An investigation into a communicative approach to English language teaching in governmental and nongovernmental primary schools in Bangladesh

In recent years, the government of Bangladesh has worked to implement a language policy to foster a communicative approach to English language teaching (ELT) beginning at the primary level in the Bangladeshi educational system. Since such a top-down decision was made, no studies have been conducted to substantiate the presence of a communicative approach in Bangladeshi English classrooms. The aims of this study are thus (1) to probe to what extent in-class English teaching and learning practices reflect a communicative approach to ELT; more specifically, this thesis motivates "indicators" used in the documentation of "present realities" in governmental and BRAC primary school English classrooms, (2) to assess the communicative ability of Bangladeshi students enrolled in these schools, and (3) to examine the relationship between the aforementioned indicators and students? communicative ability. Results suggest that a communicative approach is not strongly reflected in ELT in governmental and BRAC English classes; that Bangladeshi ELLs enrolled in governmental institutions are better communicators than those enrolled in BRAC ones; and that there appears to be a positive relationship between the presence of communicative approach indicators and students? communicative ability. This study also highlights certain instances of intersection between Western research methods and Bangladeshi (i.e., non-Western) culture-or, more appropriately, problems arising during research due to an outsider epistemic location.

AN INVESTIGATION INTO A COMMUNICATIVE APPROACH TO ENGLISH LANGUAGE TEACHING IN GOVERNMENTAL AND NONGOVERNMENTAL PRIMARY SCHOOLS IN BANGLADESH by Todd Haviland McKay A thesis submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Master of Arts Department of Linguistics The University of Utah December 2011 Copyright © Todd Haviland McKay 2011 All Rights Reserved T he Uni v e r s i t y o f Ut a h Gr a d ua t e S c h o o l STATEMENT OF THESIS APPROVAL The thesis of Todd Haviland McKay has been approved by the following supervisory committee members: Johanna Watzinger-Tharp , Chair 08/24/2011 Date Approved Rachel Hayes-Harb , Member 08/24/2011 Date Approved Aniko Csirmaz , Member 08/24/2011 Date Approved and by Edward Rubin , Chair of the Department of Linguistics and by Charles A. Wight, Dean of The Graduate School. ABSTRACT In recent years, the government of Bangladesh has worked to implement a language policy to foster a communicative approach to English language teaching (ELT) beginning at the primary level in the Bangladeshi educational system. Since such a top-down decision was made, no studies have been conducted to substantiate the presence of a communicative approach in Bangladeshi English classrooms. The aims of this study are thus (1) to probe to what extent in-class English teaching and learning practices reflect a communicative approach to ELT; more specifically, this thesis motivates "indicators" used in the documentation of "present realities" in governmental and BRAC primary school English classrooms, (2) to assess the communicative ability of Bangladeshi students enrolled in these schools, and (3) to examine the relationship between the aforementioned indicators and students‟ communicative ability. Results suggest that a communicative approach is not strongly reflected in ELT in governmental and BRAC English classes; that Bangladeshi ELLs enrolled in governmental institutions are better communicators than those enrolled in BRAC ones; and that there appears to be a positive relationship between the presence of communicative approach indicators and students‟ communicative ability. This study also highlights certain instances of intersection between Western research methods and Bangladeshi (i.e., non-Western) culture-or, more appropriately, problems arising during research due to an outsider epistemic location. To the children of Bangladesh-both enrolled and at large-my committee members for their encouragement and tutelage (and just getting me through this), my folks, and Mehran Mazinani the Leopard. TABLE OF CONTENTS ABSTRACT ............................................................................................................... iii GLOSSARY OF ABBREVIATIONS AND ACRONYMS ......................................... vi 1 INTRODUCTION ..................................................................................................... 1 1.1 Overview .................................................................................................... 1 1.2 Education in Bangladesh ............................................................................. 4 1.3 Purposes of Thesis .................................................................................... 11 2 REVIEW OF THE LITERATURE .......................................................................... 14 2.1 Overview .................................................................................................. 14 2.2 Defining Communication and the Communicative Approach .................... 15 2.3 Relevant Second Language Acquisition Research ..................................... 17 2.4 Indicators of a Communicative Approach.................................................. 27 2.5 Research Questions (RQs) ......................................................................... 29 3 METHODOLOGY .................................................................................................. 30 3.1 Overview .................................................................................................. 30 3.2 Instruments ............................................................................................... 31 3.3 Participants ............................................................................................... 42 3.4 Procedures ................................................................................................ 44 4 RESULTS AND DISCUSSION .............................................................................. 47 4.1 Overview .................................................................................................. 47 4.2 RQ 1 ......................................................................................................... 47 4.3 RQ 2 ......................................................................................................... 52 4.4 RQ 3 ......................................................................................................... 59 5 CONCLUSION ....................................................................................................... 67 APPENDIX ................................................................................................................ 72 REFERENCES........................................................................................................... 75 vi GLOSSARY OF ABBREVIATIONS AND ACRONYMS ACA ............................................................Assessment of communicative ability ALM ..................................................................................Audio-Lingual Method BANBEIS ............... Bangladesh Bureau of Educational Information and Statistics BRAC ............................................... Bangladesh Rural Advancement Committee CAI ................................................................. Communicative approach indicator CI ........................................................................................ Comprehensible input CLT .............................................................. Communicative Language Teaching CO .................................................................................... Comprehensible output ELL ................................................................................. English language learner ELT ............................................................................... English language teaching GoB ............................................................................ Government of Bangladesh L1 ................................................................................................... First language L2 ............................................................................................... Second language NCTB ................................................... National Curriculum and Textbook Board NFE ..................................................................................... Nonformal Education NGO ..................................................................... Nongovernmental organization NNS .......................................................................................... Nonnative speaker NS .................................................................................................. Native speaker OG ............................................................................................ Observation guide PEC ....................................................................... Physical environment checklist RQ ............................................................................................ Research question SLA .......................................................................... Second language acquisition SSI ............................................................................... Student-student interaction TFT ....................................................................................... Teacher-fronted time TL ................................................................................................ Target language CHAPTER 1 INTRODUCTION 1.1 Overview The goals of English language programs around the globe are numerous and varied; some seek to improve the reading and writing abilities of English language learners (ELLs) so that they might better navigate the often rough waters of higher education, while the goal of others is to provide content-area and language knowledge simultaneously (e.g., content-based instruction). Another broad goal of English language programs (not unrelated to the two just mentioned), and the one most relevant to the present thesis, is that of developing the communicative ability of ELLs; in other words, their ability to convey and interpret a message via written or spoken modalities to another person. In order to develop the communicative ability of its ELLs and strengthen human capital, the government of Bangladesh (GoB), in cooperation with the National Curriculum and Textbook Board (NCTB), implemented a language policy whose directive was to foster a communicative approach to English language teaching (ELT) beginning at the primary level in the Bangladeshi educational system (Hamid & Baldauf, 2008). At the root of this language policy, and motivating this process of building Bangladeshi human resources through a communicative approach to ELT, is, at least on 2 the part of the GoB, a desire to play a more prominent role in the international arena (NCTB, 2003). The prime minister of Bangladesh herself, Sheik Hasina, recognized that English can aid the country in gaining more solid footing on the global stage, stating that "with a view to promoting employment abroad and encouraging transfer of technology, emphasis will be laid on teaching English along with the mother tongue" (The Daily Observer, 2002). This new stance on ELT marked a significant change from the way English had previously been taught in Bangladesh. The implementation of a communicative approach to ELT is evidenced in the bhumika (or „preface‟) of English for Today, the English language textbook used in primary schools across the country, which clearly states that Bangladeshi ELLs will learn English in the framework of a communicative approach through "interaction […] individual work, pair work, and group work" (English for Today, 2009). While the goal of developing the communicative ability of Bangladeshi ELLs is a worthy pursuit, the difference between adopting a specific approach to ELT and taking steps to implement that approach in the classroom should be observed. Long (1990) noted in his study on group work and interaction that "classroom observations consistently show little continuity in what teachers actually do and what they think they are doing" (p. 32). A look at the research pertaining to the current English language proficiency of Bangladeshi ELLs helps to elucidate the disparity in Bangladesh ELT between theory and in-class practice; in other words, despite government insistence on use of a communicative approach that could bolster Bangladeshi ELLs‟ communicative ability, research suggests that this insistence has borne few results. 3 The gap between theory and in-class practice, as it pertains to the communicative approach, has been observed at primary, high school, and university levels in Bangladesh. Choudhury (2006), working with English language students at a private university in Bangladesh, noted that even after 12 years of English education, students cannot communicative effectively and "are failing to develop an acceptable level of English proficiency" (p. 85). Others, in their comparison of English test grades and Secondary School Certificate English grades at the high school level, remarked upon "an alarming rate of failure of rural students in English" (Hamid & Baldauf, 2008, p. 21). Lastly, Rahman (1999) stated that "the situation is particularly serious in nongovernmental rural schools, where over 95% of students at class 6 and 8 are failing to reach the expected standards of proficiency" (p. 16). The studies mentioned are representative of the widely found incongruity between students‟ expected and actual level of English proficiency. These studies suggest that, while a rank-and-file decision sought to implement the communicative approach in English language classrooms at different levels in the Bangladeshi educational system, such a decision did not spur in-class changes that ultimately bore a positive change on ELLs‟ communicative ability. Research suggests that this divergence could be ascribed to (1) older, more traditionally teacher-centered pedagogical practices employed in the classroom (Choudhury, 2006; Hamid & Baldauf, 2008; Prather, 1993), (2) a mix of English- and Bangla-medium students in the same English class (Choudhury, 2006), (3) the employing of paraprofessional English teachers and literature pundits instead of English language teachers (Choudhury, 2006; Prather, 1993), (4) student reservations to 4 speaking in class (Choudhury, 2006), and (5) a misapplication of precious resources towards English language education (Hamid & Baldauf, 2008) among others. So far, this thesis has touched on the GoB‟s push for more communicative Bangladeshi ELLs and highlighted studies that substantiate a gap between ELLs‟ expected and actual level of English communicative ability. While these studies have identified factors that could be inhibiting the implementation of a communicative approach in the English language classroom, more studies are needed that "acknowledge and document present realities through classroom-based research" (Nunan, 1986, p. 142) in Bangladesh. 1.2 Education in Bangladesh Language policies that aim to develop the communicative ability of Bangladeshi ELLs are a subset of larger educational goals being actively pursued by the GoB. The GoB, following the Convention on the Rights of the Child (1990), affirmed education as a basic human right, and has since fought to increase the basic literacy and numeracy of the Bangladeshi population. Such an effort, however, has proven difficult as children and adolescents often forgo education as a result of widespread poverty in the country to contribute to their family‟s modest monthly income. A 2011 World Bank estimate put more than half of the 135 million Bangladeshi population below the poverty line (an estimate measured by income, ability to meet basic human needs, and consumption); recent statistics from the 2009 Human Development Report indicated that the human poverty index of Bangladesh (which incorporates living a long and healthy life, access to education, and standard of 5 living) is 36.1%, earning it a rank of 112th among the 135 countries for which the human poverty index was calculated. And, making the connection between poverty and education more explicit, Ibrahim (2002) conjectured that abject poverty is one of the leading causes that explain drop-out rates among adolescents. Due to this fact, enrollment and attendance rates are dismal at best and drop-out rates are high. United Nations Development Goals Indicators 2004 (updated June, 2006) showed that the primary net enrollment rate was 94%-although only 76% completed class 5, and the drop-out rate was roughly 23.6%. Education Watch (2008) put the attendance rate at the primary level at 67.7%. Despite enrollment and attendance rates that have improved over the years in response to efforts channeled towards improving the educational situation in Bangladesh, these statistics are nevertheless saddled alongside those indicative of poor attendance and low survival rates. 1.2.1 The Structure of the Educational System in Bangladesh Prior to discussing just how the GoB is working to offer more extensive educational coverage in its plan for nationwide literacy and numeracy, now would be a prudent time to provide the reader with a snapshot of Bangladesh‟s educational structure; that is, what are the different grade levels, and what types of educational institutions are available in Bangladesh? Even though there are different types of institutions in Bangladesh, including madrasah, governmental, nongovernmental, and international or private institutions, the two relevant for this thesis are governmental and nongovernmental (respectively referred to as formal and nonformal in Bangladesh)1. 1 I make the distinction here between formal and governmental, and nonformal and nongovernmental so as to distance any preconceived notions of "formal" and "nonformal" (i.e., formal and informal) from these 6 Furthermore, for reasons that will follow, this snapshot limits itself to the primary-secondary sequence of the overall educational structure. According to the Bangladesh Bureau of Educational Information and Statistics (BANBEIS), primary education in Bangladesh covers grades 1-5. Students typically enroll in a primary school at the age of 6 and graduate at approximately 10 years old. Once students graduate from primary school, their next step varies from one student to the next; many students continue on in their studies and enroll in governmental educational institutions. Others, influenced largely by poverty, do not continue their studies; these students may choose to look for work in the hopes of providing monetary support to their families (Ibrahim, 2002). 1.2.2 The Role of the NGO in Bangladesh The enrollment, drop-out, and attendance rates, mentioned above, point to yet another incongruity, one between the goals of the GoB and what is actually happening in primary institutions around the country. As the GoB pursues a broader application of primary education nationwide, many students skip classes and drop out before completing the primary-school cycle. Many of those dropping out of school are doing so to obtain a job to help their families foot the bills for daily expenses (Ibrahim, 2002). This conflict between the GoB‟s broader educational goal and the reality of the educational situation is further complicated by the inability of the GoB to reach many rural communities in the hopes of providing education. Bridging the gap between the two types of institution. In other words, nongovernmental schools are simply called "nonformal" in an attempt to distinguish them from governmental schools, but do not forcibly imply "informal" teaching and learning practices, nor do governmental schools adhere to "formal" ones. 7 goal of the GoB to extend education to all on the one hand, and the inaccessible rural populations of the country on the other, requires the support of nongovernmental organizations, or NGOs. Holloway (1998) and others (e.g., Ibrahim, 2002; Prather, 1993) portrayed NGOs as government helpers, recognizing that they complement GoB efforts to ensure a minimum level of education among the Bangladeshi population. NGOs typically have financial, administrative, and personnel and human resources at their disposal (that the GoB does not) that allow them to access the hard-to-reach areas of Bangladesh (Ibrahim, 2002; Prather, 1993). In gaining access to the more rural areas of the country, NGOs are thus able to reach communities where schools are unavailable to primary school-aged children. Furthermore, if the GoB is committed to providing a minimum level of education across the country, in view of their resource handicaps they must rely to some extent on the abilities of NGOs to carry out educational operations in remote or rural areas of the country where the Bangladeshi school-aged population is not served, or poorly served, by the governmental educational system (BRAC Education Programme Midterm Review, 2004-2009). With 12-15% of external assistance to Bangladesh flowing through NGOs, their support is crucial for promoting basic literacy and numeracy skills in all corners of the country (Prather, 1993). 1.2.3 NGO and Government Overlap NGO support is a verifiable bulwark should the GoB truly wish to raise the educational bar in Bangladesh; yet, as Ibrahim (2002) noted, "while there are many NGOs involved in the delivery of NFE [nonformal education] programmes, only a few 8 are important players, as they advocate new models, provide curriculum materials and have established networks under partnerships" (p. 4). Of the handful of key players with their fingers in nongovernmental education that Ibrahim lists, one is the Bangladesh Rural Advancement Committee, or BRAC. BRAC got its start in education in the 1970s in response to the concerns of mothers who were taking adult literacy courses that their daughters would miss out on education (BRAC Annual Report, 2009). In order to educate the daughters of these women and other youth in rural communities, BRAC created its nonformal system (i.e., nongovernmental system). Nath (2002), a senior research statistician and coordinator for BRAC Research and Evaluation Division, provided a brief overview of important features in BRAC‟s nongovernmental system that differ from most governmental ones. Table 1 captures the similarities and differences between governmental and nongovernmental systems. The resources at the hands of NGOs, and the stretched resources of the GoB, stipulate to a certain degree differences in administration, teacher training, the availability of in-class manipulatives, and curriculum. NGOs, by definition altruistic in the sense that they are driven by a "vision," stress education alongside empowerment inputs and employ a larger number of female teachers than do governmental primary schools. Differences certainly exist between governmental and nongovernmental primary schools, and these should by no means be brushed aside; however, certain aspects of nongovernmental schools were aligned with those of governmental schools in order that nongovernmental, primary-school graduates more easily transition to governmental institutions at the secondary level. 9 9 Table 1: Capturing School Differences Governmental Primary Schools BRAC Nongovernmental Primary Schools Class Size About 60 students per class. About 30 students per class (Nath, 2002). Male: Female More even ratio between male and female students. About two-thirds of students are female (Nath, 2002). Teachers Both male and female teachers with varying schooling and certification. Mostly female teachers with about 10 years of schooling who are drawn from the local community (Lovell & Fatema, 1989; Nath, 2002). Primary School Length Primary schooling from grades one through five (i.e., five years of primary schooling). Primary schooling from grades one through five; however, this is accomplished in four years by shortening vacation time (Nath, 2002). Location Primary schools in both urban and rural areas. Primary schools in predominantly hard-to-reach rural areas (Holloway, 1998; Ibrahim, 2002; Prather, 1993). Enrollment 89.64% in government primary schools (Holloway, 1998). 92% in NGO nongovernmental schools (Holloway, 1998). Survival Rate About 54% of students drop out before completing primary school cycle (BRAC Education Programme Midterm Review, 2004-2009). About 94% of students in BRAC primary schools remain enrolled throughout primary school cycle (BRAC Education Programme Midterm Review, 2004-2009). Resources Lack resources necessary for providing primary education to remote populations (Holloway, 1998; BRAC Education Programme Midterm Review, 2004- 2009). Have more resources, including age, space, time, pedagogical, and curricular resources (Ibrahim, 2002; Prather, 1993). 10 One of the overlapping aspects of governmental and nongovernmental primary schools is the use of NCTB-based English language textbooks and the same English for Today textbook, which again frames ELT in a communicative approach and pushes for interaction, individual work, pair work, and group work (English for Today, 2009). According to Nath (2002), BRAC uses its own textbooks, based on the NCTB curriculum for grades 1-3, and uses English for Today for grades 4 and 5. In sum, the GoB‟s goal of extending primary education coverage is shared by NGOs. Despite the existence of other primary institutions in Bangladesh (e.g., madrasahs and private institutions), private institutions entail fees and other expenses that well exceed the means of many Bangladeshi families. Furthermore, with such a large percentage of the Bangladeshi population residing in rural areas of the country- which, recall, lay outside the GoB‟s reach-and NGOs working to extend primary education in rural areas in which the population is underserved by mainstream education, this thesis focuses on the two "leading" primary institutions in Bangladesh. What makes an investigation into the English teaching and learning practices more appropriate across governmental and nongovernmental primary institutions in Bangladesh are, apart from GoB-NGO cooperation and English curricula based on a communicative approach, (1) a dramatic increase in the number of NGO schools in Bangladesh since the 1990s (Ardt, Hastings, Hopkins, Knebel, Loh, & Woods, 2005 report an increase of four times), and (2) differences between governmental and nongovernmental institutions. As the number of NGO-run primary schools continues to augment in rural parts of the country, disparities in the quality of education provided by these schools-and the reason for these disparities-begin to figure more prominently 11 into the concerns of the GoB, NGO and other educational researchers, and parents (Ardt et al., 2005). BRAC, as was previously mentioned, is the lead provider of nongovernmental primary education among NGOs, with approximately 76% of all NGO schools being BRAC schools (Kabeer, Nambissan, & Subrahmanian, 2003). 1.3 Purposes of Thesis One of the incongruities, highlighted among others by Choudhury (2006) and Hamid and Baldauf (2008), is the distance between Bangladeshi ELLs‟ expected level and actual level of English communicative ability, in spite of official insistence on use of a communicative approach in the English language classroom. Recall that the GoB made its top-down decision to implement a language policy foregrounding a communicative approach to ELT in the hopes of bolstering the communicative ability of Bangladeshi ELLs. Of concern then for the GoB is why ELLs‟ level of English proficiency is failing to meet expectations. As was previously noted, research conducted so far suggests that this failure could be ascribed to (1) older, more traditionally teacher-centered pedagogical practices employed in the classroom (Choudhury, 2006; Hamid & Baldauf, 2008; Prather, 1993), (2) a mix of English- and Bangla-medium students in the same English class (Choudhury, 2006), (3) the employing of paraprofessional English teachers and literature pundits (Choudhury, 2006; Prather, 1993), (4) student reservations to speaking in class (Choudhury, 2006), and (5) a gross misapplication of precious resources towards English language education (Hamid & Baldauf, 2008) among others. However, more classroom- 12 based studies are needed to help isolate in-class practices that are hindering students‟ communicative development. In this light, and in consideration of the assisting role of the NGO in administering primary education (viz., BRAC) to areas out-of-reach of the GoB, this thesis has three purposes: (1) to investigate to what extent in-class English teaching and learning practices reflect a communicative approach to ELT; more specifically, this thesis will motivate a set of "indicators" that can be used to assist in the documentation of the "present realities" (Nunan, 1986, p. 142) in governmental and nongovernmental primary school English language classrooms; (2) to assess the communicative ability of Bangladeshi primary school students enrolled in governmental and nongovernmental institutions; and (3) to examine the relationship between the aforementioned indicators and the communicative ability of Bangladeshi ELLs. This section of my thesis has helped to clarify that the educational situation in Bangladesh is a dire one. The GoB is working to improve the literacy and numeracy of its population, and this work has been encumbered by poverty and the inability to access many rural parts of the country without enlisting the help of NGOs. Galloping alongside this work is the GoB‟s desire to boost the communicative ability of its ELL population that Bangladesh might figure more prominently on the international stage. While achieving national literacy and numeracy, and ELLs with more formidable communicative ability, are worthy endeavors, the GoB-and the NGO, by extension- need to be wary of not only the quantity of education, but the quality of education being provided. In other words, it would be unreasonable to assume that a Bangladeshi ELL can develop communicative abilities without first being exposed to or experiencing in the 13 classroom the means with which to attain them. It is my sincere desire, therefore, that the results of this thesis will be used to inform English instructional practices, administrative decisions, and teacher training in governmental and nongovernmental primary institutions in Bangladesh. It is also my hope that this thesis will provide further insight into the in-class implementation of a communicative approach to ELT and the level of communicative ability of both governmental and BRAC primary populations for another reason-that, as our understanding of various factors at play in Bangladeshi ELT converges, the beginnings of a model for implementing a communicative approach to ELT in Bangladeshi primary classrooms will take shape. The scope of this thesis is limited, and does not dismiss the importance of determinants external to the physical primary-school classroom in Bangladesh that result in the realization of English language classes (e.g., administrative, personnel, and financial resources). Nonetheless, this thesis, in investigating the in-class implementation of a communicative approach to ELT (and students‟ level of communicative ability), assumes an inside-out avenue to understanding the implementation of this particular approach. From this understanding, specifically that of the communicative approach English teaching and learning practice in governmental and BRAC primary schools, future researchers, administrators, and English language teachers will be equipped with the wherewithal to more effectively-and perhaps efficiently-implement a communicative approach to ELT. CHAPTER 2 REVIEW OF THE LITERATURE 2.1 Overview It is not enough to state that a communicative approach to ELT is being used in the classroom. A communicative approach is supposedly at work in English language classrooms in Bangladesh, but research suggests that this is not the case. Inherent in the term "communicative approach" is the word "communication," which in and of itself is understandable to many English native speakers (NS) and nonnative speakers (NNS); yet, when "communication" is viewed through the lens of ELT and second language acquisition (SLA) perspectives, it must be carefully defined, for it is from an understanding of what the goal of communication is-and what communication entails- that a "communicative approach" is fashioned. Upon arriving at a definition of "communication" and a better understanding of what constitutes a "communicative approach," this section of my thesis will then turn to SLA research to motivate use of a communicative approach in Bangladeshi governmental and nongovernmental primary schools. I will draw on this same research for the purpose of isolating a set of indicators that can be used to examine the in-class implementation of a communicative approach. 15 2.2 Defining Communication, Communicative Approach, and Communicative Language Teaching The term "communication" is given a number of definitions in SLA research; however, essential in all of them is the need for more than one person to engage in social interaction wherein meaning is created, conveyed, and interpreted. Breen and Candlin (1980) stated that "communication and learning to communicate involve participants in the sharing and negotiation of meanings" and that a participant needs "to be able to interpret the meanings of others and express his own meanings" (p. 92). Similarly, Littlewood (1981) drew a comparison between communication and a social "transaction" in which "two parties are involved" (p. 66). Based on these definitions, and the utmost necessity for more than one person to be involved in a communicative exchange, it becomes clear that what distinguishes communication from simple language production can be winnowed down to direction and access. In order for communication to take place, there must be a two-way exchange between at least two interlocutors. Also, what an individual produces must then be accessed by another participant in the communicative exchange. Richards and Rodgers (1987) took a firm stance on the purpose of language, stating that language falls service to "interaction and communication" (p. 71). It was precisely this take on the role of language that led to the development of a communicative approach to ELT. Previous ELT methods, such as the Audio-Lingual Method (ALM), also involved language production, but this particular method had close ties to behaviorist thinking in that its primary concern was with the formation of habits through drills and repetition. What distinguishes a communicative approach from the 16 ALM-like the distinction between language production and communication above-is the construction and interpretation of meaning in veritable, real-time social interaction. To achieve this end, language must involve "unpredictability and creativity […] and be judged as successful or not on the basis of behavioral outcomes" (Canale & Swain, 1980, p. 29). ALM sought to equip language learners with a bandoleer of pre-packaged utterances instead of skills for navigating dynamic social interactions. How then does one appropriately distinguish between a communicative approach, communicative competence (Canale & Swain, 1980; Hymes, 1972; Savignon, 2001), communicative performance (Canale & Swain, 1980), and communicative language teaching (CLT)? Richards and Rodgers (1987) claimed that a communicative approach is the same as communicative language teaching (p. 65), but I would argue that CLT refers more explicitly to pedagogy and design (the how of language teaching); a communicative approach, on the other hand, seeks to identify a set of assumptions that flow from a theoretical understanding of the essential social function of language (the why behind a specific flavor of language teaching). The notion of communicative competence, coined by Hymes (1972), stems from his perceived inadequacy of Chomsky‟s (1965) distinction between linguistic competence and performance. Hymes argued that it was not enough to view competence and performance with respect to a language speaker‟s knowledge of grammatical rules; rather, a speaker must be able to manipulate grammatical knowledge within the appropriacy dictated by the context in which the communicative exchange occurs2. In 2 It is important to note that Hymes‟ model of communicative competence addresses the native speaker in its conception-both a practical and theoretical problem in the field of second language acquisition, as it does not incorporate an understanding of interlanguage variety amongst even native speakers (Blyth, 2003). 17 short, communicative competence, according to Canale and Swain (1980), is the "relationship between linguistic competence (rules and grammar) and socio-linguistics competence (rules for use)" (p. 6). Canale and Swain, building on this definition of communicative competence, recognize that competence as knowledge is itself not readily observable; a language speaker‟s competence, linguistic or communicative, can be realized only through social interaction "in the actual production and comprehension of utterances" (p. 6), or communicative performance. 2.3 Relevant Second Language Acquisition Research 2.3.1 Communication Inside and Outside of the Classroom For a communicative approach to exist in an English language classroom, it is absolutely crucial that classroom activities mirror, to the extent possible, the type of communication that one is likely to encounter in social interaction outside of the classroom (Canale & Swain, 1980; Nunan, 1986). This being said, a communicative English language classroom should embody those aspects of social interaction that would occur outside of the classroom: the chance to construct relationships (Littlewood, 1981), to practice producing with the language in genuine communication (Canale & Swain, 1980; Littlewood, 1981; Nunan, 1986; Swain 1993) and in a nonthreatening environment conducive to language learning (Long, 1990; Long & Porter, 1985), and to negotiate meaning (Ellis, 2000; Gass & Varonis, 1985; Long, 1990; Long & Porter, 1985; Pica, 2002; Pica & Doughty, 1985; Varonis & Gass, 1985). This differentiation from one native speaker to the next then makes the goal of realizing "nativelike" competency unrealistic let alone definable (Auger & Valdman, 1999). 18 Some might argue that it is unrealistic to try to approximate classroom-based learning to real-world social interactions. However, those aspects of real-world communication that are said to constitute a minimum in social interaction are also present in classroom learning environments that push meaningful communication to the front lines. These aspects are primarily the presence of the teacher and other learners-hence, other interlocutors-and the production of utterances that must contain meaning, meaning which is then accessed, interpreted, and responded to by another interlocutor. Breen and Candlin (1980) stressed that the "communicative classroom can serve as a forum" (p. 95) wherein students interact both with each other and their teacher, and Littlewood (1981) similarly credited the classroom as a "real social context in its own right, where learners and teacher enter into equally real social relationships with each other" (p. 44). 2.3.2 Pair and Group Work As communicative exchanges through language involve a minimum of two people (Cheng, 1980; Thompson, 1996), it follows that any English language classroom wherein the focus is on communicative activities should emphasize pair work and group work (types of collaborative learning). Ciotti (1969), in her study of different small group formations, defined the small group as being "the most fundamental social unit for it is through this medium that the individual accomplishes tasks and satisfies socio-emotional needs" (pp. 78-79). This emphasis on pair work and group work is reflected by the overwhelming prevalence of certain toolbox activities in communicative classrooms, namely simulations, role plays, and discussions (Qinghong, 2009; Thompson, 1996). 19 There is a large body of research that documents the benefits of group work. One of the benefits of pair and group work is that it substantially increases the amount of "talk time" for students to produce with language in the classroom (Allright, 1984; Cheng, 1980; Long, 1990; Long & Porter, 1985; Pica, 2002; Swain, 1993). 2.3.2.1 Production and acquisition. Language production through spoken and written modalities has been shown to be a critical variable in the acquisition process. Swain (1974, 1985, 1993) noted in her research with students in French immersion programs in Canada that, while these students had receptive language abilities equivalent to their first-language (L1) counterparts, they performed less ably on measures of productive language skills (i.e., speaking and writing). Based on these findings, she posited her comprehensible output hypothesis that states that second-language (L2) learners, in order to develop speaking and writing skills, need opportunities to practice producing with the language, to negotiate input by interacting with both texts and other learners, to receive explicit instruction on form, and to test and receive feedback on hypotheses. 2.3.2.2 Changing roles. The use of pair work and group work in the communicative classroom has a dual purpose; it serves to increase the amount of talk time for students-thereby improving their chances to acquire language-and it shifts the focus in the classroom from the teacher to the student. Long (1990) noted that the ALM, Audio-Visual Method, and the Grammar-Translation Method were predominantly "teacher-centered," provided "minimal input" for student learning, and devoted large chunks of in-class time to "pseudo-communication" (pp. 31-32). In order for a communicative approach to take hold in the English language classroom, students must 20 be afforded opportunities to practice producing with the language-accomplished through the use of pair and group work-but also teachers must learn to "wean students off of teacher dependency" (Littlewood, 1981, p. 18) and learn to put themselves on the level of the students as a co-communicator rather than a "controller-in-chief" (Cheng, 1980, p. 61). Lee and VanPatten (2003), citing Finkel and Monk (1983), described the physical setup of transmission-oriented classrooms and the "Atlas Complex," wherein teachers bear the weight of student learning on their shoulders and view students as vessels for receiving knowledge. Classrooms that advocate a communicative approach require that teachers shed themselves of this lofty character and put themselves on an even keel with their students. This requisite for a communicative classroom is evidenced in the literature when researchers refer to teachers in a communicative setting as "facilitator" (Larsen-Freeman, 1986), "co-communicator" (Littlewood, 1981), "guide" (Cheng, 1980), "counselor" (Richards & Rodgers, 1987), and "negotiator" (Breen & Candlin, 1980). Proponent teachers of this approach must be able to guide students as they complete activities, share learning with their students, and, once an activity is underway and students know what it is they need to do, take a step back and be as "unobtrusive as possible" (Cheng, 1980, p.60). The teacher should step in only when it is absolutely necessary to assist in the completion of an activity and suppress the "inclination to supply gaps in lexis, grammar, [and] strategy" (Richards & Rodgers, 1987). Furthermore, returning to the notion of "talk time," teachers who are used to more traditional lockstep modes of instruction are depriving their students of valuable practice with the target language by doing the majority of the talking. A study by Flanders (1970) 21 showed that teachers in lockstep instructional formats sometimes talk for more than half of the allotted class time. As teachers, it can be difficult to turn the reins over to students; yet, teachers should be aware that while teacher utterances tend to be more grammatical than student production (though this is not always the case), Swain (1993) noted that teacher-led discussion generates "simple, syntactically short utterances; students must be allowed to engage in extended discourse that will push their linguistic competence to the limit" (p. 162). This is not to say, however, that the teacher cannot be a source of meaningful communication either with the students as a group or with students individually. The simple fact of the matter is that if a teacher does most of the in-class talking for whatever reason, students are allocated limited language production time and are being exposed to only a single variety of English, which, according to Long and Porter (1985) is "highly conventionalized" (p. 209). 2.3.2.3 Potential pitfalls in pair and group work. By working in groups, students can bring their own individual opinions, feelings, and other differences to the language-learning table. It is particularly this diversity of individual learners that Breen and Candlin (1980) argued "authenticates communication in the classroom" (p. 97). Lockstep formats wherein the teacher spearheads conversation and interaction gloss over individual differences that provide authentic communicative situations (Long & Porter, 1985). Criticism on behalf of teachers has been that conducting group work in the classroom places demands on time that would be better spent transmitting knowledge and leads to confusion with "only a handful of students doing the work" (Paulston & Britanik, 1995, p. 80). This belief in the futility or pains of conducting group work is not altogether unfounded, and teachers in a communicative classroom must be careful that 22 group work is conducted carefully, in a way that does not steamroll individual differences that provide the necessary diversity for authentic communication and ensures individual group member accountability. Indeed, Long (1990) warns teachers that "badly organized group work is not better than badly organized lockstep work" (p. 37). Blumenfeld, Marx, Soloway, and Krajcik (1996) supported Long in this claim, and I am confident that all teachers who have conducted group work in their own classrooms have had difficulties on occasion getting students to work together and to share the workload of a particular project or activity. By asking students to work in groups, teachers are not immediately guaranteed successful cooperation among group members, nor can they rest assured that language acquisition will occur. Blumenfeld et al. offered a short list of factors that teachers should heed when conducting group work in their classroom, including group norms, tasks, and the composition of the group. Problems that can arise in conducting group work include failure to contribute (accountability), overly assertive students usurping production time available in the process of completing group activities, and students ostracized from the group (p. 38). Another concern surrounding the use of group work is that learners, being at variable stages in the acquisition process, will develop inaccurate or incomplete representations of grammatical knowledge as they communicate with one another. As learners work to boost their communicative ability in the target language (TL), many teachers fear that grammatical misconceptions may go uncorrected, thus leading to the stabilization or fossilization of a nontarget variety of the TL. While teacher-fronted discussion or interaction is (typically) more grammatical overall (Pica, 2002), this is not to say that learner-learner interaction in groups is forcibly detrimental to the 23 representation of grammatical knowledge in learner interlanguage (Selinker, 1972) systems. There are a number of studies germane to the relationship between group work and the maturation of grammatical skills. Swain (1993) described four ways in which language production can assist in language acquisition, one of which is that language production can force language learners to move from processing message meaning in communication to syntactic information as they "recognize gaps in their language" (p. 161). Pica (2002; cf., Long & Robinson, 1998) made a similar observation, describing how communication breakdowns in the communicative process can lead learners to "focus on form" (p. 5) to figure out what prompted misunderstanding between learners. However, learners are not always drawn to syntactic information in the communication and interaction process. When students have difficulty noticing grammatical forms, at this point "there is a tendency for them to develop incomplete or incorrect representations in their interlanguage development" (Pica, 2002, p. 4). When grammatical errors go uncorrected, learners might construct inaccurate representations in their interlanguage systems, but this process does not necessarily earmark a "decline in grammatical accuracy" (Long, 1990). Cheng (1980) stated in her work with high school students in Hong Kong that "once the small-group relationship is established, students will help, prompt, encourage, and correct one another" (p. 62). A study by Donato (1994) showed that learners, incapable of producing a particular grammatical structure individually, were able to produce it with the help of the group. Ellis (2000) described this process as "scaffolding" (p. 209) wherein learners assist one another to achieve tasks and complete functions in the TL. After scaffolding in a group, 24 according to Ellis, learners are more likely to internalize the scaffolded information and be able to perform future tasks and functions "unassisted" (p. 209). Lastly, Long and Porter (cf., Porter, 1983) demonstrated that student miscorrection is not a serious threat in the classroom, as only .3% of errors produced by fellow students went miscorrected. Conducting group work in the communicative classroom allows learners to express their individual differences and opinions with fellow learners. Working in groups can substantially increase the amount of "talk time" or practice that students have with the TL that can facilitate the language acquisition process. The classroom can serve as a forum for language learning provided that teachers, having initiated group work- and done so in a manner that anticipates potential problems that can arise in group work (e.g., those discussed by Blumenfeld et al., 1996)-learn to turn the floor over to student learning. Group work, in that it is composed of a minimum of two learners, "authenticates" (Breen & Candlin, 1980, p. 97) communication in the classroom. 2.3.2.4 Negotiation of meaning and input. While practice producing with the target language is one aspect that can spur the acquisition process, another essential component of language acquisition that group work provides is the "negotiation of meaning," which Pica (2002) defined as "when one interlocutor‟s message appears to another interlocutor to be unclear, incomprehensible, or incomplete in its meaning" (p. 4). Succinctly put, negotiation of meaning refers to the process of overcoming communication breakdowns, or "non-understanding routines" (Varonis & Gass, 1985), that crop up during communication. In the process of negotiating meaning, learners use confirmation checks, clarification requests, and comprehension checks to overcome nonunderstanding routines and make input more comprehensible (Long, 1980); that is, 25 learners use a set of strategies when a misunderstanding in communication arises to tailor the input they receive from another learner or co-communicator. Along with comprehensible output (Swain, 1993), comprehensible input (Krashen, 1977) has been identified as a variable critical in the language acquisition process. In fact, Long (1983, 1996) proposed his interaction hypothesis in response to the way in which negotiation of meaning, which occurs in group-work interaction and communication, facilitates language acquisition. The input hypothesis states that "language acquisition is facilitated when learners obtain comprehensible input as a result of negotiation." The interactional features used to overcome communication breakdowns during the course of communication occur both in group work among learners and in teacher-learner dyads. Just as meaningful communication can occur in teacher-student interaction, these types of interaction detract from the total amount of "talk time" available in class-thus the amount of time students have to practice with the TL-and result in fewer communication breakdowns that subsequently prompt both teachers and learners alike to engage in the negotiation work necessary to generate comprehensible input and, by extension, language acquisition (Pica & Doughty, 1985). While NS-NNS dyads (including teacher-student) dyads, this formation is not "resource-practical one" in ELT (Long & Porter, 1985). This type of formation could only be made in the English language classroom provided that the teacher work individually with students, or that NSs are brought in from outside the classroom. Varonis and Gass (1983), however, have shown that the amount of negotiation work that occurs in NNS-NNS interactions is more frequent than in lockstep formats or NS-NNS formations, and Gass and Varonis (1985) stated that it is "precisely NNS-NNS pairs that 26 offer NNSs the greatest opportunity to receive CI [comprehensible input] and produce CO [comprehensible output] through negotiation" (p. 161). Further support for NNS-NNS formations come from Porter (1983) and Long, Adams, McLean, and Castanos (1976). 2.3.2.5 Pair and group work and the affective filter. One of the final benefits of group work discussed in this thesis refers to the affective filter hypothesis (Krashen, 1977), which states that negative emotional (affective) responses to environmental factors bear a negative impact on the language acquisition process. When the affective filter is up, it can block processes that could otherwise facilitate language acquisition. Long and Porter (1985) accurately depicted the rising of the affective filter in a language-learning setting when they state that "students stress skyrockets when called to answer a question in the public arena with speed and accuracy" (p. 211). Group work can help mitigate the negative affective responses that learners might experience in the classroom by affording a milieu in which they are allowed to work at their own pace (as opposed to that set by the instructor)(Cheng, 1980), are freed from the pressure of having to respond with a high level of accuracy (Long & Porter, 1985), and are given a "non-threatening forum" within which to practice producing with the language (Varonis & Gass, 1985, p. 71). In groups, learners are also less likely to respond negatively-and, therefore, with a raised affective filter-to feedback that they receive from other learners in their group because all group members are working to improve their communicative ability and practice producing with the language. Varonis and Gass referred to this bond between language learners as a sort of "shared incompetence" (p. 71). 27 2.3.3 Classroom Setup The physical setup of the classroom can also be helpful for breaking down teacher-student and student-student tensions, or raised affect, in the classroom. Single-row setups in the classroom can sometimes reinforce teacher-centeredness in more traditional classrooms where the teacher stands or is seated at the head of the class while students are seated in rows and columns reminiscent of military formations. This classroom setup supports information transmission wherein the separation between the teacher and the group of learners lends itself to unidirectional teaching. Littlewood (1981) recommended more informal seating arrangements in the classroom to abate the dangers of "excessive teacher domination" (p. 94) and reinforce the idea in communicative classrooms that teachers and students are on equal footing as co-communicators. If the communicative classroom is to be compared to a forum, whenever possible, then, the layout of the classroom should uphold a social environment that facilitates teacher-student and student-student interaction. 2.4 Indicators of a Communicative Approach The purposes of this thesis are once again (1) to investigate to what extent in-class English teaching and learning practices reflect a communicative approach to ELT in governmental and nongovernmental primary school English language classrooms (as determined by a set of three indicators), (2) to assess the communicative ability of Bangladeshi primary school students enrolled in governmental and nongovernmental institutions, and (3) to examine the relationship between the aforementioned indicators and the communicative ability of Bangladeshi ELLs. 28 The GoB, while perhaps in early stages of the development of in-country ELT, is faced with the dual challenge of building the communicative ability of Bangladeshi ELLs while, at the same time, stretching what resources it has to provide a minimum education to students in both accessible and hard-to-reach areas of the country. Along with the assistance of NGOs, among which BRAC is a key player, the GoB is successfully chipping away at nationwide illiteracy and innumeracy; however, research has shown that students at different levels across the spectrum of the educational system are failing to meet expected standards of English proficiency. As the GoB, and the NCTB-who is responsible for the development of the English language curriculum used by both governmental and nongovernmental primary institutions-advocate use of the communicative approach in the English language classroom, this thesis will investigate three factors that surface frequently in the literature as being fundamental in a communicative approach to ELT: (1) teacher-student interaction (Breen & Candlin, 1980; Canale & Swain, 1980; Cheng, 1980; Long 1990; Long & Porter, 1988; among others) (2) student-student interaction (Allwright, 1984; Blumenfeld et al., 1996; Cheng, 1980; Ciotti, 1969; Ellis, 2000; Gass & Varonis, 1985; Horne, 1970; among others) , and (3) the physical setup of the classroom (Littlewood, 1981). Through the investigation of these three factors in governmental and nongovernmental primary schools in Bangladesh, this thesis will aim to answer the following research questions: 29 2.5 Research Questions (RQs) RQ 1- Has a communicative approach to ELT taken hold in English language classrooms in governmental and BRAC primary schools, as evidenced by the following three indicators: (1) teacher-fronted time, (2) student-student interaction, and (3) the physical setup of the classroom? RQ 2- Is there a difference in the communicative ability of students enrolled in governmental versus BRAC primary schools as measured by Assessment of Communicative Ability (ACA)? RQ 3- What is the correlation between students‟ scores on the ACA and the three indicators (above) identified as essential to ELT within the framework of a communicative approach? CHAPTER 3 METHODOLOGY 3.1 Overview Through qualitative and quantitative means, this study makes use of (1) two classroom observations-an observation guide (OG) and physical environment checklist (PEC)-whose data is used to address the presence or absence of the three communicative approach indicators (CAIs) and (2) the ACA, which is administered in a one-on-one interview format and whose data is used to gauge the communicative English ability of Bangladeshi ELLs in both BRAC and governmental primary institutions. This section contains a more detailed description of (1) these instruments, as well as a brusque rationale supporting their usage and underlying their development, (2) the various participant populations (i.e., the classes, including teachers and students present during classroom observations, and participants who assisted individually with this research by assenting to take the ACA), and (3) the procedures adhered to during the implementation of all the aforementioned data-collection engines. 31 3.2 Instruments 3.2.1 Classroom Observations 3.2.1.1 OG. While the term "classroom observation" might easily be confused with a single data-collection tool, the term "classroom observation" in this study encompasses both the OG3 and the PEC. The OG, loosely put, provides the researcher (and anyone else who should wish to view it) with a sketch of in-class English teaching and learning practices; in other words, it provides information pertaining to in-class proceedings, such as who is talking when and what are the teacher and students doing, that lends itself to a visualization of the class observed. The OG, in addition to the date, class time, school type (i.e., BRAC or governmental), grade, overall class duration, and sex of the teacher, allows the researcher to document at 2-minute intervals (1) who is talking when and with whom, and (2) what the teacher and students are doing. To the extent that the documentation of in-class proceedings at two-minute intervals provides a comprehensive view of the class observed, the information collected with the OG speaks to two of the three CAIs (a point to which I shall return momentarily). The OG includes "T," "S," and "O" tokens (cf., Table 2), which can be circled to designate the classroom situation at a given 2-minute interval. The "T" token is shorthand for "teacher-fronted time (TFT)", which is defined as those instances of classroom time in which students are not working in pairs and small groups; that is, time that is less optimal for language learning according to communicative approaches to ELT. The "S" token is shorthand for "student-student interaction (SSI)", strictly defined 3 The OG used in this study is adapted from Estacion, McMahon, and Quint‟s (2004) Physical Environment Form (PEF). 32 Table 2: A Section of the OG Min. TFT SSI Other Notes 2 T S O 4 T S O 33 here as time in which students work in pairs or small groups.4 The "O" token is shorthand for "other," essentially an emergency category in the event that an in-class occurrence fits neither in the "T" nor in the "S" categories. The OG also includes a "Notes" section, as can be seen in Table 2, which allows for the researcher to take more detailed notes concerning in-class proceedings. Should the researcher using the OG then note, 6 minutes into a 50-minute class, that a teacher was lecturing in the students‟ native language, the researcher would circle the "T" token to indicate that class time at that moment was teacher-fronted (CAI 1). That the teacher was lecturing, and lecturing in the students‟ native language, cannot be captured by circling any of the three tokens. A "Notes" section was deemed compulsory to accommodate a more in-depth understanding of in-class English teaching and learning practices. Similarly, should a researcher note 16 minutes into the same 50-minute class that students are working in pairs (CAI 2), the researcher would then circle the "S" token. However, student-student interaction is defined as class time devoted to either pairwork or small-group work; that is, whether or not students are indeed working in pairs or small groups cannot be deduced from the "S" token. Therefore, the "Notes" column was again seen as a crucial component to the OG. 3.2.1.2 PEC. The PEC addresses CAI 3-the physical classroom setup. The PEC allows for the researcher to document, among other things, the patterns of desks in the classroom (e.g., semicircular formation or table seating) and the equipment or 4 Although some researchers go so far as to specify optimal task types for improving language acquisition in pairwork and small-group work (cf., "one-way" and "two-way" tasks in Long & Porter, 1985, for example), this study sets opportunities to work in pairs or small groups as a minimum proviso for a communicative approach to ELT. 34 resources available for both students and the teacher.5 Taken together, the OG and PEC yield data with which interpretations bearing on the presence of communicative approach fundamentals can be constructed (RQ 1), these fundamentals being those captured by the three CAIs. 3.2.2 A Rationale for Classroom Observations Classroom observations can offer more direct insight into teaching and learning practices. Estacion, McMahon, and Quint (2004) classify classroom observations as a resource-intensive method for obtaining data that provides valuable information that cannot be obtained in other ways. In their study, Estacion et al. investigated whether or not teachers were modeling or implementing certain cognitive and meta-cognitive learning strategies in the classroom. As Long (1990) mentions, "classroom observations consistently show little continuity in what teachers actually do and what they think they are doing" (p. 32). Teachers are often unaware of some of their teaching behaviors. As such, teacher self-reports, such as interviews, surveys, or questionnaires, run the risk of generating an inaccurate portrayal of in-class teaching and learning practices. This study makes recourse to observations in order to get at an accurate depiction of English teaching and learning practices indicative of a communicative approach in governmental and BRAC primary schools in Bangladesh. 5 The type or availability of resources for both students and teachers in primary school classrooms in Bangladesh does not figure prominently into the present study. Rather, as Bangladesh is a developing country, how resources might affect classroom proceedings, or, more specifically, the implementation of certain English teaching methodologies, was of ancillary interest to the researcher. 35 3.2.3 ACA The OG and PEC jointly provide evidence (or lack thereof) for all three CAIs- teacher-fronted time (CAI 1), student-student interaction (CAI 2), and the physical setup of the classroom (CAI 3)-which attest to the manifestation of a communicative approach to ELT in English language classrooms in governmental and BRAC primary schools (RQ 1). To determine whether or not there is a difference between the communicative ability of students enrolled in governmental versus BRAC primary schools (RQ 2), the ACA was developed. Referring back to the definition of communication above, communication, according to Canale and Swain (1980), can be realized only through real-time social interaction-"in the actual production and comprehension of utterances" (p. 6) with a minimum of two people (Cheng, 1980; Thompson, 1996). It is worth mentioning here that the notion of "real-time social interaction" as a face-to-face exchange is somewhat parochial and passé; in other words, social interaction takes place in asynchronous and virtual domains as well, such as via e-mail or social-networking sites. However, for the purposes of this study, Canale and Swain‟s limited definition of the medium through which communication occurs is useful. "Communication" is thus operationalized as (1) students ability to interpret a message and (2) students‟ ability to convey and interpret messages through interpersonal interaction. The interpretive and interpersonal abilities of communication speak to two of the three modes of communication defined by the Standards for Foreign Language Learning put forth by a four-way alliance of language organizations: the American Council on the Teaching of Foreign Languages, the American Association of Teachers of French, the 36 American Association of Teachers of German, and the American Association of Teachers of Spanish and Portuguese (National Standards in Foreign Language Education Project, 1999). The Standards for Foreign Language Learning outline what are called the Five C‟s of language learning-communication, cultures, connections, comparisons, and communities-wherein communication embodies interpretive, interpersonal, and presentational abilities. Given the anticipated English level of Bangladeshi ELLs at both governmental and BRAC primary schools, the presentational mode of communication was excluded from the operationalization of communication used in the development of this study‟s ACA. In other words, alluding to the expected-actual English proficiency gap highlighted in the Introduction of this study and having thoroughly examined the English for Today textbooks used by both BRAC and governmental primary schools, it was thought at the time the ACA was conceived that any demonstration or performance lending itself to the presentational mode of communication would be beyond the capability of such low-level learners. As such, multiple tasks were advanced for interpretive and interpersonal abilities but not the presentational ability. It could be helpful to view the overall interrelation of the above component definitions and their respective tasks as inclusive or hierarchic levels (cf., Figure 1). At the first level is students‟ overall communicative English ability. At the second level, and nested within the first, are interpretive and interpersonal abilities-the two bifurcations of communicative ability defined and operationalized in this study. To obtain usable data with which to address students‟ interpretive and interpersonal abilities-and hence their communicative ability-the five level-three tasks were developed: classroom commands, 37 Figure 1: The above figure represents a collusion of communicative ability (Level 1), its component definitions (Level 2), and the different tasks (Level 3) used to get at the first two levels. 38 body parts, and picture identification tasks for interpretive ability; and dialog and two information gap tasks for interpersonal ability. It is important to note that, although six tasks make up the ACA, there are only five groups of tasks because there are two information gap tasks. 3.2.3.1 Task development and description. Tasks were developed in consultation with the English for Today textbooks used in both BRAC and governmental primary schools, the Utah State Office of Education‟s Elementary Core Curriculum for World Languages 7-12, and with an eye to information-exchange (or information gap) task characteristics hypothesized to promote the negotiation of meaning, and thus language acquisition (cf., Ellis, 2000, among others). The English for Today textbooks were taken into consideration during ACA task development for multiple reasons, chief among these is the pervasive use of this particular textbook across both types of primary institution (i.e., in grades 2-5 in both urban and rural locales). The Elementary Core Curriculum for World Languages 7-12, embodying multiple goals and standards for equipping students with the requisite linguistic and cultural knowledge for modern, 21st-century communication, adopts a definition of the interpretation of spoken and written language that aligns closely with that of this study. For example, one component of standard 2 (level 1, goal 1) requires that language learners be able to "respond to routine requests in the classroom and in public places" (p. 17); as such, one interpretive task for the first component definition of communication used in the development of this study‟s ACA asks students to respond to English classroom commands. Other standards of the Core Curriculum for World Languages were acknowledged in the development of other ACA 39 tasks, as those standards were allied to definitions of interpretation and interpersonal communication similar to those adopted in this study. "Interpersonal communication" tasks (tasks four through six) were developed in consideration of those example tasks described in the Core Curriculum for World Language in addition to the information-exchange task characteristics described by Ellis (2000), these being tasks that… (1) have a required information exchange; (2) involve a two-way […] exchange of information; (3) have a closed outcome; (4) are not familiar to the interactants [here, Bangladeshi ELLs]; (5) involve a human/ethical type problem; (6) involve a narrative discourse mode; and (7) are context-free […] and involve considerable detail. (p. 200)6 Individual ACA tasks are described in more detail below. The administration of these tasks is laid out in the Procedures section, yet an overview of all ACA tasks offered here might allow for a better one-to-one mapping of the individual tasks and the literature or materials (e.g., English for Today textbooks and aspects of information-exchange tasks) perpended in their conception and creation. 3.2.3.2 Task one. Tasks one through three require that a particular message be interpreted by the concerned student. In task one, students hear a total of six classroom commands, such as "raise your hand" and "open your book." The student, in hearing a command, must respond accordingly. In response to the above two classroom commands, students (provided they understand) would raise their hand or open their book. (A more detailed explanation of the administration of this task and all subsequent tasks will be detailed in Procedures. The testing protocol can also be viewed in the 6 It should be noted that not all of these task characteristics are embodied in the information-exchange-type tasks used in the ACA. The students who assented to take the ACA are young students (i.e., roughly between 6 and 15 years of age) with little previous exposure to English; thus, task characteristic (6) involving narration, for example, was not characteristic of information-exchange-type tasks used in the ACA. 40 Appendix.) Responses are recorded as being either correct or incorrect on a response sheet. 3.2.3.3 Task two. In task two, students hear a total of eight commands, wherein they are required to touch different parts of their body. For example, the researcher might say "Touch your head." Students (provided they understand) would then touch their head. Other body parts used in this task include "eye," "ear," "nose," "hand," "mouth," "foot," and "leg." Responses are recorded as being either correct or incorrect on a response sheet. 3.2.3.4 Task three. For this task, students are given a laminated card with six rows of images, three images per row (six potentially correct/incorrect answers). In each row, one image is missing (the rightmost or third image). A number of individual laminated cards were also provided. The researcher says the name of each laminated image in each row. The student, upon hearing the name of the missing image, chooses the image from amongst the individual ones and places it in the blank space provided. Responses are recorded as either correct or incorrect on a response sheet. 3.2.3.5 Task four. The fourth task is an introductory dialog where the student is required to interact-or "meet"-the researcher. The researcher is tasked with initiating verbal exchanges, although students are free to be creative with their responses and can, if so desired, ask questions of the researcher. There are a total of six "turn-starters," and turns can be skipped should the student‟s response render a subsequent turn-starter inappropriate. Student responses are recorded as either appropriate or inappropriate and documented on a response sheet. 41 3.2.3.6 Task five. This is an information gap, or information-exchange-type activity. The researcher is given sheet A and the student sheet B. On each sheet are images, some of which are colored and some are not. Sheet A has the colored images that are not colored on sheet B. Sheet B has the colored images that are not colored on sheet A. A handful of crayons or colored pencils are placed on the table between the student and the interviewer. The interviewer and student must help each other to color the images the correct color. Responses are recorded as either correct or incorrect; however, as this is a two-way communicative task, how students interacted with the researcher to arrive at a correct or incorrect response was closely observed and documented. 3.2.3.7 Task six. Task six, like task five, is an information-exchange-type activity. For this activity, students are given a set of pictures: four large pictures and two smaller pictures. Each large picture has a blank space either "in" or "above" (or "on") it where each of the two smaller pictures can be placed as per the researcher‟s instructions. The student also has two smaller pictures already placed either "in" or "above" two of the four large pictures. Additionally, the researcher has four large pictures and two smaller pictures. On two of the researcher‟s large pictures, two small pictures were placed either "in" or "above" the large picture (those held by the student that have not been placed on either of the two larger pictures). The researcher has two smaller pictures in hand (those that have been placed on two of the student‟s large pictures). The researcher and student must help one another to place the pictures they have in their hands in the correct location on the large pictures. Responses are recorded as either correct or incorrect; however, as 42 this is a two-way communicative task, how students interacted with the researcher to arrive at a correct or incorrect response was closely observed and documented. 3.2.4 A Rationale for the ACA Other means for assessing governmental and BRAC primary school students‟ communicative proficiency, including assessments, were examined before the decision was made to develop an assessment (i.e., the ACA) specifically attuned to the needs of this study. Given the population of primary school students assessed-Bangladeshi ELLs-and the aforementioned definition of communication adopted in this study, other means for assessing students‟ communicative proficiency were disregarded. A broad-strokes proficiency assessment, for example, could not be adapted for this particular population in this particular environment (broadly, Bangladesh in governmental and NGO-fronted primary educational settings). 3.3 Participants 3.3.1 BRAC Primary School Participants BRAC students fall into two categories: those who were present during a classroom observation (both students and their teacher), and those who took part in one-on- one interviews, wherein the ACA was administered. Beginning with those present during classroom observations, six English classes were observed: two grade 2 classes, two grade 4 classes, and two grade 5 classes.7 These classes were located in schools in 7 These classes were chosen for several reasons: (1)-students in BRAC primary schools do not receive English instruction in grade 1; and (2)-grade 3 classes were not presently being held in the BRAC catchment area in which the researcher was conducting classroom observations, as grade 3-aged students were either not available, not sufficient in number, or not at a grade 3 education level. 43 the capital of Bangladesh, Dhaka, in the Kurail Slums and Middle Badda areas. Each class held between 25 and 33 male and female8 students between the ages of 9 and 15. Students had varying amounts of exposure to English. For example, those in grade 5 who previously attended a BRAC primary school took English courses in grades 2-4. Those in grade 2 do not receive English instruction in grade 1. All students were native speakers of Bengali coming from poor families in the neighboring community. Five students from each class were chosen at random by their teacher to take the ACA (17 female and 13 male), being 9-15 years old (M = 11.03 years). Six teachers also participated in this study by consenting to have their English class observed by the researcher. All six teachers were female between the ages of 20 and 35. Teachers are native speakers of Bengali who reside in the local community. 3.3.2 Government Primary School Participants Government primary school participants, like BRAC participants, fall into two categories: those who were present during a classroom observation (both students and their teacher), and those who took part in one-on-one interviews, wherein the ACA was administered. Beginning with those present during classroom observations, three English classes were observed: one grade 2 class, one grade 4 class, and one grade 5 class at the National Government Primary School in the Mirpur area of Dhaka. Each class held between 34 and 52 male and female students from 6-11 years of age.9 Students had 8 This is worth mentioning because some schools in Bangladesh-namely, the madrasahs-are all-male schools. 9 The difference in age range between BRAC and government primary schools is accounted for by the nature of BRAC‟s nonformal education program. 44 varying amounts of exposure to English. All students were native speakers of Bengali coming from poor families in the neighboring community. Five students from each class were chosen at random by their teacher to take the ACA (5 female and 10 male). Students were between the ages of 6 and 11 (M = 9.26 years).10 Three teachers also participated in this study by consenting to have their English class observed by the researcher. All three teachers were female between the ages of 20 and 35. Teachers are native speakers of Bengali who reside in the local community. 3.4 Procedures 3.4.1 Classroom Observations Following discussion with the relevant teachers, head teachers, and other school administrators at either BRAC or governmental institutions, an observation schedule was outlined. The schedule detailed the start times of English classes, the grade to be observed, the date on which a particular class and grade would be observed, and the location of the class.11 When a given English class was scheduled to be observed, the PEC was completed 5-10 minutes before the start of the class. Using a clipboard and pencil, the formation of desks or seating and the availability of classroom resources was documented 10 The age range between BRAC and governmental primary-school student, as was mentioned previously, is one of the differences between these two types of institutions. Without moving into too detailed a discussion of linguistic features and language acquisition (cf., Larsen-Freeman, 1975), or the relationship between age and acquisitional "ability" (cf., Lenneberg, 1967), this study would like to acknowledge the fact that, in the case of both older (e.g., a 15-year-old BRAC participant) and younger (e.g., a 6-year-old governmental participant) participants, interpretations could be made in favor of either amount of exposure to English, on one hand, and age as it pertains to a decline in acquisitional "ability" on the other. 11 In the case of BRAC, there is often not a single school structure that houses grades 1-5, for example. Instead, grade 2 might be in one location and grade 4 in another around Dhaka. 45 on the PEC. After students arrived to class, the researcher gave a short introduction of himself and spoke informally with students in their native language (Bengali) in order that they might feel more comfortable having a foreigner present during their English class. During this time, the purpose of the researcher‟s visit and observation were also explained to students at an appropriate level in Bengali; stock jargon in the field of language acquisition, for example, was not used. The researcher then conferred with the teacher in question concerning where would be the best place to sit and observe the upcoming English class. Once an English class started, the researcher documented classroom proceedings every 2 minutes on the OG using a pen or pencil. Two-minute intervals were tracked using a stopwatch. As for the students who assented to take the ACA, five were chosen at random by the teacher whose English class was observed prior to the classroom observation. These students were interviewed one by one either at the back of the classroom or in an open area just adjacent to the classroom.12 Interviews lasted anywhere from 12-25 minutes. There was considerable variation in test-taking duration depending on a given student‟s communicative English ability and the grade in which the student was enrolled; in other words, a grade 5 student who enrolled in a governmental primary institution at grade 1 most probably has a larger bulk of English exposure compared to a grade 2 student at the same institution. 12 See later discussion regarding the intersection of Western research methods with Bangladeshi culture. In many cases, the one-room classroom was the only viable option in which students could take the ACA. One-on-one, student-researcher interaction in an isolated location is contra Bangladeshi cultural norms, and, in the crowded Dhaka-city area, altogether impossible. With all of the one-on-one student interviews (i.e., ACA administration), both at BRAC and governmental primary schools, onlookers were present. This is a mere one of a myriad of difficulties that Western researchers are likely to confront in similar, notably "non-Western" countries or other area. 46 The ACA was administered individually either on the floor or at a small desk-type area afforded by the school (e.g., a trunk or short table). Once students sat down with the researcher, it was explained to them-in Bengali-that they were going to be asked a few questions and told a few statements in English and that they could discontinue taking the ACA at any point should they feel uncomfortable. Additionally, students were told, in the event that they did not understand an English question or statement, that they could respond with "I don‟t know," "I can‟t," or "I don‟t understand" in Bengali, and that failure to respond appropriately or accurately to an English question or statement was by no means life-threatening. Students‟ responses were recorded on a response sheet with sections delineated for each task, and tasks were administered in random order. At the beginning of each task, instructions for completing the task and an example or model were provided to students. Students were encouraged to take their time in responding and could ask for a given English question or statement to be repeated up to three times. Individual task instructions and their respective examples or models (given in Bengali, of course) were spoken according to a script13 to eliminate inconsistencies in administering the ACA from student to student. 13 The ACA script was developed under the help and guidance of Md. Shamim Yusuf in the BRAC Education Program (BEP) at BRAC headquarters in Dhaka, Bangladesh. CHAPTER 4 RESULTS AND DISCUSSION 4.1 Overview I now turn to a discussion of the research findings of this study, proceeding chronologically with RQs in the order in which they were presented in Chapter 2. This section in addition highlights instances of intersection between what might (loosely) be termed Western research methods with a notably non-Western Bangladeshi culture. In other words, what some might call problems from a certain perspective others may very well respect as areas in which Western researchers need to shoulder a certain degree of flexibility. Diving into these junctions can be like opening a can of worms; nonetheless, I believe their presence conveys a certain weight-not only in the interpretation of results to follow but for all Western researchers entertaining ideas of research in Bangladesh (particularly in the educational sector)-and thus merits here space for consideration. 4.2 RQ 1 Has a communicative approach to ELT taken hold in English language classrooms in governmental and BRAC primary schools, as evidenced by the following three indicators: (1) teacher-fronted time, (2) student-student interaction, and (3) the physical setup of the classroom? 48 Any answer to the above question rests on certain amount of subjective interpretation and can be arrived at only by way of description. To put it another way, while many researchers agree that teacher-fronted time (Breen & Candlin, 1980; Canale & Swain, 1980; Cheng, 1980; Long 1990; Long & Porter, 1988), student-student interaction (Allwright, 1984; Blumenfeld et al., 1996; Cheng, 1980; Ciotti, 1969; Ellis, 2000; Gass & Varonis, 1985; Horne, 1970), and classroom setup (Littlewood, 1981) are basal in communicative approaches to ELT, to the best of my knowledge any quantification of the extent to which each of these must be present (or the proportion of class time devoted to each) in a given English class has not been attempted. For example, in a 50-minute class, how much small-group work or pairwork must be present for this class to align with a communicative approach? Is 10 minutes of two-way information exchanges in small groups sufficient? Although the perfect communicative approach recipe calling for small-group, pairwork, and physical setup ingredients has not been proposed (and any attempt to do so would hopefully, at least initially, cause a few eyebrows to be raised), I believe that advocates of this particular approach or language methodology votaries can rest assured that some student-student interaction must be present in a given class for said class to be communicative in nature. Additionally, while the ideal desk formation in a communicative approach to ELT is elusive, rest assured that particular arrangements could be agreed to work against getting students to work in pairs or small groups. Having said as much, results show that teacher-fronted time (CAI 1) is still the predominant mode in both BRAC and governmental primary school settings. Starting 49 with BRAC schools, it can be seen from Table 3 that the overwhelming majority of class time is teacher-fronted. Indeed, of the six BRAC primary school English classes observed, two of these (one grade 2 and one grade 4) made use of pair- and small-group work; four of these classes were entirely teacher-fronted, with no talk time elicited from pairs and small groups. Of the government primary school English classes, although fewer were observed, two were entirely teacher-fronted, while one (grade 4) allocated roughly six percent of class time to pair and small-group work. To revisit the above RQ, the information provided in Table 3 sheds light on two of the first three CAIs-(1) teacher-fronted time and (2) student-student interaction. As student-student interaction is defined as time spent either in pairs or small groups, it can be seen from the above data that both BRAC and governmental primary school English classes are intensely teacher-fronted. Turning now to the third CAI-(3) the physical setup of the classroom-a look at the class-to-class desk arrangements noted on PECs reveals that none of the BRAC classrooms contain desks; instead, students sit in a U-shape arrangement on the floor. However, students in government primary school English classes all sit at small picnic-style tables, at which four to six students can be comfortably seated. It should be noted that the U-shape seating in BRAC classrooms does not forcibly constrict the administration of pair and small-group activities. Long (1990) reminds us that that "badly organized group work is not better than badly organized lockstep work" (p. 37). On the same note, the picnic-style seating assumed in governmental primary schools does not necessarily imply that a communicative approach to ELT is at play in a given English class. Instead, U-shape seating might flow from a lack of funding for 50 Table 3: Classroom Minutes School (Grade) Min. (T)14 Min. (S) % Min. (T) % Min. (S) BRAC (2) 40 6 86.95 13.04 BRAC (2) 48 0 100 0 BRAC (4) 40 0 100 0 BRAC (4) 38 4 90.47 9.52 BRAC (5) 28 0 100 0 BRAC (5) 42 0 100 0 Gov. (2) 30 0 100 0 Gov. (4) 32 2 94.11 5.88 Gov. (5) 30 0 100 0 14 The actual class duration in minutes-as opposed to just the proportion-was provided to inform the reader of the fact that the durations varied from one class to the next. 51 purchasing desks and picnic-style seating in government schools might contrarily betray an awareness of the importance of pair and small-group work in communicative English classes, in spite of which there is a notable absence. In sum, though the amount of class time allocated to pair and small-group work necessary for a communicative English class is not specified, it can be argued that, in order for an approach to ELT to be communicative, the class should not be completely devoid of pair and small-group activities, as they are in four of the six BRAC English classes observed and two of the three government ones. Additionally, desk arrangements in BRAC classrooms (U-shape) do not incontrovertibly contravene the implementation of pair and small-group activities-yet, taken at face value, the data gleaned during classroom observations (i.e., the OG and PEC) would lend itself to the conclusion that a communicative approach to ELT has not altogether rooted in either BRAC or governmental primary school English classes. Nonetheless, there was evidence of student-student interaction (CAI 2), though little in relation to the proportion of teacher-fronted time. Backing the above claim is a picture of in-class proceedings stemming from a closer look at notes taken during classroom observations. Recall that a "T" token signifying a 2-minute interval as "teacher-fronted" does not inform us as to what exactly was happening at the time the token was circled; that is, was the teacher in question lecturing? Or, were students working individually in their notebooks or on their slateboards? After obtaining the overall proportion of teacher-fronted time and student-student interaction, notes taken at 2-minute intervals were coded then analyzed for 52 frequency. In other words, this subanalysis sought to finger the predominant activities that contributed to an imperiously teacher-fronted English class. Table 4 reveals what was happening most frequently when in-class proceedings were noted at 2-minute intervals. The most frugal contributor to teacher-fronted class time in BRAC English classes was individual work, which covers here students writing (I-W) in their notebooks or on their slateboards and reading individually (I-R). Lecture, which covers teacher-fronted question and answer sessions and explanations in Bangla of particular English sentences or grammatical constructions, constituted over a third of the total class time. In governmental classes, individual work is also the largest consumer of class time, contributing 38% to the total teacher-fronted time. Of the two subsets of individual work, the principal task was, again, individual writing activities in notebooks or on slateboards. The large chunk of class time apportioned to lecture and individual work smacks of older, perhaps more traditionally teacher-fronted language teaching methodologies, such as the ALM and Grammar Translation Method. 4.3 RQ 2 Is there a difference in the communicative ability of students enrolled in governmental versus BRAC primary schools as measured by the ACA? To determine whether or not a difference exists in the communicative ability of these two populations of students, ACA data will be analyzed at three different levels: overall communicative English ability (Level 1), interpretive and interpersonal task groups (Level 2), and the various tasks within the interpretive and interpersonal task groups (Level 3). 53 Table 4: Lecture and Individual Work BRAC Gov. Lecture 35% 34% Individual (I) 48% 38% I-W 27% 26% I-R 18% 9% 54 Beginning with overall communicative ability (Level 1), an analysis of the results of the ACA show (cf., Figure 2) that Bangladeshi ELLs enrolled in governmental primary schools are significantly better at communicating in English. The group difference is significant across all ACA tasks (Mann-Whitney U=131.5, p=.024)15 with BRAC students obtaining 80% correct on the ACA and governmental students 87%. The definition of communication, as operationalized in this study, is comprised of two component definitions: students‟ ability to (1) interpret messages and their ability to (2) communicate messages at an interpersonal level. Taken together, interpretive and interpersonal abilities make up students‟ overall communicative English ability (Level 1). However, just as a particular token of the OG glosses over more specific information pertaining to in-class proceedings, the above information perhaps masks informative data at the task and task-group levels (i.e., Levels 2 and 3). Therefore, whereas it can be shown that Bangladeshi ELLs enrolled in BRAC primary schools are altogether more communicatively adept than governmental ELLs, closer inspection of student performance at the task and task-group levels can add dimension to an understanding of English teaching and learning practices in these two types of primary institution. Turning to a Level 2 analysis of student performance of interpretive and interpersonal tasks, results show (cf., Figure 3) a significant difference in interpretive ability between BRAC and governmental students. The group difference is significant at p=.012 (Mann-Whitney U=121.5) for interpretive tasks (not for 15 As parametric statistical test procedures involve population parameters, and given the uneven samples across BRAC and governmental students (and observed classes), the nonparametric Mann-Whitney U was used in the analysis of ACA data. 55 Figure 2: Mean proportion correct averaged across all tests by school type; bars represent +/- 1 SD. 56 Figure 3: Mean proportion correct averaged across all tests by interpretive and interpersonal task groups for school type (+/- 1 SD). 57 interpersonal tasks: p=.088) with BRAC students obtaining, on average, 83% correct on the ACA and governmental students 90%. These results are interesting because, while students might perform better on interpretive ACA tasks than interpersonal ones, it is the joint performance on both task groups that yields an overall "high" or "low" communicative ability; in other words, the interaction between the two is important and should contextualize any interpretation of the results. Moving from a look at interpretive-interpersonal task-group results, I now turn to an even closer look at individual task results-the students‟ performance on the various commands, picture identification, body parts, dialog, and information gap tasks. Recall that the interpersonal task group is composed of information gap tasks (of which there are two) and the dialog task, and the interpretive task group is composed of commands, picture identification, and body parts tasks. If we compare the results between each task across BRAC and government populations (cf., Figure 4), findings are insignificant for body parts (Mann-Whitney U=200.5, p=.503), picture identification (Mann-Whitney U=203, p=.527), dialog (Mann-Whitney U=195.5, p=.437), and the two information gap tasks (Mann-Whitney U=158, p=.104). However, results are significant for classroom commands (Mann-Whitney U=111, p=.004). Based on these findings, it is thus students‟ performance on the classroom commands interpretive task that carries the effect of significance across Levels 1 and 2. It should be reiterated here that it is the interaction of student performance on all tasks that contributes to an overall significance at higher levels. Again, any interpretation of results of classroom commands task performance stipulates consideration for the overall context of the ACA. To put it more plainly, while classroom command task results carry 58 Figure 4: Mean proportion correct averaged across all tests by task for BRAC and governmental primary schools (+/- 1 SD). 59 the significant effect across higher levels, in any interpretation of students‟ communicative English ability, performance on this task cannot be altogether disengaged from performance on other tasks. The objective of this portion of the study was not to target interpretive ability and interpersonal ability as two separate abilities, but as interpretive and interpersonal abilities as equally important and mutually influential constituents of communicative ability. The above analyses look temporarily at individual levels in isolation, but in any big-picture conclusions refer to the interrelatedness of all three levels. In answer to RQ 2-the difference in communicative ability of BRAC and governmental primary school students-governmental students are significantly better at communicating in English, insofar as the ACA gets at the heart of communicative English ability. Furthermore, governmental students are significantly better at interpretive tasks than BRAC students, and, even more narrowly, at interpreting classroom commands. 4.4 RQ 3 What is the relationship between students' scores on the ACA and the three communicative approach indicators identified as essential to ELT within the framework of a communicative approach? Figure 5 gives some indication as to the relationship between students‟ scores on the ACA and CAIs. In order to obtain "plotable" data for these indicators, something other than descriptive data was needed. Thus, a composite score was developed that assigned values to both teacher/student talk time and the physical classroom setup. If a 60 Figure 5: Test proportion correct against composite communicative indicators by class; (G) signifies governmental, while (B) and (b) signify BRAC primary schools (e.g., grade 2 (b) is one grade 2, while grade 2 (B) is another). 61 given class had a nongroup desk arrangement, a 0 was assigned to this indicator (CAI 3). If a given class had a desk arrangement that promotes group work, a 1 was assigned to this indicator. This value was then averaged across the proportion of student-student interaction (CAI 2), which lies in complementarity to the proportion of teacher-fronted time (CAI 1). For example, if an English class observed was bereft of pair or small-group activities, the proportion of student-student interaction would thus be 0. However, if this same class had group-type desk arrangements (e.g., picnic-style seating, like that of the government classes), it would also receive a score of 1, that, when averaged with 0, would leave this particular class with a composite communicative indicator of 0.5. Similarly, if a class had group-type seating (earning it a value of 1) and devoted 50% of class time to pair and small-group activities (earning it a value of .5), this class would receive a composite score of .75. In this way quantifiable data was obtained in which to draw a comparison between students‟ ACA scores and the presence of CAIs in an English classroom. Unfortunately, due to the odd and fairly small sample size of observed English classes (i.e., three government classes and six BRAC classes), neither parametric nor nonparametric statistics could be used to determine whether or not there is a statistically significant relationship between the presence of indicators and students‟ ACA scores. However, despite this inconvenience the above data (and that presented in response to earlier RQs) seems to confer in a specific direction. Referring back to RQ 2, it can be shown that governmental students are significantly better at communicating in English than BRAC students. With this in mind, consider the grouping of governmental versus BRAC classes in Figure 5. All observed 62 governmental classes received a composite communicative score of at least 0.5, whereas BRAC scores cluster between 0 and 0.1-relatively considerably lower composite scores. Knowing that governmental students are significantly better at communicating in English, and governmental classes received high composite ratings, the data seems to suggest that there is a positive relationship between the communicative ability of students and the presence of CAIs. In other words, the significant difference in communicative ability on behalf of governmental primary school students seems to originate from more "established" CAIs in the classroom, including desk formation and the proportion of teacher-fronted versus student-centered talk time. In summary, while results are not conclusive, there does indeed appear to be a positive relationship between indicators and students‟ communicative English ability. 4.5 Western Research Methods and Bangladeshi Culture In this section, I would like to comment briefly on instances of intersection between what one might coin Western research methods with a notably non-Western Bangladeshi culture encountered during this research. I originally thought that administering my ACA would be a simple task; however, carrying out ACA testing protocol proved problematic at multiple junctures due to the disjoint between Western methods of testing and Bangladeshi culture. In other words, Bangladeshi culture (not to be backgrounded) conflicts with Western testing methods (cf., Hamid, 2010, for more Bangladesh-specific information on reconciling Bangladeshi culture with Western empirical research methods). 63 One problem encountered centers on the physical testing environment. Western testing methods would require as many variables as possible to be controlled in an ideal testing environment. Take for example a one-on-one, interviewer-student test. The ideal version of this test according to Western standards would most likely control for excessive ambient noise (e.g., construction noise external to the testing room) and outsider interference among other factors. However, if we transpose this type of test into a Bangladeshi primary classroom setting, these variables would be all but impossible to control for. In the primary classroom (at least in the BRAC setting), testing must be conducted in the one-room classroom, as this is where students are most comfortable, and testing carried out external to the classroom would simply draw a larger crowd of bystanders (the second variable). This type of interview conducted in the one-room classroom setting in Bangladeshi culture leads to the following problems: (1) other students peer over the shoulder of either the interviewer or the student, and (2) testing protocol interferes with classroom proceedings and has an effect on both student attention and the teacher‟s ability to teach. Both of the above problems are crucial for several reasons. If students observe another student taking the test (problem [1]), it could become difficult to tease out what the subsequent students know and are able to demonstrate during the test from what they may have gleaned during their observation of the previous student. Furthermore, with the knowledge that other students (and oft the teacher) are watching, the student may find it difficult to speak openly what he or she does and does not know, refrain from looking to observing students or the teacher for help, or feel uncomfortable or pressured. 64 Problem (2) is also crucial for those conducting classroom-based research in Bangladesh, as networks and partnerships are so important for ensuring that research proceeds smoothly. If, for instance, a researcher incurs a particular reputation among a certain population, whether it be among school teachers-or, much worse, other researchers, school administrators, or NGOs-future research may not be allowed or will otherwise be strongly discouraged by one of these populations. A non-Bangladeshi researcher‟s presence could interfere with classroom proceedings in yet another way. As more teacher-centered approaches were used prior to a communicative approach (i.e., perhaps even less pair and small-group activities were employed), Bangladeshi English teachers might consider teacher-centered activities, such as lecture and question-and-answer sessions, more appropriate for the language classroom in general. To put it another way, these teachers might consider certain activities more conducive to an effective learning environment and others less so, and the researcher‟s presence in the classroom could have led these teachers to adopt more teacher-centered activities on account of their being more indicative of effective teaching and learning in their eyes. It could very well be the case that, on a typical observer-free day, more pair and small-group activities are used proportional to teacher-fronted ones. Apart from the problems presented within the physical testing environment, another problem encountered concerns the use of accent. Because teachers in BRAC primary schools are predominantly women from the local community (Lovell & Fatema, 1989), few (if any) have received education at the university level or have spent time abroad in which their English proficiency developed. That being said, teachers spoke English during their English lessons with "heavy" Bengali accents. As a researcher from 65 the U.S., and a native American English speaker, I speak with an American English accent (i.e., not British or Australian English). Students may have very well heard an American English accent for the first time during the ACA testing administration. On more than one occasion, I was interrupted by an observing teacher, who would repeat what I had said in her Bengali English accent. For example, if I were to say "touch your eye," the observing teacher would repeat this command for the student to hear. And, quite frequently (albeit not surprisingly), this was easier for the student I was interviewing to understand. After enough interruptions from observing teachers, I tried out a few commands in one of my test tasks in a Bengali English accent-and I found that students found my English much easier to understand. In the case of classroom commands, after hearing a Bengali English classroom command, the student in question would do what he or she heard. These were two problems confronted during testing administration that I believe come to bear on the results presented above or that should, at the very least, be held in tow in their consideration. From this test-administration experience, several take-home messages can be elicited. First, Western testing methods, at least those pertaining to one-on- one interviews, are not fully compatible with Bangladeshi culture, as exemplified in one-room classroom situations (i.e., BRAC primary schools). Second, Western researchers should be fully aware of the context in which they hope to conduct their research. As an example, my American English accent proved problematic for the young Bangladeshi primary students. Therefore, an American researcher unable to assume a Bengali English accent might decide that, in asking a certain question, students do not know the answer. However, this is not forcibly the case. Students may very well know 66 the answer, but due to problems of exposure, may never before have encountered an American accent (or any other, for that matter). They might simply not understand a particular question on account of the researcher‟s accent. Lastly it is my strong belief that researchers operating within a Western paradigm need to be flexible in conducting research in Bangladesh. Western research methods may clash with Bangladeshi culture, and so the researcher must be willing to forgo the ability to control for all extraneous variables during data-collection procedures. CHAPTER 5 CONCLUSION 5.1 Take-Home Messages This study provides an initial take on the ELT situation in BRAC and governmental primary school classrooms in Bangladesh, attempting to document for the first time whether or not a communicative approach to ELT, which the GoB advocates, is manifest in English classes. What merits such an attempt, despite acknowledged differences between BRAC and governmental primary schools, is the prevalence of these two types of school across the country. With so many students attending either BRAC or governmental primary schools, an investigation into the quality of English language classes was called for-the timely question of quality versus quantity. It was established in the review of the literature that certain cornerstones must be in place before a communicative approach can be adopted in ELT: (1) teacher-student interaction (Breen & Candlin, 1980; Canale & Swain, 1980; Cheng, 1980; Long 1990; Long & Porter, 1988) (2) student-student interaction (Allwright, 1984; Blumenfeld et al., 1996; Cheng, 1980; Ciotti, 1969; Ellis, 2000; Gass & Varonis, 1985; Horne, 1970) , and (3) the physical setup of the classroom (Littlewood, 1981). These cornerstones were operationalized as CAIs to guide in the documentation and collection of data for this study. 68 The goals of this study were to document the presence of CAIs, gauge students‟ communicative English ability via the ACA, and examine the relationship between ACA scores and indicators. The findings of this study demonstrate that English classes in both BRAC and governmental primary schools are largely teacher-fronted (RQ 1), that governmental students are significantly better at communicating in English than BRAC students (RQ 2), and, in spite of irregular sample sizes, that there seems to be a positive correlation between the presence of CAIs and students‟ English communicative ability as measured by the ACA (RQ 3). This study also brought to light several hurdles encountered during ACA test administration and that future researchers operating within Western paradigms are likely to encounter. First among these are, depending on the perspective assumed, problems that can arise from the physical environment in which testing (namely one-on-one, interview-type testing) takes place, and, secondly, issues of accent or exposure. These hurdles lead me to put forth several take-home messages that I believe any Western researcher pursuing similar research in Bangladesh would be unwise to fail to consider. First, Western research methods may not be fully compatible with Bangladeshi culture, especially those that avail of similar testing procedures, in that certain variables that we as Western researchers are urged to consider are there unavoidable. Secondly, before research is undertaken in Bangladesh, the researcher needs to imagine as thoroughly as possible how the proposed research program will jibe with the local culture or context. To cite an example from this study, the researcher‟s own American English accent was foreign to students and led to inconsistencies in ACA administration. Lastly-and most importantly-those wishing to conduct research in Bangladesh need to 69 be flexible, forgoing the perhaps compulsive desire to control for things like ambient noise and onlooker presence during interviews and other data-collection procedures. That or the ambit of Western research methods need to be reconsidered, modified, or re-evaluated for or in consideration of certain contexts-here, Bangladesh. 5.2 Suggestions for Future Research In light of the findings of this study, and having examined these findings and problems encountered during data collection, I would like to briefly suggest directions for future research. First among these is a study that replicates the one presented here-a study that documents the presence of indicators and students‟ communicative ability in both urban and rural settings. The participants in this study were primary-school students at BRAC and governmental schools in Dhaka, the capital city (i.e., urban). Similarly, any replication should also seek to incorporate more schools of both type in both urban and rural locations. One of the shortcomings of this study, particularly as it pertains to the correlation between the presence of a communicative approach and students‟ communicative ability (the third RQ), is that the small sample size prevented use of parametric and nonparametric inferential statistics; that is, no statistical formula could be used to obtain either significant or insignificant results. As such, to what extent the findings of this study are generalizeable across governmental and BRAC primary schools is uncertain, and results presented herein should be cautiously interpreted. Secondly, many of the problems encountered during data collection stem from the interplay between Western research methods and Bangladeshi culture. Some of these problems were anticipated (e.g., a slow-moving and slack bureaucracy) while others were 70 unforeseen (e.g., the American English accent issue). Given these problems, more research is needed that is sensitive to the joint where Western research methods and Bangladeshi culture come together. More studies on this topic could lead to more "valid" research in the future-or, at the very least, the smoother implication of research design and data-collection procedures. A third suggestion is for studies that examine educational policy or the implementation of a communicative approach to ELT somewhere between policy decision-makers (the highest link[s] in the chain) and the English language classroom (the bottom-most link[s]). For example, the few pair and small-group activities observed in this study might not be attributed to any fault of the English teachers in question to administer pair and small-group activities, but rather to those responsible for teacher-training. Teacher trainers might entertain inaccurate or incomplete notions of what constitutes a communicative approach to ELT (and how to teach communicatively) that they subsequently pass on to English teachers in governmental and BRAC primary schools. Relatedly, this could be a classic cart-before-the-horse example-that is to say, BRAC and governmental primary school classrooms might not be ready for a communicative approach to ELT, given the recent shift in English teaching and learning paradigms. More studies are needed then that assay the appropriacy of such an approach to this particular ELT context. Lastly, studies that examine the relative "strength" or "presence" of a communicative approach in English classrooms as measured by other indicators would be beneficial. One shortcoming of this study, looking to the results of the third RQ, is that the value assigned to the presence or absence of group-promoting desk arrangements (in 71 coming up with a composite communicative score) is somewhat heavy, thereby regulating the gap between the grouping of BRAC and governmental classes (see Figure 5). A study that replicates this one with more BRAC and governmental schools could help in this respect. Table 3 shows that the proportion of student-centered time in the one governmental class (i.e., class 4) is less than the proportion of student-centered time in the BRAC classes (classes 2 and 4), viewed separately. One would thus anticipate these BRAC classes to sport a higher composite communicative score; yet, this value averaged with a 0 for nongroup-promoting desk arrangements is responsible for the left-edge grouping of all BRAC classes in Figure 8 (and the right-edge grouping of governmental ones). In sum, the present study was vital for a closer examination at the ELT situation in Bangladesh. It provided a useful look into both classroom proceedings in BRAC and governmental classrooms and students‟ actual communicative ability in English. Perhaps most importantly, this study, very exploratory in nature, caused a myriad of questions to rise to the surface concerning ELT at the primary level in Bangladesh. These questions will hopefully inspire and carry future research in Bangladesh, yielding answers that can better assist in decisions regarding English educational policy, teacher training, and instructional practices across the whole of the educational cycle, from nongovernmental to governmental and from madrasah to private education. APPENDIX Testing Protocol Interpreting a Message Task I Testing Instructions: For this task, a book and a picture are laid out in front of the participant. The participant is told that he or she will hear some English commands. The participant must respond to the command. The participant does not need to say anything for this task, simply obey the command they hear. This task should also be modeled for the participant. The researcher says, for example, "Open your book." The participant then opens his or her book. Below is a list of classroom commands. Participants may take as much time as they need to respond, and the command may be repeated once. Commands should be read at random.  Open your book  Close your book  Look at the picture  Raise your hand  Stand up  Sit down Task II Testing Instructions: For this task, participants are given another set of commands. This task deals with understanding commands in relation to different parts of the body. Participants are given a command, such as "Touch your ____." Participants must then touch the part of their body in question. This task should also be modeled for the participant. The researcher says, for example, "Touch your head." The participant then touches his or her head. Below is a list of parts of the body. Participants may take as much time as they need to respond, and the command may be repeated once. Commands should be read at random.  eye  ear  nose  hand  mouth  foot  leg  head 73 Task III Testing Instructions: For this task, participants are given a piece of paper with six rows of images, three images in each row. In each row, one image is missing. Participants must listen to the researcher and place the correct image (that read by the researcher) in the missing space in each row. Each row is comprised of a set of related images. For example, one row might have images of a cow and a horse. The third image-the missing one-might be that of a sheep. This task should also be modeled for the participant. The researcher reads, for example, "Cow, horse, sheep." The participant chooses the image of the sheep from a store of images and places it in the missing space. Participants may take as much time as they need to respond, and the list may be repeated once. Missing images should be random. Interpersonal Communication Task IV Testing Instructions: For this task, participants are led through a common greeting exchange, or greeting dialog. This task requires the participant and the researcher to interact with one another. The researcher asks the participant a series of questions, questions that are typical of first-time greetings, such as "How are you?" The participant then provides a response and possibly a follow-up question, such as "Fine. And you?" This task should also be modeled for the participant. The researcher says, for example, "How are you?" The participant then responds. Below is a script of questions outlining a greeting exchange.  R: "Hello!" S: [student response]  R: "How are you?" S: [student response]  R: "What‟s your name?" S: [student response]  R: "How old are you?" S: [student response]  R: "What class are you in?" S: [student response]  R: "Bye!" S: [student response] Task V Testing Instructions: This is an information gap task. The researcher is given sheet A and the participant sheet B. On each sheet are images. Some of the images are colored, some are not. Sheet A has the colored images that are not colored on sheet B. Sheet B has the colored images that are not colored on sheet A. A handful of crayons or colored pencils are placed on the table between the participant and the researcher. The researcher and participant must help each other to color the images the correct color. This task should 74 be modeled for the participant. For example, the research says "Color the mango yellow." The participant then colors the mango yellow on their sheet. The participant can then say, "Color the apple red." The researcher then color the uncolored apple on his or her sheet red. Commands should be read at random. Task VI Testing Instructions: This is an information gap task. The researcher is given sheet A and the participant sheet B. On each sheet are pictures of a box, a hut, a bus, or a bag. There are two blank spaces in and above each picture. These spaces are provided for participants to place images either in or on the picture in question. Each sheet (A or B) has two pictures with the image either in or on the picture in question. Without showing the sheet to one another, the researcher and participant must help one another to put the image either in or on the relevant picture (e.g., to put the cat in or on the hut). REFERENCES Allwright, R. (1984). The importance of interaction in classroom language learning. Applied Linguistics, 5 (2), 156-171. Ardt, K., Hastings, C., Hopkins, K., Knebel, R., Loh, J., & Woods, R. (2005). Report on primary education in Bangladesh: Challenges and success. Produced by the Bangladesh Bureau of Education Information and Statistics (BANBEIS) and the Bangladesh Ministry of Women and Children Affairs (MWCA), 1-23. Auger, J., & Valdman, A. (1999). Letting French students hear the diverse voices of francophony. The Modern Language Journal, 83, 403-412. Blumenfeld, P. C., Marx, R. W., Soloway, E., & Krajcik, J. (1996). Learning with peers: From small group cooperation to collaborative communities. Educational Research, 25 (8), 37-40. Blyth, C. (2003). The sociolinguistics of foreign-language classrooms: Contributions of the native, the near-native, and the non-native speaker. Boston: Heinle. Breen, M. & Candlin, C. (1980). The essentials of a communicative curriculum in language teaching. Journal of Applied Linguistics, 1 (2), 89-112. Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics, 1 (1), 1-47. Cheng, W. (1980). Communicative language teaching: Theory and practice. CUHK Education Journal, 8 (2), 59-67. Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge: MIT Press. Ciotti, M. C. (1969). A conceptual framework for small-group instruction in high school. Foreign Language Annals, 3 (1), 75-89. Ellis, R. (1985). Teacher-pupil interaction in second language development. In S. Gass and C. Madden (Eds.), Input in second language acquisition (pp. 69-85). Rowley, MA: Newbury House. 76 Ellis, R. (2000). Task-based research and language pedagogy. Language Teaching Research, 4 (3), 193-220. Gass, S. M., & Varonis, E. M. (1985). Task variation and nonnative/nonnative negotiation of meaning. In S. Gass and C. Madden (Eds.), Input in second language acquisition (pp. 149-161). Rowley, MA: Newbury House. Hamid, M. O. (2010). Fieldwork for language education research in rural Bangladesh: Ethical issues and dilemmas. International Journal of Research & Method in Educati