Methodology

The data in this book are primarily from two research projects in which I was the Principal Investigator. In both the research projects, I had the good fortune of working with a team of bilingual research assistants. Thus during the course of this book, I refer the whole team as “we.” The first project, Building English Competencies in Bilingual Underachievers: A Baseline Study of Singapore’s Learning Support Program (OER 28/08 VV), was an investigation of pedagogy in Singapore’s Learning Support Program. The second project, “Use of First Language in Teaching and Learning Chinese and English” (OER16/11VV), attempted a Proof of Concept in the Learning Support Program (LSP), which entailed using a translanguaging approach to teach reading in English. The Proof of Concept also involved attempting to use English to teach Chinese in Chinese Mother Tongue classes; however, this part of the Proof of Concept is not included in this book. The baseline study of the LSP was completed in 2008 and the Proof of Concept in 2016. For easy reference, I will refer to the first study as “the baseline study” and to the second one as the “Proof of Concept.”

In the both the research projects, the classes in the Learning Support Program are referred to as “multilingual English classrooms.” This phrase, which is also in the title of the book, is meant to be an inclusive term that captures the linguistic diversity in Singapore’s classrooms. It refers to classrooms in an English medium school system where in one class there could be children from 4–5 different linguistic backgrounds. In the brief literature review provided below, I discuss the research of scholars who support bilingual language teaching in content areas (like science and math) and also research which is generally on including the home language of the child in the school system. Though the learning support classes analyzed in this book only teach English, I have also discussed scholars who recommend bilingualism to teach content as this could be a resource for Singapore, where all the core subjects are taught in English. There are numerous subject classes in Singapore, e.g., math and science classes, which are equally diverse regarding the language background of the children.

Monolingual vs Bilingual Pedagogy for ELLs

Garcia and Kleifgen (2010) review relevant literature to conclude that bilingual approaches, in which the child’s home language is used, are more effective in teaching students to read than English-only approaches. In fact, learning to read in the home languages promotes reading ability in English. They write: “what is evident from this research is that the use of the students’ home language is crucial for their long term cognitive growth and academic achievement in English.”

It would not be efficient to refer to all the literature that Garcia and Kleifgen (2010) have already reviewed. However, one study merits a fuller discussion. In one of the largest-scale longitudinal studies, involving 210,054 students, Thomas and Collier (2001) described the variety of education services provided in the USA for language minority students and the long-term academic achievement of students in each of these types of programs. They focused on eight major program types:

  • 90–10 two-way bilingual immersion (or dual language)

  • 50–50 two-way bilingual immersion

  • 90–10 one-way developmental bilingual education

  • 50–50 one-way developmental bilingual education

  • 90–10 transitional bilingual education

  • 50–50 transitional bilingual education

  • English as a second language (ESL) taught through academic content

  • English mainstream

The way Thomas and Collier (2001) tracked the impact of language distribution in preschools on later academic achievement is noteworthy. They found that “English language learners immersed in the English mainstream because their parents refused bilingual/ESL services showed large decreases in reading and math achievement by Grade 5, equivalent to almost 3/4 of a standard deviation (15 NCEs), when compared to students who received bilingual/ESL services”. Thomas and Collier (2001) explain that a difference of 5 normal curve equivalents (NCEs) is a major difference that requires action on the part of the government. In fact, they found that both a 50–50 distribution of languages in preschool and a 90–10 distribution of languages (starting with 90% home language and 10% English leading up to a 50–50 distribution) were better for students’ later English language achievement scores.

Thomas and Collier’s (2001) is a longitudinal study evaluating the impact of bilingual vs monolingual instruction in preschools. The LSP, on the other hand, is a program starting in grade 1, and we do not have data on the type of preschools that the children attended before they entered primary school. We do know, however, that preschool attendance in Singapore is extremely high: according to a Lien foundation report, 99% of 6-year-olds who come to grade 1 have had at least 1 year of preschool education. At the same time, according to the same Lien foundation report, the quality of preschool education in Singapore is not as high as that in other developed countries like Finland and New Zealand. In a list of 45 countries ranked on the basis of 9 variables that define high quality in early childhood education, Singapore ranks 29th (Starting Well 2012). More importantly, not much is known about language distribution between English and Mother Tongue in preschools in Singapore.

Even in the later stages of a bilingual child’s academic career, scholars recommend an approach that involves translanguaging (Celic and Seltzer 2011; Lin and Lo 2016). According to Celic and Seltzer (2011), translanguaging is an important pedagogic approach through which ELLs can access the common core curriculum in the USA. They emphasize that translanguaging is not merely a simple scaffold that can be removed as soon as the child becomes proficient in English. On the contrary, translanguaging is a sustainable practice that builds academic achievement and celebrates the identity of minority students. Their guide gives teachers of elementary, middle, and secondary school classes ideas regarding how to incorporate translanguaging in various disciplines. Their contention is that translanguaging aids in academic achievement. Specifically, translanguaging “offers bilingual students the possibility of being able to gather, comprehend, evaluate, synthesize and report on information and ideas using text-based evidence” (Celic and Seltzer 2011).

Similarly Lin and Lo (2016) emphasize the importance of dialogic interaction and translanguaging by analyzing data from two disparate grade 10 science classrooms. In these classrooms, the students, whose L1 was Chinese, were learning science in and through English, their L2. The teacher who used Chinese to connect scientific knowledge regarding “food substances” with the students’ everyday world was deemed more effective than the teacher who used only English to teach content. The more effective teacher also used triadic dialogue (IRF instead of IRE) to elicit languaging from the students leading to a more interactive and engaged class.

Though the literature supports bilingual education at the preschool (Thomas and Collier 2001) and later (Celic and Seltzer 2011) stages in a child’s academic career, this is not the case in Singapore. In Singapore the approach to teaching reading in learning support classes not only assumes that the students and the teacher are monolingual English speakers, it is also based on assumptions of Second Language Acquisition. The approach ascribes an L1 and L2 to each child which not only essentializes hybridity but assumes that the home language of the child causes interference in English acquisition resulting in fossilized errors which must be removed through skills training. The reality is that most of the children in the LSP are simultaneous bilinguals and so are many of the teachers though their proficiency in each of their languages varies.

In a study of 30 fifth grade Spanish-English bilingual students working on a joint activity, Martin-Beltran (2010) raised exactly this point. The author pointed out the though SLA also looks at interaction as a way of language learning, the focus is on one-way learning of the target language and not two-way learning of more than one language. SLA conceptualized L1 and L2 as separate, sequential, and linear which is a problem in globalized cultural spaces like Singapore where individuals are simultaneous bilinguals. Though Cenoz and Gorter (2011) point to a new trend in SLA in which “the focus on the language per se has shifted to an increasing interest in the learner, the communicative interaction, and the context in which the interaction takes place” (pg. 357), this turn has yet to reach Singapore’s classrooms.

Given the metrolingual languaging practices of young people in Singapore, the LSP needs an approach based on dynamic bilingualism or translanguaging. Metrolingualism is a term proposed by Otsuji and Pennycook (2010) to describe “the ways in which people of different and mixed backgrounds use, play with and negotiate identities through language; it does not assume connections between language, culture, ethnicity, nationality or geography, but rather seeks to explore how such relations are produced, resisted, defied or rearranged.” Singapore’s language in education policy tends to essentialize the ethnic group, language, and culture of students, whereas the globalized linguiscape of Singapore displays immense syncretism. Children in Singapore’s schools come from homes where multiple languages are spoken in diverse registers by adults who themselves could be of mixed parentage. Also, as pointed out in Chap. 2 the nationality of one parent may not be Singaporean. These aspects lead to a “hyperdiverse” classroom which, as the later chapters will show, pose a challenge to the design of a translanguaging program.

The Research Projects:

In order to explore current pedagogy in the LSP and thereafter recommend changes to it, I initiated two research projects:

  1. 1.

    A baseline study of pedagogy in the LSP

  2. 2.

    A Proof of Concept to change pedagogy in the LSP

Five schools, listed in the first column of Table 3.2a, volunteered to participate in a baseline study of pedagogy in the LSP, which commenced in 2008 and was completed in 2011. A total of 19 hours of video data were collected from LSP classes in these five schools.

Table 3.2a Summary of classroom observations in monolingual classes
Full size table

Table 3.2a documents the observations that were conducted in typical LSP classes.

In the second research study, we attempted a Proof of Concept using translanguaging in three other primary schools. Titled “Use of First Language in Teaching and Learning Chinese and English” (OER 16/11VV), this project analyzed the use of Mother Tongue in teaching of English in LSP classes, and the use of English in the teaching of Chinese. For the purposes of this book, only data from the three schools in which LSP classes were observed have been analyzed. These schools have letter names (Schools W, F, and C) to distinguish them from the five schools in the baseline study. In schools W, F, and C, a new method was used called “Raise the BAR (Bilingual Approach to Reading)” in which Malay and Chinese were used judiciously and systematically to teach vocabulary, grammar, and discourse in English.

Before I move forward, a definition of the term “Proof of Concept” and how this is different from a formal intervention is imperative. Proof of Concept is a term used ubiquitously in the hard sciences like engineering but rarely in applied linguistics. It refers to a pilot study, essentially exploratory in design and outcomes, conducted to test out a concept. The advantage of a Proof of Concept is that in terms of design, its boundaries are more fluid than the boundaries of a formal intervention. For instance, a formal intervention would require a pre- and posttest which, given the nature of the children in the LSP, might not tell us much about the learning of these students. All children in the LSP are low achievers, and some even have special needs. Standardized testing for this cohort in reading will result in low numbers which would not add more to what we already know about these children.

My research design for the Proof of Concept was more organic, collaborative, and bottom-up and attempted to test the concept of translanguaging with low achieving students. The research design emerged as a result of repeated conversations with the principals and LSP teachers of schools W, F, and C. All three schools were keen to try translanguaging with their LSP students though each had slightly different concerns. For instance, in school W, since the majority of the students were of Malay ethnicity and Mdm Yati was a fluent Malay-English bilingual, we decided to focus on the Malay students. Also, Mdm Yati was confident about improvising regarding when and how she would bring in Malay to teach English. Also, Mdm Yati had been informally using Malay with her LSP students to help them learn English. On the other hand, Ms. Selene and Ms. Angela were entirely new to translanguaging pedagogy. They wanted to see some videos of how translanguaging was done before they attempted this new pedagogy. Also, though Mdm Yati preferred to use the curriculum she was already using, Ms. Selene and Angela were keen to use the Ginn readers that we recommended. Thus for each of the schools, we tried to customize the research design in an organic and collaborative way while, at the same time, preserving the spirit of pedagogical translanguaging.

Table 3.2b gives details of the observations conducted in the three schools where translanguaging through Raise the BAR was introduced in the LSP classes. This project started in 2012 and was completed in 2015. A total of 13.9 h of video data were collected from three primary schools.

Table 3.2b Summary of classroom observations in schools where translanguaging was used
Full size table

In total 33 h of video data, 18 h from the baseline study, and approximately 14 h from the Proof of Concept, have been analyzed in this book.

The participants included 8 teachers, 5 in the baseline study and 3 in the Proof of Concept, and approximately 75 students. Regarding the number of students, I can only provide an approximate number as there was a high level of absenteeism in the LSP classes, and in most of the classes, there were a few children who had not come to school. For instance, in School W, the pilot school, there were six students in the LSP class. However on days 2, 4, 5, and 8, there were only four students in class, and on day 7 there were five students. This kind of pattern was quite common in schools F and C also. All the students are in the LSP after they failed a reading test conducted by Singapore’s Ministry of Education; thus they are all similar in terms of their low proficiency in reading in English. All the teachers in the LSP are trained by the Ministry of Education on how to teach reading skills to low proficiency students.

The LSP classes are hyperdiverse, meaning that in each class there are children whose homes are Mother Tongue dominant, children whose homes are Singlish/English dominant, and children whose homes are truly bilingual in that they hear Mother Tongue and English 50% of time. In addition, due to the high rate of marriages with foreigners in Singapore, many of the children in the LSP have one parent who speaks very little English. According to the Department of Statistics in Singapore, 36% of marriages in 2016 involved a citizen of Singapore marrying a foreigner (https://www.msf.gov.sg/media-room/Pages/Statistics-on-transnational-marriages.aspx). The implication of this statistic is that if the mother of the child is a non-Singaporean, for instance, from China, Malaysia, or India, this will change the linguistic ecology of the home. We do not have any survey data on the home languages of the children in the LSP class. A survey of the parents in the LSP was not possible due to budgetary and ethical concerns. Thus the information about specific students’ home language is from comments made by the teacher. As each class is extremely small, the teachers did know each of the students well and had met some of the parents.

Piloting Raise the BAR

I begin with the procedure undertaken in school W, as this was the pilot school for “Raise the BAR.” School W is in an area with a Malay majority and most of the students in this school are of Malay ethnicity though there were a few students of Chinese and Indian descent. Many of them come from homes and communities which are highly Malay dominant. The LSP class in school W that was selected for piloting Raise the BAR had five Malay and one Chinese student.

Since the LSP teacher, Mdm Yati, was fluent in Malay and English, she volunteered to lead the group of Malay students. Ms. Cindy, a Chinese-English bilingual, would teach the sole Chinese student in this class in a separate room. After numerous meetings with the staff it was collaboratively decided that the child’s dominant language would be used as a scaffold to aid comprehension and teach key vocabulary items during the reading of Loughead’s (2006) story: The Grasshopper and the Ant. Since this was the first school where we were trying out a bilingual approach, we did not have videos on the basis of which we could train Mdm Yati. However, Mdm Yati expressed confidence in using a translanguaging approach as she was already using Google translator with some of her Chinese students in other LSP classes. Her view was that she would like to learn a systematic approach of using L1 to teach English. In an earlier part of this chapter, I have expressed concerns about the use of the terms L1 and L2 with simultaneous bilinguals. In school W, Mdm Yati, after 15 years of teaching experience, thought that the majority of children could be described as L1 Malay and L2 English. Since Mdm Yati wanted the flexibility of improvising for her students, we left it up to her as to when and how the L1 would be brought into the classroom as a scaffold. The Grasshopper and the Ant was a familiar text for the students as they had encountered it in their mainstream English class. During the course of the observations, the text was read to the students multiple times.

Implementing Raise the BAR

Schools F and C had a mix of mainly Chinese and Malay students with the exception of one class where there were two Indian students. In these schools, according to the teachers, there were some children who were English dominant though the majority of children came from homes where they mainly heard Mother Tongue. In collaboration with the school, we decided to focus on only the Chinese and Malay student groups. For the duration of Raise the BAR, the two Indian students were assimilated into other LSP classes. Thus there are no Indian students in this study.

In all three schools, we adapted lesson plans, prescribed by Singapore’s Ministry of Education, with ideas from the CUNY-NYSIEB (City University of New York-New York State Initiative on Emergent Bilinguals) framework and discussed this with the school and teachers before they agreed to execute these lesson plans. Developed by researchers in the City University of New York, the CUNY-NYSIEB is a guide regarding how translanguaging can be used in the classroom with bilingual children who have low literacy in the medium of instruction (Garcia et al. 2013). Specifically, we wanted to “scaffold the instruction for these students, ensuring that there is ample oral work and discussion of the academic concepts” (Garcia et al. 2013).

In schools F and C, the Proof of Concept ran for approximately 1 month during which four books were taught to the class with 1 week dedicated to each book (See Appendix for the list of books used in schools F and C). In both schools, the same literacy practices were conducted for about a month in 2013. In the first instance, the teachers identified a set of books they wanted to teach the class during the Proof of Concept. For each of these books, the research team and the LSP teacher together prepared detailed lesson plans which involved teaching key lexical items and targeting comprehension. All the books were familiar to the students. A set of leveled questions was prepared by the research team that the teacher or the Principal Investigator asked during whole class elicitation. The research team discussed these lesson plans with the teachers and revised the plans according to their suggestions. In keeping with the advice of the teacher, the Principal Investigator also taught some of the classes. The Principal Investigator, in this part of the methodology, was inspired by the benefits of dialogic reading during shared book approach (Lonigan and Whitehurst 1998). Though the shared book approach is a method used mainly with monolingual children, in Raise the BAR, it was customized to suit the needs of a bilingual cohort. Table 3.2c summarizes the training procedures and literacy activities in Raise the BAR.

Table 3.2c Training and literacy activities for raise the BAR
Full size table

Typically the teaching of each book involved 1 week of classes starting on a Monday and ending on Friday. During the Monday class, taught by the Principal Investigator in English, the children were asked factual and inferential comprehension questions. Though all eight books were familiar to the students, they were still confused about who the main characters were and what the plot was. The purpose of questioning was for the PI to ascertain the level of comprehension the students possessed.

Tuesdays and Wednesdays were bilingual days when translanguaging was encouraged. On these days a huge digital poster was flashed on a power point showing a child’s head and “Mother Tongue ON” written on the side. The moment the children entered the classroom, they saw this large poster and knew that this particular LSP class was different from their usual classes and that in this class they could use their Malay or Chinese along with English. A bilingual approach was used to teach the more difficult vocabulary, grammatical structures, and comprehension questions. At these junctures the class was divided into smaller groups based on whether the children came from Malay- or Chinese-speaking homes. Each group was led by a bilingual teacher, one of whom was the LSP teacher herself. In both schools the LSP teacher was Chinese and she led the Chinese-English bilingual group. The Malay-English bilingual group was conducted by the lead research assistant for this project who is fluent in both languages. On each Thursday, the Principal Investigator gave an oral demonstration of the how the story could be retold in English only. On this day the digital poster with “Mother Tongue ON” was not flashed on the power point. The practice of modeling was requested by the LSP teachers who felt that since this literacy task was unfamiliar and difficult for their students, the Principal Investigator should first offer a demonstration. The demonstration/performance was set up in the form of a picture elicitation task in which the research team used digital pictures flashed on a power point as a stimulus. For each storybook, we deleted the prose text and converted each page into a digital picture. On some of the digital pages, we inserted the key vocabulary items that the children were taught during the reading of a specific story. To support the oral retelling, the Principal Investigator also used props like moustaches, glasses, etc. and used her background in theater to entertain and engage the students.

Raise the BAR was profoundly disruptive. Language learning in Singapore, despite a bilingual education policy, is based on the concept that the bilingual is two monolinguals in one. Both the English classes and Mother Tongue classes are largely conducted without any hybrid language practices. Raise the BAR disrupted the traditional approach to language learning and introduced translanguaging as a new pedagogy. In her introduction to this book, Lin has commented that “Translanguaging pedagogies can be differentiated as spontaneous or planned. However, these are best conceived as lying on a continuum rather than as strictly binary options” (pg -1-). Our experience during the Proof of Concept substantiated this comment. Though we came into the schools with plans regarding how translanguaging could be systematized, the teachers took it upon themselves to choose the moments when the distribution of languages in the class could be altered. As this book will show, though some of these decisions were spot on, some did not result in the desired learning outcome.

Garcia and Kleifgen (2010) explain that development in a bilingual is not linear but dynamic. The difference is that in a linear conception of bilingualism, languages are autonomous units in a bilingual’s brain. However, in actual fact, the brain of a bilingual is like an all-terrain vehicle (ATV). An ATV can have three or four or more wheels and can be maneuvered over many types of terrains. Thus an ATV is different from a monocycle or a bicycle. In other words the brain of a bilingual has been conceptualized like a bicycle with the two wheels representing two languages whereas it functions more like an ATV. Gracia and Kleifgen (2010) recommend that “Effectively educating emergent bilinguals, even in programs that teach through the medium of English, must include and support the dynamic bilingual practices by which bilinguals construct knowledge and understandings.”

The authors discuss the differences between code-switching and translanguaging by pointing out that translanguaging includes code-switching. However, there are certain practices within the broader term, translanguaging, that are not available in the relatively limited term, code-switching. For instance, the literacy practice where the teacher in explaining a concept in English and the students are taking notes in another language cannot be called code-switching. This literacy practice is clearly one that demonstrates the nonlinear development of biliteracy in two languages and comes under the broader term translanguaging. During the course of data analysis, I will take this discussion of the difference between code-switching and translanguaging further. I will demonstrate that though methodologically there are many limitations in the way code-switching has been used to analyze bilingual transcripts, a few basic methods are still relevant and that the concept of translanguaging is open to more hybrid methodologies of coding.