AN INVESTIGATION INTO THE GAP BETWEEN KOREAN UNIVERSITY STUDENTS’ GRAMMATICAL AWARENESS AND PERCEPTION

Despite the consistent emphasis on grammar instruction in English classrooms in South Korea, studies regarding grammar instruction have not yet been extensively conducted. The present study aims to discover the gap between learners’ grammatical awareness and their perception of major grammatical items. A total of 60 EFL learners from two local universities in South Korea participated in the study and were divided into two groups, a high-level and a low-level group. A set of tests was utilized to examine learners’ grammatical awareness and their perception of six major grammatical items—tense, prepositions, articles, voices, morphology, and vocabulary. The results demonstrated that there was a significant difference in the scores of tense, article, and voice for grammatical awareness between the high-level and the low-level group. Also, both groups scored high for the category of voice while they received low scores for vocabulary and morphology. In addition, they showed a significant difference in the scores for the perceived difficulty of articles and voice. The high-level group perceived voice as the most difficult, whereas the low-level group perceived articles as the most difficult. These findings demonstrate a gap between the learners’ grammar awareness and perception and highlight a need to design an individualized curriculum for the effectiveness of teaching as well as self-initiated studying.

previous studies concerning learners' grammar difficulty and their performances based on an error analysis of their writing.Then, there will be a review of the results and a discussion in terms of the learners' awareness of major grammatical features as well as perception of the degree of difficulty depending on the learners' linguistic proficiency level, following the description of data collection and analysis of this study.Such an investigation intends to fill the gap between the teachers' and students' perceptions of the difficulty of grammar.In this sense, this study will help delineate learners' challenges in the development of L2 writing and design a writing curriculum tailored to learners in different levels.

Error analysis and grammatical difficulty
When ESL or EFL teachers read and give feedback on students' writings, they need to keep in mind the distinction between the errors and the mistakes.The clear distinction between these two notions can be accounted for by acknowledging the concepts of language competence and language performance.Researchers (Corder, 1981;Gass & Selinker, 2001;Mourtaga, 2004;Yusel, 2007) have explained that errors are made by the learners' inadequate knowledge of the target language in language competence and cannot be self-corrected, while mistakes are caused by the slips of tongue in language performance and are self-corrected.In addition, Corder (1981) emphasized the importance of error analysis for ESL writing instruction, stating that "the systematic errors that are made by learners tell us something about the learner's interlanguage, or underlying knowledge of the rules of the language being learned" (p.10).Since ESL or EFL learners cannot identify their own errors in writing themselves, teachers play a very important role in helping them reduce errors by teaching them the correct language.In order to plan lessons and develop teaching materials that are well tailored to learner's proficiency levels in writing classes, it is essential for language teachers or researchers to collect and analyze a range of errors made by learners.
Researchers have proposed that error analysis makes a significant contribution to ESL writing.Corder (1981) saw error analysis as an important means of evaluating the learning process of writing from different perspectives as noted in Zhang (2011).First, for teachers, the learners' errors could tell them how far towards the goal the learners have progressed and, consequently, what remains for the learners to learn.Secondly, for researchers the errors provide evidence of how language is learnt or acquired and what strategies or procedures the learner is employing in his/her discovery of language.Thirdly, for learners, committing errors is a focused way the learner has of testing his/her hypotheses about the nature of the language he is learning (Zhang, 2011).In line with Corder's proposal, Lee (2011) has a similar point of view, suggesting that error analysis is very useful for tapping into evaluating a learner's linguistic competence, figuring out what types of errors learners have, and designing a writing class depending on their linguistic proficiency levels.
Recent works on error analysis have been investigated along with the notion of grammatical difficulty, and have attempted to define grammatical difficulty with the idea of easy rules and hard rules from different approaches.Berent (1985) and DeKeyser and Sokalski (1996) defined grammatical difficulty with respect to comprehension and production, and noted that the degree of L2 learners' comprehension and production could vary depending on different grammatical features.Berent (1985) conducted a study to investigate whether there is a significant difference between ESL learners' production and comprehension for the different types of conditional sentences.The results showed that real conditionals were the easiest to produce, but the most difficult to comprehend, but in case of past unreal conditionals, they were the easiest to comprehend, but the most difficult to produce.DeKeyser and Sokalski (1996) noted that for English speakers learning Spanish conditional forms of a verb is easy to comprehend, but difficult to produce, while Spanish direct objects are easy to produce, but difficult to comprehend.Collins et al. (2009) considered easy rules and hard rules in relation to an L2 acquisition perspective.They distinguish easy rules from hard rules by the extent to which the rules are acquired early or late.According to their work, easy rules refer to features that are acquired early, while hard rules refer to those that are acquired late.In their study, for example, English progressives were considered easy rules because ESL learners tend to acquire this feature early.On the other hand, English simple past is considered a difficult feature since it tends to be acquired late.Scheffler (2009) considered grammatical difficulty from L2 learners' perception.He examined how the difficulty of grammar rules pertained to the perceived usefulness of L2 instruction by using a questionnaire with a 5-point Likert scale targeting 50 Polish EFL students.The results showed that the learners felt that they benefitted greatly from class when a large number of grammatical features they perceived to be difficult to learn were taught in class.
A discussion about grammatical difficulty is still in progress on a number of fronts, including the inherent complexity of the feature (Hulstijn, 1995), linguistic form, semantic meaning, pragmatic use (Larsen-Freeman, 2003), objective/subject difficulty (DeKeyser, 1995), and implicit/explicit knowledge (Ellis, 2006;2008).Although many theoretical accounts of the issues of grammatical difficulty have been proposed by researchers, empirical works from the learner's perspectives are still thin on the ground.Since empirical studies play a role in complementing theoretical accounts, the present study attempts to provide corroborating evidence from L2 learner's perspectives.Specifically, this study explores grammatical awareness and perceived difficulty regarding English writing from the Korean university-level learners' perspective and depending on their linguistic proficiency.Furthermore, the results regarding the relationship between grammatical awareness and perceived difficulty can provide a useful guideline for L2 pedagogy and lesson plans in writing classes.

Previous studies regarding written errors made by EFL learners
Despite some well-known weaknesses that error analysis entails, such as paying too much attention to language errors, neglecting the positive reinforcement of correct language, overemphasizing the production data rather than the comprehension of language, and failing to identify learners' use of avoidance strategy (Brown, 2007), studies based on error analysis have been widely conducted on account of their significance in learners' development or progress of language.Based on the findings of the previous studies of the written errors made by Hong Kong Cantonese learners, Chan (2010) noted that the errors were normally associated with relative clauses, plural marking, and topicalization, and seemed to be attributed to first language transfer.Green (1991) investigated the topic-comment structure produced by Hong Kong English learners and concluded that the overuse of this structure may be accounted for by the influence of the mother tongue.In addition, Budge (1989) discovered Hong Kong learners' written errors for plural marking (e.g.adding -s, es) and described them as evidence of typological transfer which is also affected from their first language.
Some studies analyzed written errors within an extensive framework that overviews most of the major features of writing.For example, Zawahreh (2012) examined the written errors of Jordanian English learners in various categories, such as morphology, syntax, and vocabulary.It was discovered that Jordanian English learners made errors in terms of the disagreement between the subject and the main verb, wrong verb tenses, as well as the incorrect use of vocabulary.Additionally, they tended to omit the main verb and prepositions necessary for the sentence structure.Wu and Garza (2014) also looked into the types and attributes of English written errors through the emails produced by the learners in the EFL context within the taxonomy of grammar, lexis, semantics, mechanics, and word order.It was revealed that most written errors were based on the interlingual errors rather than intralingual or developmental errors.The errors that they made were majorly grammatical, and among them ones regarding subject and verb agreement were most frequent.They accounted for 12% of the errors, followed by errors concerning sentence fragments and sentence structures.
Nevertheless, although English curriculum in secondary schools of Korea has been constantly involved with teaching grammar with a focus on enhancing learners' reading comprehension skills, there have been only a few studies that investigated grammatical errors frequently made by Korean learners, as pointed out by Kang and Kim's (2014) meta-analysis.In spite of the paucity of studies concerning the investigation of Korean learners' grammatical errors, it has been rather frequently observed that Korean learners have difficulty in employing the English articles appropriately in their writing (Choi, 2011;Chung & Lim, 2005;Lee, 2007;2008;Master, 1987;Park, 2009;Song & Park, 2001).For instance, Lee (2007) investigated the patterns of the wrong uses of definite and indefinite articles (e.g.a, an, the), produced by learners in the beginning, intermediate, and high-level levels.She found that the beginners had a low degree of knowledge of article usage, while the intermediate learners employed the definite article correctly.However, the advanced learners showed an unexpected pattern of article uses in that they often made errors at the basic level where the answers could be readily expected based on the general descriptions of the usage of articles.Also, Lee (2008) shared the congruent results from investigating the errors made by the advanced nonnative English teachers.Along this line, Park (2009) conducted a case study to analyze the types and causes of English article errors made by a Korean advanced learner of English and reported considerable cases of article errors were induced by a lack of attention as well as insufficient knowledge of article usage.Further, Chung and Lim (2005) specifically researched Korean English learners' knowledge as to the usage of the articles related to the conversion of noun countability, depending on different age groups of learners.They found that the overuse of the definite article was predominant among university-level learners, whereas middle school students employed the indefinite article more commonly.
Additionally, as for Korean learners' errors in writing, Choi (2011) also found that Korean university level-learners made errors most frequently in terms of vocabulary use.They showed difficulty selecting the appropriate vocabulary as well as constructing grammatically correct noun phrases.It was also shown that there were several errors associated with the wrong usage of articles in the noun phrases.Jung (2006) investigated university level-learners' errors in writing and found grammatical errors in constructing sentences in the passive voice.They often tended to make errors of employing appropriate be-verbs (e.g. am, are, is, was, were, etc.).
As for the studies of learners' perception, there are two main proposals with relation to perception and production in L2 acquisitions, specifically that perception precedes production (Fledge, 1995;Best, 1995;Best et al., 2001) or production may precede perception (Sheldon & Strange, 1982).Previous studies have mostly focused on the production aspect of grammatical error analysis in EFL writing, while research on learners' perception in writing has been comparatively neglected.One recent study by Jahangir (2016) investigated Pakistani EFL learners' perceptions toward grammar in writing, and discovered that among three grammatical features, subject-verb agreement, tenses, and articles, the one the learners perceived as the most difficult was subject-verb agreement (49%), then tense (43%), and the least difficult was articles (40.5%).In accordance with actual writing performance, this study showed a similar result to Hourani (2008), indicating that subject-verb agreement is the most frequent type of error made by Pakistani EFL learners.Although the relationship between perception and production in writing in various EFL contexts should be examined, very few studies on the perception from the students' perspectives have yet been done.
Considering the lack of studies that have investigated Korean EFL learners' grammatical awareness and their perception of the degree of difficulty regarding specific grammar features, the present study aims to fill this gap by examining the following research questions: 1. Is there any difference in EFL learners' awareness of grammatical features depending on the linguistic proficiency levels?If so, which grammatical features are they?2. Is there any difference in EFL learners' perception of grammatical features depending on their linguistic proficiency levels?If so, which grammatical features are they?

METHOD Participants
A total of 60 Korean EFL students from various majors in a local university were chosen to participate in this study.They were divided into two groups, a high-level and a low-level group, based on TOEIC scores.The TOEIC scores were between 200 and 300 for the low-level group and between 700 and 850 for the high-level group.Among the 60 participants, 20 were male and 40 were female students.All participants were university freshmen.They reported that they had never studied abroad.

The procedure and instruments of data collection
To obtain a general understanding of Korean EFL university-level learners' awareness and perceptions of English grammatical features depending on two different language proficiency levels, two research instruments were employed to collect data for this study: a grammatical awareness test and a writing perception questionnaire.The grammatical awareness test and perception questionnaire were created for the study based on previous studies regarding L2 writing (Dulay et al. 1982;Thornbury, 1999;Choi, 2011).For these two tests, the students were asked to identify the correctness/incorrectness as well as the degree of difficulty regarding six major grammatical features, respectively: tense, prepositions, articles, voice, morphology (word form), and vocabulary (word selection).These six grammatical features were adopted from the major linguistic categories of errors created by Dulay et al. (1982) and Thornbury (1999) and modified according to the frequent errors that Korean students make in their writing.
First, the grammatical awareness test was performed to determine with which grammatical items Korean EFL students have difficulty when noticing errors in a written text.Second, a writing perception questionnaire was conducted to explore how difficult Korean EFL students perceive the six English grammatical features and to determine whether these two groups show any significant differences.
Additionally, an open-ended question asking about the general challenges in L2 writing was given in the perception questionnaire and the students were requested to write freely about it.For the study, the questionnaire format was used since it has the advantage of gathering a large amount of information within a short time and provides results that are easily quantified and analyzed (Dornyei, 2003;Gillham, 2007).
For the data collection, two steps were followed.First, participants were asked to respond to the grammatical awareness test which consisted of two short reading passages.The parts of the sentences pertaining to the six grammatical features were selected, and some of them were intentionally changed into grammatically incorrect sentences for the purpose of the study.36 sections of sentences were underlined; 18 were correct and 18 incorrect.Students were then asked to mark correctness or incorrectness of the underlined parts of the sentences.Table 1 illustrates the six grammatical features with explanations and some examples from the test.
The total number of questions for the grammatical awareness test was 36, including 6 questions for each grammatical feature.A correct answer was given 1 point, making a total of 6 points for each item.For this task, thirty minutes were allocated.After finishing the awareness test, the writing perception questionnaire was conducted for another 10 minutes.To help the participants better understand the questionnaire, a Korean version was provided.The questionnaire included 6 questions of a five-point Likert scale descending from "5" as very easy to "1" as very difficult.Further, the participants were asked to write freely in either Korean or English about the challenges that they normally face in L2 writing.

Data analysis
Data obtained were then analyzed using SPSS version 23.Descriptive statistics and independent sample t-tests were performed to compare the students' grammatical awareness and perceptions depending on the different linguistic proficiency levels.In addition, the students' responses to the open-ended question were analyzed qualitatively, following the procedures of the grounded theory (Dörnyei, 2009).

RESULTS AND DISCUSSIONS
The results indicated that there were statistically significant differences between the high-level and low-level groups with respect to grammatical awareness test and the writing perception questionnaire.The following reports both quantitative and qualitative results.

Results regarding the learners' grammatical awareness test
The results of the grammatical awareness test between the two groups depending on linguistic proficiency levels are displayed in Table 2.With the proficiency levels collapsed, the results showed that the participants received the highest scores for the category of voice with a mean of 4.77 and the lowest score for vocabulary with a mean of 3.08 (voice > article > preposition > tense > morphology > vocabulary).
When the two groups were considered separately, they also showed a similar tendency in the distribution of scores, even though the high-level group performed better than the lowlevel group for all the grammatical features.Namely, Korean university-level learners were able to recognize the errors related to English voice and articles well, regardless of their linguistic proficiency levels.However, they had difficulty in selecting the appropriate vocabulary as well as using the correct word form.
The results of the current study showed some differences with previous studies regarding which grammar features Korean students made frequent errors in.For example, Jung (2006) found that Korean university-level learners commonly made written errors in the uses of voice.However, the participants of this study scored the highest on recognizing the errors related to voice.Likewise, although several researchers who conducted an error analysis on the writing of Korean university-level learners called for an attention to errors regarding the wrong uses of articles (Chung & Lim, 2005;Lee, 2007;2008;Park, 2009), the results of the present study showed that participants had the second highest score for article use.In addition, they also achieved a relatively high score for prepositions which Zawahreh (2012) regarded as difficult for EFL learners to appropriately employ.The disparities in the results may be attributed to the different formats used to measure grammatical errors, such as marking correctness/incorrectness of grammar features in the provided reading passage or checking the use of grammar features in their English writing.Based on the results, it is noteworthy that it might be difficult for language learners to utilize these three categories correctly in their own writing, although it was possibly easy to notice the errors in an awareness test.This difference might be due to the familiarity that Korean students have with this particular task type.It would be interesting to examine how the learners from different contexts would perform using the same data collection instruments.
In order to determine if there are statistically significant differences between the high-level and low-level groups on grammatical features, an independent sample t-test was performed at a significance level of 0.05.As displayed in Table 2, the results revealed that there were significant differences between two groups for the grammatical errors regarding tense, article, and voice.
The means of the tense, article, and voice features in the high-level group were much higher than those in the low-level group (4.37 vs. 3.33 for tense, 4.47 vs. 3.87 for article, and 5.40 vs. 4.13 for voice), showing significant differences between two groups.Nevertheless, no significant differences were found for the categories of preposition, morphology, and vocabulary between the two groups.It is of note that the three categories that yielded significant differences between the two groups were congruent with those that EFL learners found difficult in previous studies (Choi, 2011;Chung & Lim, 2005;Jahangir, 2016;Jung, 2006;Lee, 2007;2008;Master, 1987;Park, 2009;Song & Park, 2001).These results imply that a tailored curriculum for grammar instruction is essential for the learners with different linguistic proficiencies.

Results regarding the learners' perception of grammar
The study results showed that, with the proficiency levels collapsed, the Korean university-level learners perceived the uses of preposition as easy with a mean of 3.70 and morphology as difficult with a mean of 3.18 (preposition > article > tense > vocabulary > voice > morphology).However, when considering these two groups separately, unlike the grammatical awareness test, the differences were revealed in the learners' perception of the use of the grammatical items with respect to the difficulty.In the case of the high-level group, they perceived that articles were easy to employ, but that voice was difficult to use (article > preposition > vocabulary > tense > morphology > voice).On the other hand, the low-level group showed different results.The lowlevel students felt that prepositions were the easiest feature to utilize, but articles were considered to be the most difficult (preposition > tense = voice > morphology = vocabulary > article).From these findings, it can be concluded that learners' viewpoints of grammatical perception in writing vary with linguistic proficiency level.Further research should be performed to investigate the specific reasons why the learners from different levels felt difficulty for different grammatical features.
In order to determine if there were significant differences between the two groups of students based on perception of grammatical difficulty in writing, independent sample t-tests were performed.The questionnaire results indicated that there were significant differences for the learners' perception of the difficulty of grammatical features regarding the categories of article and voice between the highlevel and the low-level group, as shown in Table 3.For the category of article, the high-level group perceived it as easier than did the low-level group, with a mean of 3.87 and 3.10.However, for the category of voice, the result was opposite in that the high-level group reported that it was more difficult to learn English voice than did the low-level group (2.87 and 3.57 for the high-level and the low-level group, respectively).
It is worth noting that the results from the highlevel learners' perception test toward grammatical features were different from the previous findings as well.Although a majority of the studies concerning error analysis of EFL learners' writing revealed that the learners made frequent errors in the use of articles, particularly for those learners whose first language does not include them (Choi, 2011;Chung & Lim, 2005;Lee, 2007;2008;Song & Park, 2001), the results of the perception test showed that the categories of morphology and voice were found to be the most difficult.The category of morphology included questions regarding possessive case, possessive pronouns, plural forms, and infinitives, which required the learners to possess particular grammar knowledge in order to correctly select the correctness or incorrectness depending on the context.
Also, to answer the questions concerning the category of voice, the learners needed to understand the trait of each verb (such as intransitive or transitive), as well as and the form of the past participle of each verb, as discussed in Lee (2006;2007).Also, as Berent (1985) and DeKeyser & Sokalski (1996) noted, it might be easy to comprehend but difficult to produce in writing for the advanced learners.It was presumed that the advanced learners experienced difficulty on account of the complexity of these grammar rules.Nonetheless, the low-level learners noted that articles seemed to be the most difficult to utilize in the perception test, although they achieved the second highest score in the awareness test.Korean university-level learners are unfamiliar with articles as they do not exist in their first language.Although Jahangir (2016) found that the use of articles was the least difficult among the three grammar features investigated by Pakistani EFL learners, for L2 English learners whose native language does not have articles at all, there is a distinct initial disadvantage in the rate of acquisition as noted by Master (1997).Hence, it might be difficult for Korean learners to learn and employ them correctly.
The results of the current study revealed a gap between learners' actual awareness and their perceived awareness in terms of certain grammatical features.The high-level learners scored high in the category of voice, whereas, interestingly, they perceived it as the most difficult.Also, low-level learners indicated the second highest score for the category of article, yet they perceived it as the most difficult to employ in writing.It is possible that disparities displayed in the results might be due to the questions presented on the test as mentioned previously.
In addition, the qualitative analysis showed that the learners from both groups expressed varying degrees of anxiety regarding writing in English.It was discovered that they did not feel confident in the uses of English grammar and sentence structures.Furthermore, the low-level learners reported that it was too difficult to memorize all the necessary vocabulary for writing; furthermore, they also reported that they cannot put the words in the correct order.The high-level learners also revealed difficulty concerning the usage of English grammar while writing.However, unlike the results of the learners' perception toward grammatical difficulty, some students found it difficult to correctly employ articles and prepositions.Also, they expressed that it is challenging to exactly convey their intended meaning in English.Overall, the results demonstrated that there were statistically significant differences in learners' perception toward the difficulty of the uses of article and voice between the high-level and the low-level groups.Also, it was found from the open-ended question that the high-level group expressed difficulty constructing exact sentences to deliver the meaning that they intended, while the low-level group showed troubles in creating English sentences due to the complex grammar rules.Additionally, both groups displayed anxiety in English writing in general.As seen in the results, it is regarded that the learners in EFL context possess varying degrees of anxiety or uncertainty in employing their grammar knowledge for constructing sentences to convey their intended meaning.Instruction should be delivered for the learners to practice using grammar rules in their writing so that they can relieve some of the anxiety associated with writing.

CONCLUSION
The current study investigated Korean universitylevel learners' awareness and perception of a few major grammatical features.Firstly, the results demonstrated that both the high and the low groups scored high for the category of voice, followed by the category of article.However, they received low scores for vocabulary and morphology.Also, they showed significant differences in the scores of tense, article, and voice between the high and the low groups.Secondly, the high-level learners perceived the category of voice as the most difficult, followed by morphology, whereas the low-level learners perceived the category of article as the most difficult.The two groups also displayed significant difference in the scores of article and voice.
The results of this study imply that it is necessary to provide grammar lessons tailored to learners of different linguistic proficiency levels on the basis of the gap discovered between the learners' actual level of grammatical awareness and their perception of grammatical items.Although the results should be generalized with some cautions due to the small number of the participants in this study, it appears that more instruction regarding vocabulary and morphology is called for, especially for Korean university-level learners in any proficiency levels.Furthermore, as the learners expressed difficulty toward the certain grammar category, instruction regarding voice and morphology should be implemented for the highlevel learners, and that lessons targeting the use of articles is necessary for low-level students in Korea.As Kim (2015) noted, metalinguistic feedback facilitated the learners better than direct corrective feedback in teaching articles, a general description of usage for each grammar category would help the learners enhance their understanding of them.Lastly, L2 teachers should seek to find ways to make a more comfortable learning environment particularly for Korean L2 learners in writing as the students showed anxiety in L2 writing regardless of their linguistic proficiency.

REMARK
This article was developed based on the study presented in TESOL Indonesia International Conference held in August, 2016.

Table 1 .
Six grammatical items used in the grammatical awareness test

Table 2 .
Statistical results of the grammatical awareness between the two groups

Table 3 .
Statistical results of grammatical perception between the two groups