Medium-of-Instruction Policies and Practices in Late Colonial Hong Kong

Like their counterparts during the era of large-scale decolonization in the 1950s and 1960s, policymakers in the SAR inherited an education system in which the colonial language predominated at its crucial higher levels and a society in which proficiency in the language was a prerequisite for employment in the upper echelons of the public sector and the business and professional worlds (Bolton, Reference Bolton2011; Lin, Reference Lin, Lin and Martin2005). The ascendancy of English in higher education, professional training, and white-collar employment during the colonial era (Lin, Reference Lin1996) lies behind many of the MOI-related issues that have been the foci of study and debate in recent decades; and since the language's gatekeeping role in these domains has been reinforced rather than reduced since the handover (Evans, Reference Evans2010a), these questions continue to preoccupy policymakers and researchers.

It should be emphasized that the use of English as the principal MOI in Hong Kong higher education is largely uncontroversial as policy (Choi, Reference Choi2010), though often problematic in practice (Evans & Green, Reference Evans and Green2007). As in other post-colonial polities, the significance of the MOI at tertiary level lies in its often baleful influence on policies and practices at secondary level: if competence in English is a key determinant of admission to higher education and thereafter access to professional employment, schools will inevitably be pressured by parents to teach in English since it is taken as axiomatic that high levels of proficiency in the second language (L2) are more likely to result from its use as the MOI for content subjects than from its study solely as a language subject (Li, Reference Li and Kirkpatrick2002). In the last two decades of colonial rule, parental pressure together with the government's politically expedient policy of laissez-faire encouraged most secondary schools to adopt English as their official MOI for the entire seven-year course of studies (years 7–13). This was despite the fact, of which the colonial administration was fully cognizant (Evans, Reference Evans and Morrison2011a), that most students possessed neither the language proficiency nor the academic aptitude required to study effectively in what was at this time a rather remote L2 (Education Commission, 1990). Teachers’ solution to this predicament—having to provide a form of education that parents wanted but their offspring generally could not manage—was entirely understandable: to use the shared mother tongue to present and explicate the content of English-language teaching materials and to humanize what would otherwise have been an unpalatably formal atmosphere in the classroom (Johnson & Lee, Reference Johnson, Lee, Lord and Cheng1987). Students were still required to complete assignments and take examinations in English, an obligation that inevitably contributed to the much-deprecated practice of rote learning (Llewellyn, Reference Llewellyn1982), but this was a price they were prepared to pay for the indispensable, though hardly exclusive, label of English-medium graduate.

The transformation of the once-elite English-medium secondary stream and the concomitant stagnation of its Chinese-medium counterpart in the late colonial period stimulated considerable scholarly interest in issues relating to learning processes and outcomes in nominally English-medium classrooms. Given the manifold problems uncovered by this research, this transformation raised important questions about the relative merits of English and Chinese as MOIs, the optimum point at which English-medium instruction should be introduced, and, especially contentious, whether access to such teaching should be restricted to those judged to possess the requisite proficiency and ability. As we shall see, these issues have occupied a prominent place on research and policy agendas since the handover and will continue to do so as long as English remains the main MOI at the apex of the education system.

Much less prominent on these agendas has been the issue of the MOI at the base of the system, which is presumably because the long-standing use of the mother tongue at primary level is wholly uncontroversial as policy and largely unproblematic in practice. Although the colonial regime is often criticized for neglecting mother-tongue education (Lin & Man, Reference Lin and Man2009), what tends to be overlooked is its success in promoting Chinese-medium teaching at primary level. In Hong Kong, such instruction has generally entailed the use of Cantonese and written Chinese with traditional rather than simplified characters. The provision of mother-tongue primary education is notable because Hong Kong is the only location in East Asia (including China) and Southeast Asia where a local rather than a national language functions as the MOI (Kirkpatrick, Reference Kirkpatrick2011). It is also worth noting that Hong Kong has traditionally placed greater emphasis on mother-tongue teaching than many other post-colonial societies. For example, schools in most former British colonies in Africa switch to English-medium instruction at upper primary level after using an indigenous language (though not necessarily the mother tongue) in the first three or four years. In some cases (e.g., Zambia), teaching in English commences in the first year (Ferguson, Reference Ferguson2006). Indeed, even after the recent fine-tuning of the mother-tongue policy, Hong Kong's education system arguably puts greater stress on mother-tongue education than that of Tanzania, which is often upheld as a model of enlightened language policymaking (Evans, Reference Evans2011a).

Historical Perspectives on the Medium of Instruction in Hong Kong

A conspicuous lacuna in the literature inspired by the rise of English-medium schooling in late colonial Hong Kong was a historical perspective on issues relating to MOI policy and practice. The need for such a perspective was highlighted in the final year of colonial rule by the educational historian Anthony Sweeting (Reference Sweeting and Bray1997), who criticized applied linguists in Hong Kong for their “ahistorical, a priori approach” (p. 35) to the study of language questions. One consequence of the “vanishing sense of history” (p. 36) in the local applied linguistics community was that potentially illuminating insights into contemporary issues were ignored or rejected, the implication being that if policymakers had capitalized on the knowledge acquired from more than a century of English-medium teaching, some of the problems in the area of language in education might have been alleviated or perhaps even averted.

While a sense of history has yet to materialize—the antecedents of current controversies continuing to go unnoticed in the literature—there has nevertheless been a modest stirring of interest in the historical dimension of language policy in the past decade, although as Bolton (Reference Bolton2011) points out, a detailed narrative history remains to be written, with the period between the late 1880s and the late 1970s virtually uncharted.Footnote ² Sweeting himself—in collaboration with a fellow historian (Sweeting & Vickers, Reference Sweeting and Vickers2007)—ventured into the field of applied linguistics with a penetrating critique of what they saw as Pennycook's (Reference Pennycook1998) highly generalized account of the nature and purposes of colonial language policy in Hong Kong, lamenting his excessive reliance on secondary sources and selective use of conveniently accessible primary sources. These limitations, in Sweeting and Vickers's (Reference Sweeting and Vickers2007) view, pointed to the need for the history of colonial education—itself the “dark continent of imperial historiography” (p. 1)—to be reconstructed from the bottom up through tightly focused archival research into policy, schools, and curricula in particular contexts. Coincidentally, several article-length studies of this nature have appeared since Sweeting and Vickers issued this call. These include reconstructions from primary sources such as Colonial Office correspondence and contemporary newspapers of the introduction of English teaching on Hong Kong Island in the 1840s and 1850s (Evans, Reference Evans2008a), the establishment of the government's flagship English-medium school in the 1860s (Evans, Reference Evans2008b), and the formulation of MOI policy in the 1870s and 1880s (Evans, Reference Evans2008c). These articles, together with recent studies linking past and present at secondary (Evans, Reference Evans2011a) and tertiary levels (Lin & Man, Reference Lin, Man and Davis2011), reveal that issues that have preoccupied policymakers since the late 1970s, notably the apparently deleterious influence of mixed-mode instruction on students’ language development, have been a source of concern for more than a century.

Whether this emerging strand of research will be extended in the years ahead will depend on the priorities of funding bodies and the predilections of individual investigators. As we shall see, an already crowded research agenda has expanded since the handover with the inclusion of questions concerning the teaching of Putonghua and the fine-tuning of MOI policy, and thus researchers will rightly wish to concentrate on understanding the present rather than on unearthing the past. However, as Sweeting (Reference Sweeting and Bray1997) noted, the past contains valuable lessons for the present. As policymakers press ahead with the fine-tuning initiative, which will pose essentially the same questions as those that confronted their counterparts in both the late 19th and late 20th centuries, they would do well to heed them.

Promoting Biliteracy and Trilingualism

The unique one country, two systems principle established by China and Britain in 1984 permits Hong Kong a high degree of autonomy in managing its affairs, including those relating to education. The SAR's status is embodied in its constitution, the Basic Law, which provides no evidence of the Chinese government's wish to replace the colonial education system or to introduce the teaching of Putonghua, which for more than half a century has been the cornerstone of an unwavering, all-embracing language policy on the mainland: “On the basis of the previous educational system, the Government of the Hong Kong Special Administrative Region shall, on its own, formulate policies on the development and improvement of education, including policies regarding the educational system and its administration, the language of instruction. . .” (HKSAR Government, 2012, p. 71). While the Basic Law stipulates that Chinese is the primary official language of the SAR (English being relegated to a secondary official role), the conveniently ambiguous term Chinese is left undefined, as indeed it was when Chinese was accorded co-official status in 1974. As in the late colonial period, Chinese is still interpreted in Hong Kong as Cantonese and standard written Chinese (with traditional characters) when characterizing language policy in the domains of education and government.

Although policymakers were not constitutionally obliged to introduce the teaching of Putonghua, Hong Kong's political and economic reintegration with renascent China since the early 1980s meant that the decision to promote the national language in the education system was inevitable and desirable. The plan to promote biliteracy and trilingualism was formally announced in the first policy address of Hong Kong's Chief Executive, C. H. Tung, although like the mother-tongue policy, it was instigated in the final years of British rule (Education Commission, 1996). For much of the colonial era, Putonghua played a largely peripheral role in the school system, being learned either as an extracurricular activity or—as 1997 loomed—as an optional subject on the timetable (Zhang & Yang, Reference Zhang, Yang and Zhou2004). However, it was only after the handover that Putonghua became a core component of the primary and secondary curricula (Davison & Auyeung Lai, Reference Davison and Auyeung Lai2007), initially as an independent language subject and more recently as the MOI for Chinese subjects. The rationale for using Putonghua rather than Cantonese, the traditional MOI for such instruction, is that the national language, unlike the local language,Footnote ³ corresponds closely to standard written Chinese (Li, Reference Li2006; Snow, Reference Snow2004) and is therefore deemed to be a more effective means of enhancing the acquisition of literacy in Chinese (SCOLAR, 2003). Written Cantonese, though often used in the print media, is regarded in Hong Kong as being nonstandard (Li, Reference Li2006) and thus forms no part of the administration's linguistic vision.

Patterns of Language Acquisition and Use

The term vision seems appropriate here, for while the government's choice of written and spoken codes is difficult to dispute given the SAR's reunification with the mainland (Putonghua), its status as a quasi-autonomous city-state (Cantonese), and its entanglement in the tightening web of economic globalization (English), the promotion of biliteracy and trilingualism appears to be a general aspiration rather than a precisely defined policy. One source of uncertainty concerns policymakers’ expectations regarding the levels of proficiency that school graduates and the wider populace should possess in each code. In an important review article, Li (Reference Li2009) questions whether their goal is “balanced” or “functional” biliteracy and trilingualism, assuming it to be the “more realistic” latter category, which is defined as “the ability to use the three languages to varying degrees of proficiency and for different purposes” (p. 82). The Chief Executive's early policy pronouncements, however, suggest that the former may have been in view. In his seminal 1997 address, Mr. Tung stated that the administration's goal was for graduates to be “proficient in writing English and Chinese and able to communicate confidently in Cantonese, English, and Putonghua” (HKSAR Government, 1997, p. 30), while two years later he declared that the objective was “to train our people to be truly biliterate and trilingual” (HKSAR Government, 1999, p. 23). In contrast, the government's advisory body on language education sensibly observed that it was “impractical and unrealistic to expect every member of our population to attain a high level of proficiency in both Chinese and English” (SCOLAR, 2003, p. 6), the use of the term Chinese unhelpfully masking the different circumstances which surround Hongkongers’ acquisition and use of Cantonese, Putonghua, and written Chinese.

In his analysis of the policy, Li (Reference Li2009) highlighted two “unfavourable acquisitional factors” (p. 74) that apparently impede the progress of English and Putonghua in the SAR (see also Li, Reference Li and Feng2011). The first is the contention that “for the majority of Hongkongers, English has very little reality outside school premises or in their lifeworld” (Li, Reference Li2009, p. 74). This assertion is presumably based on the fact that Cantonese is the first language of around 90% of the population (Census & Statistics Department, 2012) and is therefore the usual medium of spoken communication among Hongkongers, a major marker of their identity (Kuah-Pearce & Fong, Reference Kuah-Pearce and Fong2010; Tsui, Reference Tsui, Tsui and Tollefson2007), and the language towards which they have the strongest integrative orientation (Lai, Reference Lai2007, Reference Lai2010, Reference Lai2011). Since the handover, Cantonese has penetrated the high domains that for much of the colonial era were reserved for English: it is now the usual language of communication in the civil service (Evans, Reference Evans2010a) and the legislature (Bolton, Reference Bolton2011), and—notwithstanding the increasing prominence of Putonghua—has retained its centrality in white-collar (and presumably blue-collar) employment (Evans, Reference Evans2010b).

When considering the role of English in the SAR, and particularly Li's (Reference Li2009) daring claim about its insignificance in most Hongkongers’ lifeworld, it is necessary to distinguish between the spoken and written language and to remember that a significant segment of this world involves (for adults) workplace communication and (for adolescents and adults) socially inspired communication via computers and smartphones. A recent study by Evans (Reference Evans2010a, Reference Evans2010b, Reference Evans2011b) found that business people and civil servants spend a substantial proportion of their working lives reading and writing English documents and participating in speech events whose subjects and outcomes are invariably English texts of various kinds, even if—as is often the case—such interactions are conducted in Cantonese. The importance of reading and writing in English (particularly online) was also revealed by a large-scale survey of young Hongkongers’ use of the language in their leisure time (Evans, Reference Evans2011c). While the findings confirmed Li's (Reference Li2009) perception that teenagers and young adults have little need to speak English in their everyday lives, the survey nevertheless indicated that they are exposed to a great deal of spoken English as a consequence of their watching films and television and listening to music. With regard to Putonghua use, we have little empirical evidence to test Li's plausible claim that young Hongkongers have “hardly any opportunities for meaningful practice beyond school premises” (p. 76; although see Ho, Reference Ho2008). The workplace study referred to above found that, unlike spoken English and Cantonese, Putonghua currently plays a limited role in business communication (Evans, Reference Evans2010a) and was therefore perceived by the 2,030 survey participants to be the least important of the five codes for business purposes, with written English being the most important, closely followed by Cantonese and spoken English, and written Chinese a distant fourth (Evans, Reference Evans2010b).

The second set of factors that Li (Reference Li2009) believed inhibit the acquisition of English and Putonghua include the typological distance that separates English and Chinese (the former being an Indo-European language, the latter a Sino-Tibetan language) and the linguistic differences between Cantonese and Putonghua, which lie primarily in their phonological systems. As Kirkpatrick and Chau (Reference Kirkpatrick and Chau2008) argue, the typological gulf between English and Chinese weakens the claims of English to be the target language in an L2 immersion program in the Hong Kong context, Putonghua being (in their view) a more promising candidate owing to its close correspondence with written Chinese. This connection, as noted earlier, lies behind the initiative to teach Chinese language and literature through the medium of Putonghua. At present, this initiative does not encompass every primary and secondary school; nor does it extend to content subjects, which in the majority of schools have been taught mainly in Cantonese since the handover, as the next section reveals.

Medium-of-Instruction Policy at Junior Secondary Level (Years 7–9)

The mother-tongue MOI policy represented a fundamental shift in Hong Kong's educational landscape because in the final year of colonial rule more than 90% of secondary students were attending nominally English-medium schools (Evans, Reference Evans2009). Under the policy, which was implemented in September 1998, most of these schools were compelled to switch to Chinese in years 7–9, although they were permitted to shift to English in years 10–13 since this examination-dominated stage fell beyond the nine years of free compulsory education provided by the government. Only 114 of Hong Kong's 421 secondary schools were exempted from the compulsory mother-tongue policy, having evidently satisfied the authorities that their staff and students were capable of teaching and learning effectively in English (Tse, Shum, Ki, & Wong, Reference Tse, Shum, Ki and Wong2001). As expected, the language-based bifurcation of schools proved to be highly controversial (Tsui, Reference Tsui, Tollefson and Tsui2004): many parents and students justifiably complained that the retention of an elite English stream containing the most academically able students was divisive and discriminatory (Choi, Reference Choi2003). Given the centrality of English in higher education and professional employment, students assigned to English-medium schools were perceived to enjoy an unfair advantage in life, while those forced to attend Chinese-medium schools were denied access to valuable linguistic capital and therefore the prospect of educational and occupational advancement. As we shall see, the fine-tuning initiative, which abolishes the MOI-based labeling of schools, seeks to redress this wrong.

Although the policy emanated from plans conceived in the late 1980s (Education Commission, 1990), the issuance of what was called firm guidance to schools on their MOI soon after the handover (Education Department, 1997) was widely regarded as an abrupt and unwelcome departure from the long-standing practice of granting schools autonomy in MOI policymaking and as a political move arising from China's resumption of sovereignty (Poon, Reference Poon2010). If the mother-tongue policy was indeed politically motivated, it could be interpreted as a bold (even mildly subversive) maneuver by the new administration to safeguard Hong Kong's distinctive Cantonese-based culture and identity rather than as a ploy to appease the potentially intrusive Beijing regime, which, if unencumbered by the one country, two systems framework, would presumably have sought to impose its one-country Putonghua policy. The evidence in fact suggests that the policy was motivated primarily by educational considerations, and particularly by the hardly novel notion that learning is more effective when undertaken in the mother tongue. The policy's origins and rationale are encapsulated in the motion debated and passed by the Legislative Council (1997) two months before the handover, urging the government “to expeditiously implement fully the policy of mother-tongue teaching which has already been put in place for nearly a decade so that secondary school students will be able to learn more effectively in their everyday language” (pp. 115–16). The mother-tongue policy is the SAR government's major language-in-education initiative and has therefore been the focus of considerable research in the past decade, much of it involving some form of comparison between English- and Chinese-medium instruction. This research falls into three broad categories: policy implementation, teaching-learning processes, and learning outcomes.

Research into the Implementation of the Medium-of-Instruction Policy

One strand of research has explored aspects of the implementation of the mother-tongue policy, generally by eliciting via surveys and interviews the attitudes, views, and experiences of administrators, teachers, and students. An early study by Tse et al. (Reference Tse, Shum, Ki and Wong2001) gauged the views of teachers and administrators on the issues that flowed from the transition from English- to Chinese-medium instruction. Teachers evinced a generally positive attitude towards mother-tongue teaching, but were unprepared for the policy change in 1998, despite its long period of gestation, and lacked confidence in their written Chinese, presumably due to its neglect during their English-medium secondary and tertiary studies. The investigation also highlighted a major obstacle to the promotion of mother-tongue education: parents’ apparently greater concern for their children's career prospects than for their learning processes (see also Poon, Reference Poon, Ng and Renshaw2009b). A subsequent study by Tse, Shum, Ki, and Chan (Reference Tse, Shum, Ki and Chan2007), which also incorporated English-medium teachers’ perspectives, highlighted another critical factor in successful policy implementation: the importance of considering the views of professionals in the classroom. The teachers surveyed by Tse, Shum, et al. (Reference Tse, Lam, Loh and Lam2007) generally felt that the administration's centrally directed policy was deficient in this regard and instead recommended a return to school-based MOI decision making, which (presumably coincidentally) the fine-tuning initiative has in fact effected.

One likely consequence of the devolution of decision making is the reappearance of mixed-mode instruction in nominally English-medium classrooms. The eradication of such instruction was a central aim of the two-stream MOI policy: henceforth, teachers would be required to make consistent use of either Chinese (i.e., Cantonese and written Chinese) or English (Education Commission, 1990). This requirement was directed particularly at teachers in the 114 English-medium schools, who in many cases would have had to radically alter their approach to comply with the English-only directive (Evans, Reference Evans2002). How faithfully they have observed this instruction is difficult to determine because direct methods of gathering this information such as classroom observations and recordings may not have elicited especially representative data given the Education Bureau's monitoring of language use and its threat to revoke a school's English-medium status if its teachers were found to be mixing English and Cantonese (Evans, Reference Evans2009).

The only study that has attempted to examine this aspect of the policy's implementation, albeit on a small scale and using indirect methods, found that English-medium teachers had made a determined effort to adhere to the English-only directive (Evans, Reference Evans2008d, Reference Evans2009). While the alignment of policy and practice was not watertight, it was as close as might reasonably be expected given the predominance of Cantonese and the prevalence of code mixing in Hong Kong society (Low & Lu, Reference Low and Lu2006). This evidence, together with that derived from studies of teaching-learning processes, suggests that the years between the introduction of the policy (1998) and the institution of fine-tuning (2010) may be the high-water mark in the history of English-medium schooling in Hong Kong, a period when around a quarter of local students received a bona fide L2 immersion education. This apparently stood in marked contrast to the experience of students in Chinese-medium schools who, for particular subjects or classes, switched officially to English-medium but actually to mixed-mode instruction at senior secondary level (Evans, Reference Evans2009).

Research into Teaching-Learning Processes

A second strand of research has examined teaching-learning processes in Chinese- and English-medium classrooms using direct methods of data collection complemented by interviews and surveys. The lessons analyzed by Ng (Reference Ng2007) apparently took place before 1998, since they occurred in a category of school (dual stream) that was abolished by the mother-tongue policy. Nevertheless, the research design employed by Ng (Reference Ng2007), which capitalized on the school's policy requiring teachers to give the same lessons in both languages and utilized observations, recordings, and interviews, enabled the direct comparison of the quality of teaching and learning in Chinese- and English-medium classes (see also Ng, Tsui, & Marton, Reference Ng, Tsui, Marton, Watkins and Biggs2001). Ng's findings are valuable because the fine-tuning initiative will presumably prompt some former Chinese-medium schools to offer certain subjects in both English and Chinese. Ng found that the use of the L2 hindered the smooth progression of lessons: when they used English, teachers had to devote a disproportionate amount of time to providing linguistic support (especially lexical), but when they taught in Cantonese, they were able to focus on the subject matter and inspire interaction and engagement. In contrast, classes taught in English featured recitation and memorization activities and a narrow focus on the component parts of concepts rather than on the whole and the relationship between parts and whole. Ng thus concluded that the MOI plays a “critical role in shaping a student's learning experiences” (2007, p. 168).

Unlike Ng (Reference Ng2007), Yip, Coyle, and Tsang (Reference Yip and Tsang2007) compared teaching-learning processes in the post-1998 Chinese- and English-medium streams, focusing on the effects of the MOI on teaching styles and modes of instruction in science lessons. These effects were determined by the perceptions of 17,616 students in a questionnaire survey encompassing 100 schools and observations of 10 lessons (five in each type of school). The survey results indicated that lessons in Chinese-medium schools were more interactive and creative than those in English-medium schools. These perceptions were largely corroborated by classroom observations, which, though confirming Hong Kong teachers’ preference for a didactic approach, nevertheless indicated that students taught in Cantonese were given more opportunities to participate in class discussions, answer cognitively demanding questions, and engage in group work. In contrast, students taught in English spent most of their time listening to the teacher lecturing (not always coherently) or working individually. When questioned by the teacher, they responded in a single word or phrase; when asked to work in groups, they communicated in Cantonese.

As noted earlier in this article, the mother-tongue policy applied only to years 7–9, and thus Chinese-medium schools were permitted to switch to English at senior secondary level. The impact of this transition on classroom interaction and language use was examined by Lo and Macaro (Reference Lo and Macaro2011), who also, for the purpose of comparison, studied year 9 and 10 classrooms in English-medium schools, where it was anticipated (and subsequently confirmed) that the transition would be seamless since the MOI was unchanged. The findings were derived from recordings of 60 lessons together with retrospective interviews with the 18 teachers involved. These indicated that the shift in the MOI was manifested mainly in the written medium. In terms of the spoken medium, however, only one school effected a significant change in language use, from almost 100% Cantonese in year 9 to around 80% English in year 10. In the other Chinese-medium schools there was considerable variation in the amount of English used, which corroborates Evans's (Reference Evans2009) finding about the prevalence of mixed-mode teaching in the senior years. Lo and Macaro (Reference Lo and Macaro2011) also discovered that lessons conducted in English in the MOI-switching schools were more teacher-centered (and therefore less interactive) than those taught in Cantonese, which accords with the findings of Ng (Reference Ng2007) and Yip et al. (Reference Yip, Coyle and Tsang2007).

Research into Learning Outcomes

A third strand of research has investigated the impact of the MOI on learning outcomes, thus extending a line of inquiry that commenced in the 1970s (for an overview, see Marsh, Hau, & Kong, Reference Marsh, Hau and Kong2000) and played a central role in the formulation and justification of the mother-tongue policy in the 1990s (Choi, Reference Choi2003). One of the difficulties of conducting comparative studies during this period was the prevalence of mixed-mode instruction in the unreformed English-medium schools and the virtual absence of schools in which Chinese was the official MOI. This problem largely disappeared between 1998 and 2010, when schools evidently fulfilled their government-imposed MOI obligations (at least in years 7–9). Despite this, measuring the impact of the teaching medium on student learning remains a complex task, principally because the most academically able students are allocated to English-medium schools, but also because student performance is influenced by a multiplicity of variables, which even the most ingenious investigator may struggle to control. Notwithstanding these challenges, several comparative studies have been undertaken in recent years, focusing either on outcomes in content subjects or in L2 acquisition.

A study by Yip, Tsang, and Cheung (Reference Yip, Tsang and Cheung2003) examined the impact of the MOI on achievement in science through a test comprising multiple-choice questions and more cognitively demanding free response questions, which was administered to year 8 students in 100 schools. Recognizing the disparity in student intakes between the two types of school, they categorized Chinese-medium schools into high, medium, and low on the basis of student ability, the high schools being somewhat comparable to English-medium schools. They discovered that students from the top Chinese-medium schools demonstrated higher achievement than students in the English-medium stream. The English-medium students outperformed the high Chinese-medium students on only a small number of multiple-choice items, which Yip et al. (Reference Yip, Tsang and Cheung2003, p. 324) claimed made “relatively low cognitive demands” on the test takers. Their overall conclusion was that students in English-medium schools lacked the requisite L2 proficiency to learn science effectively and were thus at a disadvantage compared to students in the Chinese-medium stream.

Yip et al.'s (Reference Yip, Tsang and Cheung2003) conceptualization of the outcomes of science learning comprises two components: academic achievement and a student's self-concept in the subject, which is explored in another article from their project (Yip & Tsang, Reference Ng2007). This component was determined by the responses of the aforementioned test takers to a questionnaire. These responses indicated that students taught in English had a lower self-concept in science than their Chinese-medium counterparts, despite being “more academically oriented” and were “recognised socially as an elite class among their peers” (Yip & Tsang, Reference Ng2007, p. 409). Yip and Tsang attributed this to the challenges associated with studying a difficult subject in an L2. The challenges posed by English-medium instruction were also highlighted in a study by Lau and Yuen (Reference Lau and Yuen2011) into the effect of the MOI on the learning of computer programming. This was measured by the performance of 219 year 10 and 11 students from nine schools on a 25-minute programming performance test. Lau and Yuen found that Chinese-medium students generally outperformed their English-medium counterparts, with low- and middle-ability students in English-medium schools being “notably at-risk” when studying computer programming in an L2 (2011, p. 194).

The benefits of mother-tongue teaching revealed by professors in ivory towers are unlikely to deflect parents in concrete towers from their belief in the advantages of English-medium instruction. Hong Kong parents do not require empirical evidence to know that learning in the mother tongue is generally more effective and that studying in English represents a considerable challenge. Their long-standing preference for English-medium teaching rests on the conviction that using English as the MOI is a better means of achieving high levels of proficiency in an occupationally valuable L2 than teaching English solely as a language subject. Recent research (as discussed below) suggests that their faith in English-medium instruction is not misplaced. This research examines the impact of the MOI on students’ vocabulary knowledge, a highly sensitive indicator of language proficiency and one that is especially amenable to measurement given the availability of well-established instruments such as the Vocabulary Levels Test (VLT; Laufer & Nation, Reference Laufer and Nation1999; Nation, Reference Nation1990).

These instruments were used by Lo and Murphy (Reference Lo and Murphy2010) in their study of students’ L2 vocabulary knowledge in two secondary schools (Chinese- and English-medium), which they claim were “closely matched” (p. 220) in terms of students’ academic ability, socioeconomic status, English-learning experience at primary level, and exposure to English outside school. Three aspects of the vocabulary of year 7 and 9 students were measured towards the end of each school year: passive, controlled active, and free active. In terms of passive and controlled active vocabulary, year 7 students in the English-medium school obtained “significantly higher total scores” (p. 225) on both versions of the VLT than their counterparts in the Chinese-medium school. Year 9 English-medium students possessed a passive knowledge of 3,338 word families compared to Chinese-medium students’ 2,170, and a controlled active vocabulary of 2,027 families compared to Chinese-medium students’ 984. Regarding free active vocabulary, which was determined by lexical frequency profiling, Lo and Murphy found that students in the English-medium school “consistently demonstrated superior vocabulary knowledge” in their compositions (p. 228). While caution needs to be exercised over these findings (it is, of course, difficult to form comparable samples and to control the manifold non-MOI-related variables that influence L2 acquisition), they are not especially surprising, for, as Lo and Murphy pointed out, English-medium students are exposed to a wider variety and greater quantity of English-language texts and have more opportunities to use the vocabulary they have acquired than their Chinese-medium counterparts. This gives them a distinct advantage in gaining admission to higher education and adapting to its academic demands.

This advantage was confirmed by Lin and Morrison (Reference Lin and Morrison2010), who found that first-year undergraduates from English-medium schools possessed significantly larger active and passive vocabularies than those from Chinese-medium schools, and by Evans and Morrison (Reference Evans and Morrison2011a), who discovered that English-medium students made a smoother transition to university studies than their less proficient Chinese-medium counterparts. The participants from Chinese-medium schools in these two studies formed a minority in their institution and constituted the elite of the Chinese-medium stream. Recognition of the difficulties experienced by Chinese-medium students in the crucial matter of university admission, which were highlighted in an influential study by Tsang (Reference Tsang2009), appears to have been the main factor behind the Education Bureau's decision to fine-tune the mother-tongue policy in the late 2000s.