Grammatical Errors in Madrasah Aliyah Students’ Narrative Texts: An Error Analysis of the Surface Strategy

Framed by Error Analysis (EA) approach, this study aimed at analyzing grammar errors in writing narrative text, with a special focus on the differences of errors made by the students of Mathematics & Natural Science (MNS) and Social Studies (SS) streams at a Madrasah Aliyah (Islamic senior high school). This study used 60 narrative essays as the data corpus, which was analyzed using the adoption of Coder‟s EA approach. The result of data analysis showed that MNS students made more grammatical errors than SS students did in writing narrative texts. A closer look at the data revealed interesting findings as follows: 1) both MNS and SS students made grammatical errors in six categories, noun, pronoun, verb, article, preposition, and conjunction; 2) there were five types of errors found in students‟ writing including mis-selection, over inclusion, and omission as the dominant ones; and 3) MNS students produced more errors in five categories; noun, pronoun, verb, preposition, and conjunction while SS students were only in one category, namely article. These findings could be used as a proof to debunk the stereotype of MNS and SS students labeling with more preference given to MNS students who are perceived to be more dominant in academic performance.


Introduction
The importance of English as a global language has been long recognized by the government of Indonesia. The language was officially approved as the first foreign language in the country in 1955 and has become the only foreign language mandatorily taught in all levels of education since then. Based on the 2013 Curriculum (K-13), which has been implemented in all levels of high school since 2014, the teaching of English as a foreign language in Indonesia (TEFL) should focus on developing language skills on four areas namely listening, speaking, reading, and writing. Furthermore, it is stated that the goal of TEFL in Indonesia is to equip students with the ability to develop oral and written communicative competence Peraturan Menteri Pendidikan dan Kebudayan, 2014;Marzulina et al., 2018;Erlina et al., 2020).
Despite the stated goal related to English communicative competence in both oral and written forms, the teaching of writing skill has been neglected in English classrooms including that at SMA (general senior high school) and MA (Islamic senior high school) levels. According to the K-13, writing instruction in SMA and MA aims to develop students" skills in writing simple texts to more complex ones in various genres including narrative text. However, in practice, writing instruction is still much dominated with activities to reinforce language structures at the sentence level. Writing is still viewed as a product activity emphasizing merely on grammatical and syntactic accuracy Erlina et al., 2019;Fithriani, 2017;Mukminin et al., 2019). As a consequence, most Indonesian EFL learners at the SMA/MA level perceive writing as a difficult skill to develop and master, particularly in terms of accuracy at the sentential level, that is, sentence grammar.
Studies have shown that Indonesian SMA/MA learners face various difficulties related to develop their English proficiency in written discourse (Inayah, & Gani, 2016;Nurhayati, 2015;Husna, Zainil, & Rozimela, 2013). In general, they were found to commit various errors, particularly grammatical ones in writing different genres as stated in the K-13 curriculum. Several studies have been conducted to analyze gramm ar errors on one particular genre of writing, namely narrative text. The focuses of discussion are diverse, which include the frequency of grammar errors emerged in students" writing (Floranti & Adiantika, 2019;Murdliyana, 2019;Hendriwanto & Sugeng, 2013), the types of grammar errors (Hendriwanto & Sugeng, 2013), the possible sources of errors (Floranti & Adiantika, 2019), and grammatical error analysis across different grades and streams (Luthfiyati, Latief, & Suharmanto, 2015).
Despite the growing number of this path of research, most of them involved SMA students as the subjects of the research. Very few studies found in literature have focused on grammar errors made by MA students, let alone on the contrastive analysis of errors made by MA students of MNS and SS streams. Since SMA and MA are under the responsibilities of two ministries; the Ministry of Education and Culture (Kementerian Pendidikan dan Kebudayaan or Kemdikbud) and the Ministry of Religious Affairs (Kementerian Agama or Kemenag) respectively, the curriculum for English instruction implemented in these two types of secondary schools might be different due to the regulation allowing education institutions under Kemenag to modify and/or develop the national curriculum in accordance with their needs (Peraturan Menteri Pendidikan dan Kebudayaan Republik Indonesia, 2018). Regarding this difference, it was interesting to find out whether grammatical errors made by MA students in writing narrative text will replicate or contradict those of SMA students. Therefore, this study was conducted to fill the gap by contrasting the grammatical errors made by MA students of MNS and SS streams in writing narrative text. In details, it focused on answering three questions: 1) what grammatical errors were made by MA students of MNS and SS streams in writing narrative text?; 2) what types of grammatical errors were made by MA students of MNS and SS streams in writing narrative text?; and 3) what were the differences of grammatical errors made by MA students of MNS and SS streams in writing narrative text?
The significance of this study stemmed from the idea that by identifying the grammar errors made by the students of the two streams, the teachers will gain a significant insight on what strategies they employ to develop communicative competence in English written discourse, particularly in writing narrative essay. Moreover, the findings of this study may work as a diagnostic tool to reveal the grammatical problems faced by the students in writing so that they could be the focus of writing instruction for Indonesian SMA/MA students in general. Finally, since there is a stereotype of MNS students always outperforming the SS ones academically, this research could be used to either validate or debunk it.

Grammar knowledge and English writing ability
In today"s global community, communicative competence in foreign languages, particularly in English is becoming increasingly important. English as a global language could be claimed to hold the most important role as a means of communication in various fields such as education, diplomacy, and international commerce. Among English language skills, writing has a special position due to its extensive usage in educational as well as professional settings. Without a good command in writing, one could not express her ideas when doing most of daily tasks such as writing essays, reports, applications, e-mails, etc. Thus, one"s writing ability in a language could be used to indicate her proficiency in the language (Li & Lin, 2007).
Writing comprises a complex mental process involving multiple skills from cognitive analysis to linguistic synthesis. Thus, to become proficient in this particular language skill takes a considerable time and effort. For foreign language learners, the challenge and difficulty in developing writing proficiency is even doubled due to various reasons ranging from the language barriers to the different rhetoric patterns deployed (Fithriani, 2018b). In addition, the ability in foreign language writing including in EFL context involves knowledge of other language aspects, particularly grammar.
Since grammar is bounded to other language skills like listening, speaking, reading, and writing, it is no doubt that good grammar knowledge is inevitably important in developing EFL communicative competence as it guides learners to use the language appropriately both in written form and in oral language skill (Fithriani, 2018a). In terms of the role of grammar in writing skill, Hedge (2005) argued the following: Effective writing requires a number of things: a high degree of development in the organization of ideas and information; a high degree of accuracy so there is no ambiguity of meaning; the use of complex grammatical devices for focus and emphasis; and careful choice of vocabulary, grammatical patterns, and sentence structures to create a style which is appropriate to the subject matter and the eventual readers. (p. 5). In short it could be concluded that good writing requires good working knowledge of grammar.

English writing instruction in Indonesian secondary schools
English is the only foreign language taught in secondary schools and higher education institutions in Indonesia. Secondary schools in Indonesia consist of three years of junior high school and three years of senior high school and fall under the responsibility of either Kemdikbud or Kemenag. Furthermore, the secondary schools under Kemdikbud are known as SMP (Sekolah Menengah Pertama) for junior high school level and SMA (Sekolah Menengah Atas) for senior high school level. Meanwhile, those under Kemenag are called MTs (Madrasah Tsanawiyah) and MA (Madrasah Aliyah) respectively. However, the curriculum for English instruction throughout these levels becomes the sole responsibility of Kemdikbud to design and issue.
According to K-13 as the currently implemented curriculum, the allotted time for English teaching at secondary school ranges from two up to six class hours (one class hour equals to 45 minutes) per week. For junior high schools, the allotted time is four class hours throughout grade 7 to grade 9. For senior high schools, the allotted time varies according to the stream based on students" academic interests. At SMA level, students are streamed into three academic interests, namely the Mathematics and Natural Sciences (MNS), the Social Studies (SS), and the Language and Culture (LC), while at MA level, the three streams are offered with an addition of the Religious Studies (RS). For the first two streams, English is compulsory and allotted at least two class hours per week. However, for Madrasah Aliyah, the curriculum structure can be developed in accordance with the needs set by the Ministry of Religion (Peraturan Menteri Pendidikan dan Kebudayan Republik Indonesia, 2018). This exception implies that there are possibilities for some differences of TEFL in the two types of senior high schools. Meanwhile, related to this academic interest-based division, there is a stereotype referring to the academic ability of MNS and SS students. People believe that MNS students are dominant in academic performance in almost, if not all, subjects (Efendi & Wahyudi, 2016;Hanafi, 2014). As a consequence, most students of secondary school prefer to take MNS compared to SS stream.
As stated in K-13, ELT at secondary schools should equip students with the ability to develop oral and written communicative competence in four areas of language skills, namely listening, speaking, reading, and writing. Furthermore, the teaching of writing provides students with knowledge, experience and strategies of writing simple texts to more complex ones in 12 text genres, namely: recount, narrative, procedural, descriptive, report, news items, analytical exposition, persuasive exposition, spoof, explanation, discussion and review. For this purpose, genre-based approach is used to introduce students the social function, the generic structures, and the language features of the genres.
Among the 12 text genres taught to SMA and MA students, narrative may be considered special and unique as it is the most frequently used and learned throughout the grades of secondary school. Narrative is defined as a literary text that tells about a series of logically and chronologically related events (Lukens, 2007). According to Anderson and Anderson (2003), the social function of a narrative text is to entertain the readers with a story containing complications or problematic events that lead to a crisis and in turn find a resolution. The generic structure of a narrative text includes 1) Orientation to introduce the participants and inform the time and the place; 2) Complication to describe crises the participants need to deal with; and 3) Resolution to show the way of the participants to solve the crises. As it talks about the past events, narrative text should be written in simple past. In addition, writing in narrative genre should include adequate details about the action and its context so that the readers can understand what is going on.

The significance of Error Analysis (EA) in EFL teaching and learning
Errors are an inseparable aspect of learning a second language, thus, it is important to obtain information how second language (L2) learners, including those of EFL make mistakes while performing oral or written tasks in the target language. One of the most popular method to conduct this particular type of research is Error Analysis (EA). Crystal (2003) defines EA as a "technique for identifying, classifying and systematically interpreting the unacceptable forms produced by someone learning a foreign language, using any of the principles and procedures provided by linguistics" (p. 165). Similar to this, Corder (1974) believes that EA is a useful technique to describe L2 learners" knowledge of the target language in order to relate it to the teaching they have been receiving. Related to the stated definitions, EA is particularly beneficial for L2 teachers and/or researchers who want to obtain information about learners" linguistic deficiency so they can address the issues in the future. In addition to identifying the errors, Richards and Schmidt (2002) state that other benefits of EA include the identification of the learning strategies used by L2 learners and the causes of the errors they make.
EA was first introduced in 1960"s by Corder (1967) as a result of his critics towards Contrastive Analysis (CA), which was claimed to fail in predicting errors made L2 learners. CA assumed that errors made by L2 learners are exclusively derived from the interference of their L1 or mother tongue, known as interlingual source. Meanwhile, in addition to interlingual source, EA made clear that errors in L2 learning also come from intralingual source, which result from faulty or partial learning of the target language. Guided by the two major purposes of EA, either to present error categories based on observable characteristics or to report the types of error observed, EA categorizes errors into four taxonomies, namely (1) linguistic category; (2) surface strategy; (3) comparative analysis; and (4) communicative effect.

James' surface strategy taxonomy
Within the framework of Error Analysis, errors could be categorized into four taxonomies. Among those four taxonomies, surface strategy has much been used as an approach in analyzing learners" errors in various EFL contexts (Mohammed & Abdalhussein, 2015;Zheng & Park, 2013;Yoon, 2012;Vasquez, 2008). Many experts have tried to classify the errors EFL learners make based on the surface strategy taxonomy. James (1998) categorizes the errors into five types as detailed below: 1. Omission, which is characterized by an absence in a well-formed utterance of an item. 2. Over Inclusion, which is characterized by the presence of an item which must not appear in a well-formed utterance. 3. Misselection, which is characterized by the use of the wrong form of the morphemes or structure. 4. Misordering, which is characterized by the incorrect use of morpheme or a group of morphemes in a pronunciation. 5. Blends which is characterized by the combination of two alternative grammatical forms to produce an ungrammatical.
He furthermore classified them based on the sources, which include; 1) Interlingual errors referring to the errors caused by the interference of learners" mother tongue on target language (TL) learning; 2) Intralingual errors referring to the errors caused by the TL itself; 3) Communication strategy-based errors, which are subdivided into the holistic strategies or approximation referring to the errors caused by learners" assumption that it is all right to use near-equivalent items to substitute the required items in TL and the analytic strategies or circumlocution referring to the errors made by expressing the concept indirectly rather than by direct reference; and 4) Induced errors referring to the errors caused mostly by the TL teaching and learning process.

Research design
This study tried to find out the grammatical errors and the types of those errors which were found on the narrative texts written by 11 th graders of MA students taking MNS and SS streams. Regarding the aims, this research thus employed quantitative descriptive design using simple quantification or cross-section survey.

Research site and participants
This study was carried in Madrasah Aliyah Laboratorium, the State Islamic University of North Sumatra Medan (MAL UIN SU) during the 2018/2019 academic year. This particular school was chosen as the research site due to some considerations. Firstly, due to the aspect of accessibility since I worked in the same institution which allowed me to conduct educational-related studies without involving too much administrative requirements to fulfill for research permission. Secondly, based on preliminary data collection, the English teacher shared similar interest to gather information about the main grammatical problems that her students encounter in their writing production and use the information in deciding what to focus for future teaching.
The population of this study included the eleventh graders of MAL UIN SU students, who were divided into four classes; two classes of MNS with a total number of 55 students; one class of SS with 35 students, and one class of RS with 28 students. Thirty students (18 female and 12 male) equally representing MNS and SS streams were selected by using stratified random sampling. Since the students of MNS and SS streams could be divided into three homogeneous subgroups based on their writing skills (high, moderate, and low), the disproportionate stratified random sampling was employed to ensure the representativeness of each subgroup.

Data collection
This study used students" writing products as the main instrument of data collection. The corpus of the study involved 60 narrative texts written by the students, each consisted of 200 to 250 words. The students were given four topics which included the Legend of Lake Toba, Cinderella, Malin Kundang, and Red Riding Hood, representing the nuance of the themes the students learnt during their study, those which were nationally and internationally popular. After selecting the topic, each student was asked to write two narrative texts in two topics of their choice from the selection with a provided time of 30 minutes to write a narrative text. It is necessary to emphasize that the two writing products were not written in one class meeting, but two with a week time interval.

Data analysis
The data were analyzed adopting Corder"s (1967) approach of Error Analysis (EA), which involves four stages, namely: writing sample collection, error identification, error classification, and error quantification. To ensure the consistency of the findings, inter-rater reliability was employed in analyzing students" writing during the identifying and classifying stages. Two writing instructors at MAL UIN SU were involved in identifying grammar errors on students" writing and furthermore classifying them into six categories namely noun, pronoun, verb, article, preposition, and conjunction. Later, their ratings were reconciled by me. At the quantifying stage, the errors were calculated in order to find out how frequently each type of errors were made by the students. To obtain the numerical data, the following percentage formula was used: in which, P : percentage of each error n1 : total error of each category ƩN : total of the whole errors made By calculating the percentage of each error, the most and the least frequent error made by the students could be identified.
In order to find out the type of errors made, James" surface strategy taxonomy was adopted in analyzing the data (James, 1998). The errors identified by the raters were labeled into five types; omission, over inclusion, misselection, misordering, and blends.

Findings
This study tried to contrast grammar errors found in narrative texts written by 11 th graders of MA students taking Mathematics and Natural Science and Social Science streams at MAL UIN SU . The result of data analysis showed that in total there were 450 errors found in 60 narrative essays written by the students with the distribution of 250 occurrences made by MNS students and 200 errors by SS students as shown in Figure 1. In relation to the questions posed in this study, the findings were presented in three parts in accordance with the questions.

The grammatical errors made by MNS and SS students
To answer the question related to grammatical errors made by MMS and SS students, the texts were first grouped into two; those written by MNS students and those by SS students. The errors identified in each of students" writing products were recorded based on the six categories explained in the data analysis section and calculated to find out the total number of errors for each category. The results are shown in Table 1 and Table 2.  Table 1, MNS students made grammatical errors in all six categories. The "verb" category occupied the highest number of errors followed by the "article" one. These two categories of grammatical errors contributed 37.6% and 19.2% of the total errors generated in the data analysis process. "Preposition" and "noun" came in the third and the fourth positions in the list of grammatical errors made by MNS students comprising 14.4% and 10.4% of the 250 errors made. The errors in "conjunction" and "pronoun" shared almost similar contribution to the total errors identified in MNS students" narrative texts. The two categories of error comprised of 9.6% and 8.8% of the total errors, thus ranked the fifth and the sixth in the list.  Table 2 shows that SS students also made grammatical errors in all six types. The "Verb" category was found to be the most dominant error on SS students" narrative texts with the contribution of 39.5% of the total errors found. "Article" and "preposition" occupied the second and the third positions of the error frequency made by SS students contributing to 25.5% and 11.5% of the total identified errors. Ten percent of grammatical errors in the narrative texts written by SS students were in using conjunction. Finally, "pronoun" comprising 7.5% and "noun" 6% of the total errors generated in the data analysis process were in the bottom two of the chart as the least grammatical errors made by SS students. It means that each category of error contributed fewer than 10% out of the total errors identified.

The type of grammatical errors made by MNS and SS students
To answer the question related to type of grammatical errors made by MNS and SS students, each of identified error made by MNS students and those made by SS students were calculated on two separate tables and classified based on James" surface strategy taxonomy (James, 1998). The findings revealed that only one category of grammatical errors was found in all five types included in the taxonomy, namely "verb". As it can be seen in Table 3 and 4, the grammatical errors in using "verb" made by MNS and SS students were scattered in the whole types proposed by James (1998) while the rest of the grammatical errors could only be classified into either four or three types. In terms of the frequency of each type of errors occurred, the findings in Table 3 revealed that "misselection" became the type of grammatical errors mostly made by MNS students in their writings which comprised of 38% of the total errors made, followed by "over inclusion" with 29.6% and "omission" with 22.4% of contributions. Meanwhile, "misordering" and "blends" were quite seldom to occur as they only contributed 5.2% and 4.8% of the total errors made by MNS students. Based on the data presented in Table 4, "misselection" errors significantly outnumbered other type of grammatical errors found on SS students" narrative texts with 42% contribution. "Omission" ranked in the second as the most frequent type of errors made which comprised of 30% of the total errors. "Over conclusion" occupied the third position, followed by "misordering" with the contribution of 6.5%. Finally, SS students made very minimum errors in the "blends" type which comprised of only 3% of the total errors.

The difference of grammatical errors made by MNS and SS students
To contrast the grammatical errors and the types of error found in narrative texts written by MNS and SS students, the findings used to answer the question 1 and 2 in this study were displayed side by side and presented in a chart for each category as seen in Figure  2 and Figure 3.
The findings in Figure 2 indicated that both MNS and SS students made grammatical errors in their narrative text in all categories. In terms of the frequency of the errors made, the grammatical errors found in MNS and SS students" texts were dominated by the "verb" ones, followed by "article" and "preposition". However, there were some slight differences in the bottom three of the chart. The analysis of the data collected from MNS students revealed that they made more errors in using "noun" compared to "conjunction". Meanwhile, SS students made errors in using "conjunction" more often than in "pronoun". "Pronoun" became the fewest grammatical errors found in MNS students" texts while in SS students" texts, "noun" was the fewest one. Furthermore, when comparing the total of grammatical errors made, it was found that MNS students made more errors in five categories, namely: "verb", "preposition", "conjunction", "noun", and "pronoun". While SS students made more errors only in "article."

Figure 2. The difference of grammatical errors made by MNS and SS students
As shown in Figure 3, the grammatical errors made by MNS and SS students in their narrative texts were scattered in the five types listed in the surface strategy taxonomy. A closer look at the findings revealed that most of the errors made by the two groups of students were "misselection". In the second rank of the type of errors made, it was found that "over inclusion" occurred more often than "omission" for grammatical errors made by MNS students. On the contrary, SS students made more errors in "omission" compared to those in "over inclusion".

Discussion
As the findings showed, the students participating in this study made numerous grammatical errors when writing narrative texts. It indicated that 11 th grade of MA students at MA UIN SU had insufficient knowledge on how to employ several English parts of speech in their written discourse. It could also be inferred from the result that the verb and the article constituents were the most challenging area for the students. Furthermore, in relation to the findings showing the types of the errors made, it could also be inferred from the findings that those errors occurred because the students either used the wrong form of morphemes or structure, added an unnecessary item to a well-formed utterance, or omitted certain items from a well-formed utterance. I argued that these challenges occurred due to the interference of the students" mother tongue on the target language, which is also known as interlingual errors because Indonesian language employs more different grammatical rules than those in English, such as the absence of interdependent participles in verbs which determine the tense of the sentence and the linguistic element that marks the plurality of a given noun.
The findings of this study, at some extent, were correspondent yet inconsistent with those of other studies discussing similar topics. First, the finding showing the verb constituent as the most challenging area for senior high school students in writing narrative texts was also reported in Floranti and Adiantika"s (2019), Luthfiyati, Latief, and Suharmanto"s (2015), and Murdlyanas"s (2019) studies. A closer look at the findings of the three studies also revealed some similarities with that in this current study, in which the most dominant verb errors were the verb tenses. However, for the second rank of the grammatical errors frequently found in students" writing, the result of this study is inconsistent with those of Floranti and Adiantika"s (2019) and Murdlyanas"s (2019) showing pronoun as the second predominant type of error found in their students" essays. Second, the result related to the types of grammatical errors corresponds to the works of Floranti and Adiantika"s (2019) who found "misselection," "omission," and "over inclusion" as the most dominant types of errors the students in their study. However, since their study used Dulay, Burt, and Krashen"s (1982) taxonomy of surface strategy as their theoretical framework, the terms used are slightly different where "misinformation" is used for the occurrence of misselection proper constituents and "addition" for the use of unnecessary elements.

MNS Student SS Students
Furthermore, since the findings of this study generally correlate to those conducted in SMA context, it could be assumed that the modification of English instruction curriculum in secondary schools under the responsibility of Kemenag does not significantly impact the outcomes in terms of students" grammatical mastery in writing narrative text.
Finally, the finding indicating that MNS students made more errors in their essays than SS students did is contrary to the result found by Sembiring and Gintings"s (2013) research showing natural science students outperformed social science students in writing recount texts. Thus, this finding may also imply that the general understanding of society related to the labeling of SS students who are considered to have lower academic competence than MNS students is not necessarily true.

Conclusion
This study investigated the grammatical errors made by the eleventh graders of MAL UIN SU students of MNS and SS streams. The results of data analysis generated some interesting findings which can be concluded into three points. Firstly, the students of MNS and SS streams made grammatical errors in their narrative texts, which can be classified into six categories, namely; "noun", "pronoun", "verb", "article", "preposition", and "conjunction". Secondly, the grammatical errors made by the students covered the five types listed in James" (1998) surface strategy taxonomy, namely: "omission", "over inclusion", "misselection", "misordering", and "blends". Furthermore, "misselection", "over inclusion", and "omission" were found to be the types of errors that the students most frequently made in their narrative writings. Finally, the contrastive analysis of the grammatical errors made by MNS and SS students revealed that the former group of students produced more errors in five categories; "noun", "pronoun", "verb", "preposition", and "conjunction" while the latter only in one category, namely "article".