The Development of English Writing in Bilingual Children Attending Weekend School in Japan During the COVID-19 Pandemic

Bilingual children schooled in the dominant societal language can acquire home language literacy with home and community support. This paper reports on the development of English writing in 17 Japanese-English bilingual children (ages 9 to 13) who attended English weekend school in Japan during the COVID-19 pandemic. We assessed their English writing using a US standardized test before and during the pandemic for two years when their weekend school switched to an online format. Many children’s age-adjusted writing scores were within the Average range or above before the start of the pandemic


INTRODUCTION
Bilingual children can become biliterate given the right circumstances.Young children who speak two languages have the necessary prerequisites to establish biliteracy (Bialystok, 2007).Having basic interpersonal communication skills (BICS) in both languages, simultaneous bilingual children are ready to acquire cognitive academic language proficiency (CALP) (Cummins, 2000) in those two languages.They are well-positioned for biliteracy even when schooled only in the societal language because their established oral ability in each language helps them to make stronger connections with corresponding written forms (Hornberger, 2003).Unlike sequential bilinguals, simultaneous bilinguals already possess societal language proficiency to function optimally at school, so they only need to focus on learning the "home language" (hL) outside school (henceforth, we will refer to hL with a lower-case "h" to indicate its minority status and as an umbrella term encompassing "heritage" languages which can be, but are not always, the home language, as we will show in our paper).Bilingual children's societal language skills can be transferred to the hL when there is adequate exposure and motivation to learn it (Cummins, 2000).Nevertheless, the road to biliteracy is not easy when hL literacy is acquired outside mainstream education.This paper explores the biliteracy potential of simultaneous bilingual children who learn their hL at home and in weekend school (also known as "complementary," "heritage," "ethnic," "supplementary," or "Saturday" schools).As part of a larger project, this paper reports on the development of English writing during the COVID-19 pandemic when Japanese-English bilingual children attended weekend school online.

HOME LANGUAGE LITERACY AT HOME
In the sociocultural approach to learning, adults initiate children into literacy practices and help them develop new skills (Gregory et al., 2004).In a simultaneous bilingual acquisition setting, such as with the One-Parent-One-Language (OPOL) approach, biliteracy may begin in infancy when parents start reading to their children in their respective languages.However, biliteracy does not naturally develop with age (Kim & Pyun, 2014).Regular literacy practices that provide rich linguistic input and hone children's CALP in the hL are necessary, particularly as children become older and increasingly dominant in their school language (Oriyama, 2016).However, many parents do not know how to help their children progress from BICS to CALP or do not have the time to do so (Tran et al., 2022).Only two out of the 28 Japanese-English intermarried families in the UK in Okita (2002) persisted with Japanese literacy learning at home until their children became teenagers.Parents who teach literacy likely have positive attitudes toward the hL and strong impact beliefs that they can teach their children to read and write (Nakamura, 2019).These attitudes and impact beliefs translate into home literacy practices which develop hL literacy skills (Kim & Pyun, 2014).
These practices include shared reading and play to reenact school literacy activities (Reyes, 2012).English-speaking parents in Japan continued shared reading in English even when their children could read independently to provide richer input contained in more complex books (Nakamura, 2019).Japanese-American parents used manga, daily journals, and Game Boy strategy guides to give their children a genuine and authentic purpose for learning Japanese (Hashimoto & Lee, 2011).Russian-speaking parents in Cyprus also adopted a child-centered, game-based approach to teach Russian (Karpava, 2021).
The home literacy environment also matters as the number of hL books is linked to hL vocabulary and use (e.g., Rydland & Grøver, 2021).Books that match developmental needs and engage children in reading or writing maximize learning (Li, 2007).Reading for pleasure in the hL develops hL reading skills (Taniguchi, 2021).While we know less about hL writing, reading skills do predict writing ability (e.g., Kim et al., 2018).Both share cognitive processes and knowledge bases (Fitzgerald & Shanahan, 2000).Word recognition predicts orthography, sentence reading predicts sentence construction, and reading fluency predicts writing fluency (Ahmed et al., 2014).In Kim and Pyun (2014), the frequency of Korean reading and writing practice predicted the writing proficiency of Korean hL learners.

HOME LANGUAGE LITERACY IN WEEKEND SCHOOL
Children also learn their hL in weekend school.Those who attend weekend school tend to have higher literacy skills than those who do not (Lao & Lee, 2009).These schools provide parents with guidance about hL literacy, and are a community of practice where families can exchange ideas and information on hL learning (Nakamura, 2019;Oriyama, 2016).They also help maintain oral skills and the minority culture (Cruickshank et al. 2020;Li & Wu, 2010).However, despite their importance in minority communities, weekend schools face various challenges, particularly limited instruction time (Lee & Chen-Wu, 2021).
Weekend schools usually operate in borrowed spaces and lack educational resources, particularly related to technology (Nordstrom & Jung, 2022).However, whenever available, technology can expand hL learning beyond the spatial and temporal limits of weekend schools.Palladino and Guardado's (2018) study of two schools in Canada shows how relatively accessible digital tools such as wikis and blogs effectively extended the weekend school experience.They allowed teachers to communicate key information (e.g., assignments) and exposed students to more written texts and multimodal content out of class and freed classroom time for active participation.Students were highly engaged in online tasks and learned independently by reading, practicing, and testing themselves online.Nordstrom (2015) observes that online synchronous weekend school lessons provided more writing opportunities because students communicated in group chats and other writing.They also utilized online translation tools and received scaffolding from their parents at home.Drawing from this larger set of bilingual resources, they could learn without peer pressure, anxiety or stigmatization.Although emergency online learning due to COVID-19 lockdowns posed an unprecedented challenge, weekend schools showed that they could adapt well.Young and White (2022) found that weekend school teachers quickly familiarized themselves with multiple educational technology platforms and developed new pedagogical practices, e.g., online reading sessions by older children to younger ones.
A challenge for weekend schools is that the teachers' pedagogical approaches and discursive practices may not appeal to children (Curdt-Christiansen, 2006).It may also be linguistically and culturally difficult for children to relate to materials designed for native-speaking children (Hashimoto & Lee, 2011;Lee & Chen-Wu, 2021).Kim and Pyun (2014) found that literacy in the hL does not relate to the number of years of weekend school attendance.Moreover, children tend to quit weekend school at older ages.Teenagers generally do not perceive it as a popular or enjoyable activity, focusing instead on college-admission activities (Lee & Chen-Wu, 2021).However, they may continue and strive for higher proficiency when hL skills can be converted into linguistic capital.Teenagers learn Swedish in a weekend school in Australia partly because their Swedish skills are recognized as part of their high school qualification (Nordstrom, 2022).In the UK, weekend schools tutor students taking their hL as a GCSE or A-level subject (Lytra, 2013;Young & White, 2022).The application of linguistic skills to future studies/careers is likely an important factor for continued hL learning.

HOME LANGUAGE LEARNING IN JAPAN
Japanese is the dominant societal language in Japan.Even a global language like English exists as a foreign language, albeit a prestigious one; it is the de facto foreign language taught in Japanese schools.Meanwhile, the teaching of other hLs hardly exists in Japanese schools; hLs may be used to help minority children learn academic subjects and promote inclusivity but are usually not taught as language subjects (Majima & Sakurai, 2021).Minority children rarely possess grade-level abilities in both languages (with the exception of the Japanese-Chinese bilingual child in Majima et al., 2019).To learn their hL, minority children must attend bogo kyooshitsu (mother tongue classes).These classes are generally organized by local government bodies and non-profit organizations (NPO) and staffed by volunteers.They are found in areas (e.g., Osaka) with a high foreign population.Usually, languages spoken by the largest immigrant groups (e.g., Portuguese, Chinese, Spanish, Vietnamese, and Korean) are offered.Some bogo kyooshitsu may teach several hLs for different minority groups or hold classes of different proficiency levels (Matsubara, 2004).These classes teach oral and literacy skills.They also help facilitate cognitive and identity development and build self-esteem (Sakurai, 2019).How these classes are run depends on educators' and parents' expectations and children's proficiency levels.However, their effectiveness and long-term viability are hampered by the lack of teacher training and resources, dependency on volunteer groups, limited instruction time, and children's mixed abilities and lack of motivation (Saito, 2005;Takahashi, 2015).
Other than the bogo kyooshitsu, another grassroot initiative is what we termed "weekend school".Calling these weekend schools "mother-tongue" classes would be inappropriate because attendees are predominantly simultaneous bilingual and bicultural children with two mother tongues, unlike The English weekend schools' strong literacy focus differs from the bogo kyooshitsu which also teaches oral skills so that children with minimal L1 ability can communicate better with their parents (Takahashi, 2015).These English weekend schools also differ from for-profit private English conversation schools and exam-focused cram schools catering to Japanese students who lack conversational ability or need to obtain an English qualification.Pauly & Yamane (1999) and Nakamura (2019) explain how parents manage English weekend schools by recruiting students, hiring teachers, and scheduling classes.Since the children speak English, lessons focus on reading and writing, which are not usually taught at home.Nakamura's (2019) interviews with parents from one of the two English weekend schools in this study showed that they are highly invested in their children's English literacy.They regularly read to their children and help with homework.The belief that higher English skills provide better education opportunities for their children underpins their efforts.

THE CURRENT STUDY
As Lee and Chen-Wu (2021) lament, we lack systematic studies on the short-term and longterm effects of weekend schooling on hL development.Hardly any research on language assessment in weekend schools exists.Assessment practices are largely unknown even in large community-based programs (Kondo-Brown, 2021).Knowledge of the weekend schools' impact on hL literacy development is required to understand their importance for minority communities and society.Previous studies on weekend schools were conducted mainly in English-speaking countries that teach Asian languages to immigrant children (Nordstrom, 2022).However, our study of English weekend schools in Japan is in a reversed sociolinguistic setting where English exists as a foreign language and Japanese is the dominant societal language.We examine older simultaneous bilinguals (ages 9 and above) to understand the hL literacy levels that this specific group of bilinguals can achieve over time.We focus on writing as it is a productive skill that takes time to learn (Shanahan, 2006).
This longitudinal study of 17 Japanese-English bilingual children in two English weekend schools in Japan occurred during the COVID-19 pandemic.While regular schools were only closed for several months in early 2020, weekend school continued online from 2020 to 2022.Although COVID-19 has disrupted education worldwide, pandemic effects on hL learning are potentially large because weekend schools already faced huge challenges before COVID-19, such as limited instruction time and a lack of educational resources.Such changes affect not only learning processes but also learning outcomes, so our study seeks to understand how the children's English writing developed during the pandemic.Specifically, our research questions are: i. To what extent are Japanese-English bilingual children attending English weekend schools in Japan maintaining their English writing proficiency during the pandemic?
ii.How did changes in English learning at home and weekend school affect their English writing development?

METHODOLOGY THE SCHOOLS
The sites for this study are two English weekend schools in Tokyo and Yokohama.They were established by English-speaking parents who wanted their children to receive level-appropriate English literacy instruction unavailable in regular schools.Both schools are run by parents with no external support or funding.Parents make the decisions and perform administrative tasks.
Classes are held in rented spaces in public facilities.Each school has about 40 children in four or five classes of different levels.The lowest class is attended by preschoolers, the highest by middle schoolers.Students are grouped by age and ability.Lessons last for 60 or 90 minutes every Saturday for 34−36 weeks a year.Attendees are mainly simultaneous bilingual and bicultural children with one English-speaking parent.Many English-speaking parents are from English-speaking countries, but some are from non-English-speaking countries (see Table 1), thus our use of "home" rather than "heritage" language.The families' diversity contrasts with weekend schools in Western countries, which generally serve one minority group.
Native English-speaking teachers with teaching certification (e.g., the American Board) and experience teaching at international or Japanese schools teach at both schools.The literacyfocused lessons include reading (e.g., phonics), discussions of written texts, vocabulary activities, and writing tasks to acquire proper spelling, punctuation, and grammar.The Tokyo school adopts a content-based approach with a specific language arts, science, or social studies theme for each term, e.g., weather science.US language arts materials are used in class with homework given for children to continue reading and writing on weekdays.The Yokohama school takes a communicative approach to engage the children, i.e., games and activities (e.g., running dictation).Each class has a chat group where teachers communicate lesson and homework content to parents.Some homework is done online with electronic books and vocabulary/spelling games.Children in higher-level classes undertake writing projects where they plan, write, and revise an essay.Teachers assess the children periodically for class placement or promotion to a higher-level class and check homework.However, no formal yearto-year assessment exists.Some older children sit for external English examinations, but not all have had their English formally evaluated.

THE PARTICIPANTS
This study is part of a larger investigation of English-Japanese bilingual and biliterate children in Japan.This paper, however, focuses only on the children who attended either the Tokyo or Yokohama English weekend school between 2020 to 2022.Table 1 shows 17 bilingual children (including three sibling pairs among the nine boys and eight girls) from 14 families.We refer to the children by three-letter pseudonyms and list their ages when they participated in three different periods of assessment.Time 1 (T1) took place before the pandemic, whereas Time 2 (T2) and Time 3 (T3) occurred during the pandemic when weekend school went online.Fourteen children are bicultural with an English-speaking non-Japanese father or mother.Many of the non-Japanese parents come from the US, Australia (AUS), the UK, or Canada (CAN), but SHO's mother, and KAN's and YUK's father are fluent English speakers from Thailand (THA) and Indonesia (IDN), respectively.Three children from two families (HAK, EMM, and KEK) have two Japanese parents with one or both being fluent English speakers.
All children were exposed to Japanese and English from birth.SHO, however, received simultaneous exposure to Japanese and Thai from birth but heard English indirectly from his parents, who used it with each other (Slavkov, 2017, has found that a minority language Nakamura and Quay Journal of Home Language Research DOI: 10.16993/jhlr.50 spoken between parents can support children's multilingualism).His English is currently much stronger than his Thai.Six children presently hear one language from each parent (OPOL).Four children hear more Japanese with an English-speaking parent also using some Japanese (More JPN).Three receive more English input at home with a Japanese parent also speaking some English (More ENG).Another four children hear their Japanese and non-Japanese parents use both languages (JPN&ENG).
The families are middle class; all English-speaking parents have at least an undergraduate degree, and three have advanced degrees.Many use English in their jobs as teachers, writers, and editors.Most Japanese-speaking parents also have an undergraduate degree; some completed high school or junior college.Sixteen children were born in Japan and have attended school here since Grade 1.The only child born in the US, LET, returned to Japan at age six to begin Grade 1 after learning Japanese at a weekend school in the US.Another child, AMS, lived in the UK from ages 3−6 but also returned to Japan for Grade 1.Given that all children are simultaneous bilinguals who started Grade 1 in Japanese schools, we do not assess their Japanese literacy, confirming instead that all have Japanese grade-level abilities through their kokugo (Japanese language) school grades.
The average length of weekend school attendance is 63 months.LIS has attended the longest (120 months), whereas LET has only attended for 18 months.Note that the four 9-year-olds in T1 and two in T2 will not have English reading lessons in elementary school yet (MEXT, 2022), while the five 10-year-olds at T1 will have just started in Grade 5.However, these 11 children started reading and writing English for 1.5−7.5 years at weekend school before receiving English instruction in elementary school.

ASSESSMENT AND DATA COLLECTION
The test objective and longitudinal assessment on children of different ages and proficiencies.In using the normative scores with seven descriptive terms ranging from Very Poor to Very Superior, our participants' writing performance is inevitably benchmarked against same-aged US children (e.g., the descriptive term Average indicates that a participant is performing in this range as established for a same-age US child).We emphasize that these descriptions are solely for evaluative purposes.The scores are only used to explain our participants' writing performance.Specifically, the age-based scores help us to compare our 17 participants ranging in age from 9−15 (see Table 1) with each other and over time.As we informed the parents, children attending English weekend schools in Japan are not expected to possess the same level of literacy as their US peers.In the absence of an instrument designed specifically for Japanese-English bilingual children, we adopt the TOWL solely for determining the range of their writing abilities over time and for in-group comparisons.
The TOWL examines "contrived" and "spontaneous" writing across seven sub-tests."Contrived" writing tests five discrete aspects of written discourse, i.e., vocabulary, spelling, punctuation, logical sentences, and sentence combining.In "spontaneous" writing, the child writes a story in response to a stimulus picture.The story is evaluated for the use of contextual conventions (e.g., spelling, punctuation, and paragraphs) and story composition (e.g., the plot and sequence).
Assessments were based on procedures in the TOWL examiner's manual, i.e., Hammill & Larsen (2009).Before the research, we underwent training as prescribed by the manual by practicing the scoring of ten sample essays, calibrating our scores to that of the scoring keys, and doing trial assessments on four children.
Before the children were assessed, we explained the scope of our research to their parents and obtained written consent for conducting the TOWL.Written consent was also obtained from parents and teachers who participated in our questionnaire surveys and interviews.The Research Ethics Committee of the second author's institution has approved our research procedures (no.2020-30).The children took the 90-minute assessment individually with one of us before or after their weekend lessons.Table 1 lists 13 children assessed in person before the pandemic in T1.An additional four were assessed in T2 after the start of the pandemic.There was a 12-month interval between each assessment.However, the interval was longer for ERS and SHO (14 and 17 months) because of their busy schedules.Nine children were assessed three times (T1−T3).
The third assessments of four children (LET, ERS, JAS, MOH) who quit weekend school in T3 are not included here.Four children only participated from T2, so they were assessed twice (T2−T3).
Assessments in T2 and T3 took place when weekend school was online.
There are two different TOWL forms, A and B. Both have seven sub-test components with different test items.They were used alternately with Form B for the first assessment, Form A for the second one, and Form B again in the third assessment.The use of Form B in T1 and T3 gives us a clear picture of how the children's writing developed as they had to write a story based on the same picture after two years.Although T1 assessments were conducted in-person, assessments in T2 and T3 were conducted via a teleconferencing tool with some improvisations for online testing.The response booklet was mailed to the family in advance with instructions to keep it sealed until testing started.During the session, the examiner monitored the child's writing via camera.Immediately after the session, the parent was instructed to return the response booklet in an enclosed prepaid envelope.We graded each child's writing separately according to the examiner's manual.Scores were compared and discussed for all 136 test items before being finalized for each child.Inter-rater discourse (as advocated by Matthews, 2021) is thus a crucial component of our study for ensuring reliable assessment results.

Questionnaires and interviews
To understand the impact of the COVID-19 situation on English learning, we asked parents to respond to a questionnaire (see Appendix) in T2 and T3.All except one of eight questions are open-ended to allow parents to describe freely changes to their children's English learning in the past year.Parental responses were coded thematically using MAXQDA.These codes were analyzed to uncover common themes related to the children's English learning.Interviews were also conducted with two teachers from each school.The Yokohama teacher was interviewed before and during the pandemic.The Tokyo teacher was interviewed before the pandemic and provided email responses to our queries during the pandemic.These interview transcripts were reviewed to support the findings from the questionnaire data.

WRITING ASSESSMENT RESULTS
We use the overall writing composite index score to analyze the children's TOWL performance.As a composite of scores from all seven subtests, it is the best estimate of general writing ability (Hammill & Larsen, 2009 Quay and Nakamura (2022).
Table 2 shows the children's overall writing scores for T1, T2, and T3, grouped according to their longitudinal performance.Five children whose scores progressed to a higher score band from T1−T3 are in the first group.Three of them greatly increased their scores on the same Form B test taken at T1 and T3: LUH improved by 12 points from Above Average (119) to Very Superior (131), EMS by 13 points from Superior (123) to Very Superior (136), and JOG by 15 points from Poor (79) to Average (94).JOG's performance was Poor in T1 as he only wrote a few sentences describing a picture.However, his scores reached the Average range in T2 when he wrote a long paragraph with many sentences.JOG's T3 scores were also Average but slightly lower than T2 because his spelling did not improve.The performance of AMS and EMM, tested during the pandemic at T2 and T3, also improved from Average to Above Average.
The second group consists of ten children whose scores slightly fluctuated but mostly remained within the same level.LIS, notably, is an exceptional writer whose scores were Very Superior over three years.Seven children consistently performed within the Average range, scoring between 90−110 for each assessment period.LET and ERS had slightly lower scores in T2, but remained within the Average range.LET's scores improved for spelling and punctuation in T2 but decreased in other subtests.ERS had lower T2 scores due to weaker performance in some subtests, particularly logical sentences.KEK was also weak in this subtest in T3, resulting in a 3-point drop in T3 from T2.HAK is the only child whose overall score fell to Below Average in T2.Her higher spelling scores were offset by lower scores for logical sentences and sentence combining in T2.However, she had an 18-point increase back to Average in T3 when her spelling and story composition skills further improved.Given that many children in our study maintained their ability at the Average level over time, we provide an example of a child's writing from this group.KAN's scores increased by six points from 102 in T1 to 108 in T3 when the same Form B was used.At home, KAN receives a mix of English and Japanese input from both her Indonesian father and her Japanese mother, so these results suggest that high English literacy skills do not depend on having a parent speaking English exclusively nor coming from an English-speaking country.Writing excerpts illustrate how KAN's writing improved.Figure 1 shows the first half of a story KAN composed in T1, which is based on a picture of a woman and her children standing in front of a burning tree in a thunderstorm.Her story contained paragraphs and various punctuation forms.It went beyond describing the picture, and depicted strong emotions in the story's characters.However, it contained spelling and grammar errors.In T3, KAN wrote a different longer story based on the same picture (cf. Figure 2).It involved a well-sequenced dialogue between three characters that built up to a fire scene.KAN used a variety of compound sentences, introductory phrases, and punctuation forms and made fewer spelling and grammar errors than at T1. Her Average score for 12-year-olds at T3 illustrates an improvement in her writing.The third group is two children (SAK and JAS) whose scores fell to a lower band.SAK demonstrated Superior writing ability in T1 with a score of 123, but fell to Above Average in T2 and T3.Although SAK's vocabulary, spelling, and punctuation scores increased from T1−T3, her ability to combine sentences and apply contextual writing conventions did not improve much.Her mother commented that SAK's writing has "stalled" and that "it is harder for her to write more complex reports."SAK's writing was already at the Superior range in T1, so maintaining this high level by hearing English at home and taking online lessons proved difficult.JAS's scores dropped 13 points from Average to Below Average from T1−T2 due to his weaker performance in the logical sentences and sentence combining subtests as for HAK and SAK.These two subtests (out of five in the contrived writing section) have a ceiling where testing stops once three consecutive errors are made.Unfortunately for HAK, SAK and JAS in T2 and SAK in T3, early errors prevented further scoring.

TEACHERS' AND PARENTS' PERSPECTIVES ON WEEKEND SCHOOL LEARNING DURING THE PANDEMIC
Learning the hL in weekend school changed drastically during the pandemic with online lessons continuing for more than two years.However, the teachers at both schools endeavored to recreate a similar learning experience for their students online.The Tokyo school teacher continued following a theme for each semester.Students read and discussed a chapter from an assigned book every week.Comprehension questions completed for homework were discussed in breakout rooms.Various learning programs (e.g., spelling apps and e-books) already in use before the pandemic supplemented online lessons.New technology (e.g., digital whiteboards) was also introduced.The children learned how to prepare and do slide presentations online.Despite the technological tools, the teacher found online teaching challenging: 1. "Online teaching does present its own problems.It is more difficult to hold whole class discussions, and as a teacher it is difficult to do on-the-spot assessments or give quick feedback." The teacher at the Yokohama school also strived to keep online lessons interactive and interesting.Besides reading, she introduced new vocabulary activities, e.g., virtual board games.Older children also did mini projects in breakout rooms.She continued giving speed writing tasks and checked the children's writing via camera.These improvisations helped the children to concentrate as MOH's mother observed: 2. "I can see Zoom isn't easy for him to stay focused.I catch him doing other things often during a class, but it seems that the more games the teacher plays, the better he does." Many parents appreciated the teachers' efforts in teaching online when no other alternative was available, as expressed, for example, by LUM's mother: 3. "I'm happy that we have been able to continue our English learning during the pandemic, and grateful for the efforts our teacher has put in to continue lessons during this time.I think it is extremely difficult to have young kids engaged and learning online every week.Obviously, in-person lessons would be preferred, but I'm happy with the way we have managed to continue learning English despite being online." According to the Yokohama teacher, parents are present and more involved during online classes.LUM's mother feels that online classes cover more material than in-person classes.HAK's mother became more familiar with weekend school homework because she printed it out.The parents of EMS and SAK feel the online classes have honed their children's reading and writing skills.ERS's mother observes that online lessons provide more opportunities for informal written communication between classmates: 4. "The school is now via Zoom, and several students chat using Line or Messenger, texting only in English.So, while formal study has lessened, chances to write in a real-time setting as a means of communication have increased." By learning online, the children acquired typing and texting skills.The online classes also provided children with valuable English interaction with interlocutors besides their parents.
During the pandemic, SHO, LUH, JOG, and LET learned more advanced content when they were promoted to higher-level classes.The online format did not affect attendance, as many parents reported that their children attended class regularly.Although attendance was not an issue, staying engaged was a challenge, particularly for children who had difficulty concentrating, such as in MOH's case (Comment 2).The Tokyo teacher reported that children with focus issues generally adapted poorly to online lessons and quit weekend school.The lack of social interaction is another concern.The parents observed that the children previously enjoyed weekend school because they could meet their friends.
The school was a safe space that created a sense of identity, belonging, and community for the mostly bicultural children and their English-speaking parents, but this "space" was difficult to maintain online.Unlike ERS (in Comment 4), not all children communicated well with their classmates using electronic means.Some children, like SHO, gradually became less enthusiastic about online classes as his father noticed: 5. "At the beginning, my child was thrilled to have online English lessons as it was a new experience.But gradually, his enthusiasm dropped.He turned off the video camera and only used voice." ERS also became less motivated about homework and no longer looked forward to weekend school according to her mother: 6. "I'm grateful that we are still able to learn online, but I know that it also has a lot of limitations, and it's become more of a chore, whereas before, she loved going to weekend school on Saturdays.Our printer is old, and it's a pain to print out the materials sometimes, and ERS doesn't care if she completes it or not.We are on the fence about continuing next year because it is the personal interaction that makes it fun and meaningful."

PANDEMIC EFFECTS ON HOME LANGUAGE FAMILY PRACTICES
Due to COVID-19, many families could not take their annual family trips abroad, which previously provided additional English exposure, as expressed by JOG's father: 7. "I don't think COVID-19 has negatively affected his learning of English other than the inability to travel to America and stay for one month this past summer.This one month of intensive English exposure each year I have found to have been a great boost to JOG's English.That said, our several months alone together at home during the state of emergency had a beneficial impact on JOG's English conversation and listening comprehension." However, despite not meeting English-speaking family and friends, 11 of 14 parents reported that the pandemic did not affect their children's English.They compensated by arranging for virtual reunions (LIS and EMS), traveling within Japan together (JOG), hiring a personal English tutor (KAN) or starting new hobbies that involved English (JAS, EMS, and SHO).School closure during the early stages of the pandemic in 2020 provided more time for speaking English at home (Comment 7), reading English books (EMM), and preparing for Eiken (a popular four-skill English examination in Japan) as described by LIS's mother: 8. "To be honest, the COVID-19 pandemic was the best possible thing that could have happened for LIS's English!She suddenly had free time that she spent reading, writing, and talking entirely in English.She also studied hard for Eiken (memorizing difficult vocabulary, writing practice essays, reading more non-fiction, completing practice tests) and successfully passed Level 1 this year." Ten children prepared and sat for the Eiken during the pandemic.Test preparation, which included learning vocabulary, answering reading comprehension questions and writing argumentative essays, likely strengthened their English literacy skills.
Other literacy practices were also implemented.Reading activities, mentioned more often than writing, appear easier to do at home.Some parents read more with their children (EMS and JOG) and introduced new books to pique their children's interest (ERS and EMS).JOG's father asked more questions during joint reading to confirm comprehension.As for writing, some parents provided workbooks for extra practice (MOH and JAS).Based on their parents' recommendations, SAK kept a daily English journal, and LIS wrote book reviews which she shared online.These home literacy practices likely contributed to the children's English writing ability.Other home literacy practices during the pandemic included Eiken test preparation, which is not offered at the weekend schools.It was common for older children to take the Eiken before the pandemic, but COVID-19 created more time for this purpose as school and extra-curricular activities were suspended.Ten children studied for the Eiken in T2 and T3 (e.g., LIS in Comment 8) and passed the test.Seven obtained the Eiken Level 2 qualification or higher (intended for Japanese high school graduates and university students), thus confirming their good English skills.With an Eiken qualification, one could apply to better middle or high schools, the goal of many older children.Although formal English education in the Japanese school system also fosters English literacy, our participants achieved a much higher level of proficiency at younger ages than Japanese students.Test-taking converts their higher English literacy skills into "capital for the future" (Nordstrom, 2022).While English skills can be applied to one's future studies or career, skills in other hLs hardly offer such benefits in Japan.Biliteracy involving languages other than English is uncommon.However, as uncovered in this study, home language practices play a crucial role when minority children achieve grade-level abilities in both Japanese and their hL (cf.Majima et al., 2019).

CONCLUSION
In answer to our first research question, many bilingual children maintained or improved their English writing proficiency during the pandemic.We discovered for our second question that the use of educational technology in weekend school before the pandemic eased the transition to online learning.Children acquired techno-literacy skills by learning online.However, parents lamented the lack of social interaction, which weakened the children's enthusiasm for weekend school.The pandemic freed time for home literacy practices as the children could read more in English and prepare for English examinations.These findings indicate that changes in English literacy learning at home and weekend school did not negatively impact the children's writing development.
Despite these encouraging results, this is a small-scale study with limitations.Since not all Japanese-English bilingual children attend English weekend school, and not all attendees were assessed, our results only show the English literacy levels that some children can achieve.Nevertheless, our findings demonstrate the importance of home literacy practices and weekend school in developing biliteracy in late childhood and early adolescence.Unlike most studies of "heritage" language learners, our focus on "home language" (hL) expands the field of investigation to include families whose hL is not necessarily a "heritage" one as exemplified by four out of our 14 families.We showed how the recognition of English language ability (through Eiken qualification) in the Japanese education system encourages hL learning in English-speaking families.Other foreign languages (e.g., Korean, Chinese, Vietnamese, and Portuguese used by the largest immigrant groups) also need similar recognition to thrive.However, even a prestigious hL like English does not ensure successful biliteracy.Both parents and children must value the hL and envision its use in the future.This study also addressed the effectiveness of weekend schools through assessments of children who attend them, a crucial methodological consideration thus far lacking in the literature (cf.Kondo-Brown, 2021;Lee & Chen-Wu, 2021).Further work in this strand of research can help position these grassroots institutions as important and necessary for nurturing bilingualism and biliteracy.

APPENDIX: OPEN-ENDED QUESTIONNAIRE FOR PARENTS
1. How has your child's English learning at home changed from the previous year?
2. How has your child's English learning at the weekend school changed from the previous year?
3. How have your efforts to teach your child English changed from the previous year?

4.
Are there any new initiatives to promote your child's English learning this year?

5.
Have the writing assessment results from the first year helped you teach English literacy in the home?
Both the Tokyo and Yokohama schools switched to online teaching in March 2020 due to the pandemic.Although regular schools resumed in-person classes soon after initial closure, the weekend schools continued online because the public facilities hosting the classes remained closed.Online classes were also preferred to avoid COVID-19 risks.They were held during the usual lesson times via a teleconferencing tool.Younger children in the Tokyo school started attending in-person classes in January 2022, but the older children in this study were still taking online lessons in late 2022.Although the Yokohama school adopted a hybrid format where younger children could attend in-person or online from April 2022, many children continued online.The older children in Yokohama continued learning online with one monthly in-person class.