Original Paper Mainland Non-English-Major Students’ Perceptions of English Academic Writing in the Taught Postgraduate Program in Hong Kong: A Needs Analysis

As higher education is being internationalized globally, it is also not rare to find degree programs delivered in English, the world’s lingua franca, in countries where English is learned as a foreign language. In mainland China, such courses are available for bachelor and master’s degrees. Accordingly, students in those programs have to meet the academic English requirements, by which writing is assumed to be the most challenging. This small-scale research was conducted among 81 mainland non-English-major students studying in the taught postgraduate program in Hong Kong, with the instruments of questionnaires and follow-up interviews. Within the framework of needs analysis, it reports their detailed perceptions of English academic writing. Results indicate that those upper-intermediate language learners are generally able to get accustomed to academic writing in English, but some writing skills, and particularly language issues (academic lexis, grammar, and style) pose challenges to their studies. The article concludes with some feasible pedagogical implications for updating the university English education system in mainland China.

characterized by the compression of time and geographical distance, the reduction of diversity through intensified trade and communication, and new social relationships marked by reduced local power and influence" (Giddens, 1990, p. 64;cited in Coleman, 2006, p. 1). It manifests itself in the inexorably increasing use of English, the lingua franca for world communication. English has become "the main foreign language of education, and international cooperation and competition" (Brumfit, 2004, p. 165), as well as "the language of science" (Ammon, 2001;cited in Bradford, 2013, p. 227).
Against this backdrop, Coleman (2006) comes up with the conception of "academic internationalization" (p. 5), which is directly associated with the introduction of English-Medium Instruction (EMI) in learning settings for non-English-speaking countries. That is in concord with the prediction by Graddol (1997) that "One of the most significant educational trends world-wide is the teaching of a growing number of courses in universities through the medium of English" (p. 45). EMI has experienced exponential growth in Europe where English is learned as a foreign language (e.g., Doiz et al., 2013a;Costa & Coleman, 2013). Since the 1950s, European countries (particularly the Netherlands and Scandinavia) have been attempting at internationalizing their Higher Education Institutions (HEIs) by successively administering EMI degree programs both at the undergraduate and postgraduate levels (Coleman, 2006). In doing so, those HEIs are capable of taking in more international students of diversity, and meanwhile producing domestic students who are more employable in the globalizing job market, which in return raises their standing in the world university rankings. Similarly, the recent years have witnessed the expansion of the practice of EMI in Asian HEIs where academic subjects are also taught in an EFL (English as a foreign language) context, for instance, in South Korea, Malaysia, China including Hong Kong, Taiwan and mainland. To conclude, "Globally, English is already the language of higher education" (Brumfit, 2004, p. 166).
In the mainland context, an increasingly great importance has been attached to English learning.
According to the policy of Ministry of Education of PRC (2001), universities' undergraduate programs should gradually create favorable environments to adopt English as the medium of instruction for both common and specialized courses, in order to catch up with the international tempo of academic progress after China's entry into WTO. In response to the government's call, many HEIs have pioneered in conducting academic courses through English that is learnt as a language subject in mainland. Those courses take various forms. Most of them are merely for one discipline or a few in one college (e.g., Cai, 2010). In this case, bilingual instruction (mixture of English and mandarin) is actually exercised instead of full English teaching. Some are operated as joint programs with western countries, among which are the well-known Ningbo Nottingham University, and Xi'an Jiaotong-Liverpool University. The newly established South University of Science and Technology of China (SUSTC) claimed to be the first university to deliver all the academic subjects exclusively in English to both the undergraduate and postgraduate groups. Admittedly, hitherto EMI is still on the periphery of the mainland's tertiary education system. One consequence is that the existing literature mostly qualitatively describes one or a few aspects of EMI settings with small-scale sampling (e.g., www.scholink.org/ojs/index.php/selt Studies in English Language Teaching Vol. 6, No. 1, 2018 23 Published by SCHOLINK INC. Ling et al., 2014). In-depth reflections and analyses of EMI practice are hardly available, which indicates the great needs for more research into this field.

Rationale for the Current Research
There is utility in looking to the better developed EMI programs for some informed research foundations. A very similar picture has been reported from the classrooms in districts such as Europe, Australia (EMI degree courses for international students), Japan, and Hong Kong: linguistic challenges apparently and immediately trouble the faculty and students most, an aspect that is worth far more subsequent investigations (e.g., Tatzl, 2011;Phakiti & Li, 2011;Evans & Morrison, 2011).
Unlike English for general purposes, the language in EMI programs is termed as English for Academic Purposes (EAP) which refers to "language research and instruction that focuses on the specific communicative needs and practices of particular groups in academic contexts" (Hyland & Hamp-Lyons, 2002, p. 2). Furthermore, the productive skills (academic speaking and writing) have long been the center of students' linguistic problems (e.g., Evans & Green, 2007). Since the limitations of the project's time and resources prevent the investigation from covering too broad the research topic, academic writing, which has been a prominent issue for tertiary students, is selected as the only area of the current research to probe into.

Theoretical Framework of the Research: Needs Analysis
Needs analysis is "a systematic way of identifying the specific sets of skills, texts, linguistic forms and communicative practices that a particular group of learners must acquire" (Dudley- Evans & St. John, 1998;cited in Hyland & Hamp-Lyons, 2002, p. 5).
As this technique "can tell us a lot about the nature and content of the learners' target language needs" (Hutchinson, 1988, p. 71), it has often been regarded as the critical initial step when developing appropriate learning syllabus and course materials (e.g., Hyland, 2006). Hutchinson and Waters (1987) put forward the theory of target needs, which could be broken down into three components: necessities (what the learners are required of by the target situation's standards), lacks (the gaps between what the target situation expects and what the learners already know) and wants (what the learners personally wish to learn). That will serve as the theoretical framework of this research.
In terms of analyzing the needs of EFL university students for academic literacy acquisition, a look at how the related research has gone so far in Hong Kong is rather significant before it comes to the situation in mainland, since Chinese is the mother tongue in both districts though it is spoken Cantonese and written standard Chinese in HK while spoken mandarin and written simplified Chinese in mainland. Chinese (L1) and English (L2) are typologically much more distant than other first languages in the EMI programs (Genesee, 2006), which means many issues of transfer from L1 in HK classes can provide implications for mainland settings. In addition, HK has a much longer history of EMI higher education and it has experienced the shift from "elite" to "mass" tertiary schooling, which have yielded more empirical research outcomes. Over the past two decades, there have been three large-scale studies of HK undergraduates' academic language needs: Hyland's (1997), Littlewood and Liu's (1996), and Evans and Green's (2007). Their informing results have provided invaluable insights into HK's local undergraduates' experiences of academic literacy when studying in a second language.
Hyland (1997) conducted a survey among 1619 undergraduates from 8 majors in 5 HK universities.
The participants admitted the necessity of academic English courses to support their content studies.
Their language-related problems were chiefly from the academic writing and speaking, as well as the accumulation of specialist vocabulary. However, this broad-stroke research only identified the students' perceived difficult macro skills of English, i.e., writing and speaking, instead of the micro level (cohesion, grammar, academic vocabulary, etc.). Thus the nature of those problems was not specified indeed.
Around one decade later, Littlewood and Liu (1996) published the results of another large-scale survey among 2156 freshmen from 4 tertiary institutes. This research mainly concerned the students' language experiences and attitudes when learning in an English environment. As improvement on Hyland's (1997) research method, this study even added objective tests to confirm students' subjective perceptions. As a result, academic writing was again assumed to be the principal difficulty for those participants. This time the questionnaire contained sub-skills of academic writing, but 7 items seemed insufficient to exactly reveal the overall nature of those challenges. Also, as pointed out by Evans and Green (2007), even the 7 items were inappropriately selected, for example, the skill of "use idioms or colloquial expressions correctly" (a skill not necessary for academic writing) was included while the important abilities to summarize or synthesize were excluded.
One of the largest investigations of this type ever undertaken so far was Evans and Green's (2007), which surveyed approximately 5000 Cantonese-speaking undergraduates from 26 departments of Hong Kong Polytechnic University. The questionnaire adopted covered detailed and proper sub-skills of academic listening, reading, writing and speaking. The major challenges were academic writing (style, grammar and cohesion in particular) and speaking. The survey's findings, accompanied by interviews with related parties, generated very rich data to identify the linguistic problems of HK's local undergraduates. But like the former two studies, the subjects in Evans and Green's (2007) are not the same cohort in the current research who are taught postgraduate students from mainland, which means their experiences could be in disparity.
Targeting at the students in the master's programs, Cai's (2013) small-scale survey reported in-depth outcomes of 50 English majors' needs for academic English support based in a university of South China. English academic writing was the only topic in Cai's (2003) survey, which is in consistence with the current one. The data derived from the questionnaires and interviews had the indication that the subjects expected some explicit instruction on generic features of writing the sections in academic articles and more importantly, the language resources for academic writing.
Since the subjects in the current research are mainland non-English-major postgraduate students studying in HK, their English proficiency and language exposure are differentiated from those English majors in Cai's (2013) survey. The current participants are anticipated to meet more difficulties with English academic writing, which makes them deserve more attention from the research field. Besides, considering the ultimate goal of this study is to make contributions to the research on mainland's EMI programs, those non-English-major subjects could be more representative of the majority of the students in the EMI settings. Hence, the present research has solid grounds for being carried out. It is mainly intended to describe the linguistic experiences (specifically the academic writing) of the non-English-major students in the taught postgraduate program, thus to specify the corresponding language requirements for studying through a second language and what their undergraduate studies can have prepared them to quicken their accommodation process for the academic career. It is hoped that the findings can give some insights for perfecting the mainland college English education system, and also show dedication to the research of the academic English field.
This research was aimed to address the following questions: a) What are students' attitudes towards English academic writing when studying in an EMI taught postgraduate program? b) What are students' perceived difficulty degrees of the sub-skills of English academic writing? c) What academic writing skills do the students wish to have been taught over the undergraduate studies?

Subjects
This research was based in the University of Hong Kong (HKU) with some considerations. For one thing, it is a well-established comprehensive university with a history of over 100 years. All the courses (except some for particular languages) are delivered in English once the students enter university, and the regular assignments, laboratory reports and theses are also written in English, which means it has rich experience in attending to nonnative students' language issues met with in the EMI mode. For another, it is also where the author is currently doing her master's degree in TESOL, so that she knows the research context well and the inconvenience during data collection could be minimalized. 81 students enrolled in the 2015 HKU's taught master's program participated in the research with willingness. They were selected as the sample mainly for the reasons: 1) They are almost finishing their degree studies, which means they are well aware of the academic writing requirements from an EMI taught master's program; 2) They all are from the mainland, who can truthfully tell the previous study experiences and offer constructive suggestions to the future students of the same cohort (important components of the research); 3) Since they were admitted to the program with the language requirement (i.e., IELTS/TOEFL scores), they could highly represent a focus group of makes the data collected reflect a comprehensive rather than discipline-specific picture. And the subjects obtained their bachelor's degrees from various universities in mainland China, which could provide richer information about their former English academic studies.

Questionnaire Design
As the major instrument for this investigation, the questionnaire was designed within the theoretical framework of needs analysis by Hutchinson and Waters (1987). A total 10 items could be divided into three sections (see Appendix for the questionnaire).
Section I (Q.1-6) (necessities) centered on the English academic writing genres the students needed to manage working on during the studies and their motivation to learn this academic skill. According to Cai's (2013) suggestion, an extra item concerning students' intrinsic interest in English academic writing was added for its significant role in second language acquisition, which could have something to do with the findings.
Section II (Q.7-8) (lacks) was aimed at revealing students' perceived difficulty degrees of the 13 sub-skills of English academic writing, which were all indicated on a 5-point scale (1=very easy while 5=very difficult). Those sub-skills fall into two categories: general academic writing skills and language issues. The items originated from the questionnaires in Evans and Green (2007) and Cai (2013). Yet some adaptations were made to them (See Table 1 Cai's (2013). Also, the fine-grained categorization of English academic writing skills in their questionnaires seems slightly too professional for non-English majors. Thus some reduction in the number could prevent the information from being too overwhelming for the questionnaire takers in an estimated 5 minutes' self-report time.
b) Adjustment of the content of the items. Given the fact that subjects in Evans and Green (2007) were undergraduates, they might not write very formal research papers which should include such sections as methodologies, data collection and analysis, and so forth. On the other hand, Cai (2013) looked into the situation of English majors in the master's program, who might not do experiments as those engineering or science majors. However, the participants were those postgraduates who had to write formal these for graduation, and also came from various disciplines. Therefore, some modifications could be identified when compared with the items in the other two questionnaires.

c)
Polish of the expressions of each item. Considering that those subjects did not specialize in English language, some use of meta-language should be avoided in order not to confuse them.
Apart from the modifications to the questionnaire, the language adopted was even converted into Chinese to make it user-friendly after a pilot study in which the subjects strongly recommended doing that to guarantee the reliability and validity of the research data.  Section III (Q.9-10) (wants) concerned academic writing skills the students wished to have been taught over the undergraduate studies after they were already aware of the writing abilities required by such a degree in EMI form. The items were extracted from Cai (2013). But not like the way to ask students to brainstorm for the necessary skills, it was designed as a multiple choice form to make it easier for the non-English-major students to finish this part.

Follow-up Interviews
A round of follow-up interviews was conducted after the survey's data were analyzed. They were meant to supplement the data derived from the questionnaires. Chinese was adopted for the interviews since the participants preferred to express themselves in their mother tongue. And the semi-structured interviews were around similar questions, for example, "Do you usually have to write a lot of assignments?", "What factors do you think pull your legs in getting used to writing academic essays in English?", "Why do you think those sub-skills are very difficult?", "Why do you believe those skills to be the most important ones worth teaching in undergraduate studies?".

Data Collection and Analysis
This research adopted both quantitative and qualitative methods. The primary source of data was from the self-report questionnaire survey, supplemented by follow-up interviews.
After the questionnaire was designed with the software of Sojump based on those for the former investigations, a pilot study was conducted among 5 subjects (who were also part of the formal survey).
Some modifications mentioned above were made, according to their suggestions. Questionnaires were then delivered through mobile phone to the taught postgraduates of 2015 after a call in the Wechat group of mainland students association. 82 students participated in the study willingly with a reward of some lucky money sent via that social network. After briefly checking their answers, one response was left out as his/her discipline was English education which was beyond the scope of this research. Thus the final sample size was 81. SPSS was employed to process the results.
After the survey analysis, 3 students were selected for the following interviews with some criteria: close IELTS/TOEFL scores (to ensure their similar English capacity), and yet different discipline backgrounds (because writing tasks vary from majors of arts to those of science). The finally chosen students were one from Master of Social Work, one Master of Early Childhood Education, and the other Master of Computer Science. All of them had an entry IELTS score of 6.5. The interviews were semi-structured with similar questions to each interviewee to make sure of the consistency of their answers on the questionnaires, and also gather richer details about their linguistic experiences in the past year and back in their undergraduate studies.
The interviews were carried out through Wechat with autonomous recording. Afterwards they were transcribed and translated into English for further analysis.

Research Question 1: Taught Postgraduate Students' Attitudes towards English Academic Writing
In Section I (Q.1-6), students were first asked to list the text types they had to write looking back on the master's degree studies (in multiple choices). Results (Table 2 presents the responses in a descending order of frequency) show that research papers gain the highest frequency (82.72%), followed by activity/experiment reports (59.26%), short summaries (55.56%) and case studies (50.62%). It seems that although this is a coursework master's degree, students still have a heavy workload of academic writing, evidenced by one participant (Master of Early Childhood Education) from the interview, "We have 6 essays, each of which has the word count of 2000, and 1 graduation thesis of more than 12000 words throughout the academic year". Apart from the voluminous writing work, one interviewee added that those academic genres were quite unfamiliar to them since they had never written them in the bachelor's degree studies, which in some way extended their accommodation period to several months (30.86% of the subjects were in this case) although still the majority (62.96%) of the students succeeded in this process within a few weeks. The major task in this section was to ask the students to assess their perceived importance of English academic writing on a Likert scale from 1 (not important) to 5 (very important). The outcome (M=4.6 out of 5, SD=0.68) is highly consistent with the findings from other studies (e.g., Evans & Green, 2007;Evans & Morrison, 2011;Cai, 2013). Academic writing is claimed to be the most important language skill in tertiary education including undergraduate and postgraduate levels, as students' written assignments are the principal determinant of their academic performance (Zhu, 2004). Some information retrieved from the interviews also contributes to this phenomenon from the perspective of the EFL context, "Often speaking and writing are the most difficult, but speaking is not as important as writing in my case, because out of class, I usually communicate with my friends in mandarin, not English".
With regard to the interest in English academic writing, those participants indicated it on a scale from 1 (not interested) to 5 (very interested). The percentage of students showing interest in writing (by conflating the responses on Scale 4 and 5) turns out to be 66.67%, occupying the majority of all. Here the item design is in need of some improvement as the corresponding one in Cai's (2013) questionnaire which further explored the type of interest whether it was intrinsic or for external factors like "the purpose of completing current graduate studies" (Cai, 2013, p. 8).

Writing
In Section II (Q.7-8), students in the beginning were asked the situation of academic writing instruction during the postgraduate year. An overwhelming majority (92.59%) of them chose either "a few hours' workshop" or "I had to handle it all by myself". It seems that HKU's taught postgraduate program is not equipped with specialized courses on EAP (English for Academic Purposes) which are provided for undergraduates through 13 weeks. The assumed causes are the tight schedules for those postgraduate curricula and potentially huge investment into the courses establishment.
As for the effectiveness of the workshop in improving students' writing abilities, one subject from the interview reported that the workshop of this sort did not help much with their actual academic writing.
The content and procedures of the workshop were rather fixed, not specific to their discipline or individual needs. What was taught covered the basic writing skills, like how to search for academic resources through the university's e-database, how to cite in APA format and the like.
The body of this section was to reveal students' perceived difficulty degrees of the 13 sub-skills of English academic writing, which were all indicated on a 5-point scale (1=very easy, 5=very difficult).
Those sub-skills fall into two categories: general academic writing skills and language issues. Table 3 presents the mean and SD scores of each item. Within the category of general academic writing skills, the items of commentaries and discussions on the data, and reviewing and critiquing the previous written work are regarded as the most difficult, with the mean scores of 3.47 and 3.43 respectively. By contrast, the skills of searching for appropriate literature using databases and library resources (M=2.95, SD=0.88), and particularly, writing references/bibliography (M=2.62, SD=1.00), are the two with the greatest ease. The interviewees' explanations for the most difficult items are that the two skills require analyzing and synthesizing the information in the literature and collected data in a quite logical and critical fashion, which calls for thinking of higher order. They needed more guidance and practice to get those skills grasped. On the other hand, the skills of resource searching and reference citation are the most basic to survive in the academic life. They are often taught explicitly if there is any instruction on academic writing.
In respect to the language issues, the item of using proper academic language and vocabulary is noted with surprise. Deemed as the most difficult language problem, it also has the highest mean score (3.73) of all the 13 writing skills. Summarizing or paraphrasing is another challenging language-related skill with a mean score of 3.41, closely followed by the ability of expressing ideas in grammatically correct English (M=3.36). The rest of the language-related skills all have the mean scores of higher than 3.0, which implies the students' low confidence in the language for writing. One interviewee said that, "Personally, language is a very big problem to me. What I write is just not something academic. I have trouble in using the formal and 'big' vocabulary. Mine are often very easy, and low…Also, I often make grammatical mistakes, although we did a lot of grammar exercise drills in high school. I just can't apply the theories learnt into real practice".
Generally speaking, language issues (the overall mean score is 3.33) were reported to be more challenging than general academic writing skills (the overall mean score is 3.18). This outcome is right opposite to the findings in Cai's (2013) study. One reasonable assumption is that as English majors, the subjects in Cai's (2013) research are linguistically more competent than the counterparts in the current one, so that they generally felt less trouble with the language in writing. However, academic language has always been a troublesome issue for English learners, as was documented in other literature (e.g., Hinkel, 2003). Despite differing demographic information of the participants, the current research's results agree with Evans and Green's (2007): language is even more challenging than the general skills.
But the author attributes this coincidence to distinct causes. Prior to tertiary education, HK students have more experience with academic writing, thus familiar with the general skills needed, whereas students from the mainland do not. Meanwhile, the English education in HK attaches more importance to meanings than forms, which leads to the consequence that students entering university feel the linguistic problems not easy to tackle. Not like HK, the mainland's English education is notoriously known to be exam-oriented, producing graduates who excel in grammar drills in the test papers but have little practice of applying the knowledge into use. Hence their actual performance of accurately and appropriately employing language is still disadvantageous, which results in their lack of confidence in this respect.  Next, students were asked to choose the 5 most important English academic writing skills they wished to have been taught over the undergraduate studies, after they were already aware of the writing performance expected by such a degree in EMI form. When analyzed, those options were sequenced and put into four categories: information searching, language, writing and structure, and thinking.

Conclusion
This small-scale investigation delineates a picture of mainland non-English-major students' experiences of English academic writing in the coursework master's degree studies in HKU, with exploration of their academic support from both the undergraduate and postgraduate programs. With the basic writing skill instruction from the current departments and almost no coverage of this academic area in the undergraduate English course, this cohort of upper-intermediate English learners is generally able to succeed in accommodation to academic written work in an EMI environment. However, some perceived obstacles are still present, embracing the general writing skills, and more strikingly, the language-related issues. Students feel it more challenging when it comes to the skills requiring more complicated higher-order thinking to synthesize and critique academic works. Yet what concerns them even more is to produce the right lexis, style, and grammar for the academic discourse. And they expect the undergraduate English course to have prepared them with the linguistic competence.
The survey results could provide some prompts for the direction of university English education reform. Cai (2013) proposes the establishment of genre-based EAP course for the undergraduates to equip the graduates with general writing skills and linguistic resources for the future academic career in EMI programs. However, given the long-standing mainland university education style of test-driven language instruction, it seems too radical to introduce a completely brand-new course by breaking the existing curriculum system. In this sense, it is more reasonable to gradually embed in the current university English course the "task-based and content-driven" pedagogical approach (Evans & Green, 2007, p. 14). This instructional method emphasizes language analysis, and its application that the current mainland English learners lack most.
Admittedly, there are some limitations to this project. Restricted to time and resources available, it had a small sample size, which could only reveal an in-depth focus instead of a panoramic landscape of the phenomenon. A variety of data sources are desirable, including not only the perceptions from the learners, but also the lecturers and program's administrators at both the postgraduate and undergraduate level. Also, another method of delivering the questionnaires face to face with the subjects is preferred to Wechat, a virtual network that may have a negative impact on the quality of collected data. Finally, as a needs analysis, it should not be one-off, since activities and perceptions evolve continuously, so that conclusions are to be checked and assessed constantly.