLs; however, they did use some self-assessment. In a different study, South Korean elementary and secondary school EFL teachers were asked to assess videotapes of sixth-grade students’ group activities (Butler, Reference Butler2009b). The teachers varied substantially in their overall evaluations.

The first findings of an exploratory study involving EFL teachers in Hungary showed that they applied two types of classroom assessment practices (Hild & Nikolov, Reference Hild and Nikolov2010). First, they gave rewards for best performances on tasks; rewards were cumulative in nature. Good performers earned small rewards on a regular basis, whereas their less successful peers felt that they deserved nothing. This practice meant that able and more motivated learners got positive feedback and the lack of reward functioned as negative feedback for their less successful peers. Second, the other type of assessment teachers applied involved achievements on classroom tests, most frequently in reading and writing. This dual practice reflected what teachers saw as the aim of assessment: motivation and feedback on actual development in English.

Portfolios and self-assessment may contribute to autonomy and better learning. Some experts advocate the use of portfolios, especially in Europe. Hasselgreen (Reference Hasselgreen2005) considered what the CEFR (2001) and the European Language Portfolio offered to YLs. She explored ways in which they were applied in Norway. However, as Little (Reference Little2007) stated, portfolio assessment was often seen by teachers and learners as extra work. Overall, not enough is known about actual uses of portfolios in YLs’ classrooms.

Two studies looked into YLs’ self-assessment in Korean elementary schools (Butler & Lee, Reference Butler and Lee2006, Reference Butler and Lee2010). The first one examined the validity of students’ self-assessments of their oral performance in EFL in two types of assessments; whereas the second study explored the effectiveness of self-assessment on a regular basis for a semester during English classes.

The above discussions on teacher-based assessment lead to an important point connected to assessing YLs. Edelenbos and Kubanek-German (Reference Edelenbos and Kubanek-German2004) coined the term diagnostic competence. In their view, it is “the ability to interpret students’ FL growth, to skillfully deal with assessment material and to provide students with appropriate help in response to this diagnosis” (p. 260). This is a definite challenge in the YL classroom, especially in contexts where huge classes are typical, as in diagnostic assessment teachers are expected to work in one-on-one contact with the children rather than in a whole class.