An Error Analysis of Research Project Abstracts Written by Thai Undergraduate Students

Errors in writing are unavoidable while students are trying to develop their writing skills. There have been several studies on identifying writing problems or errors in students’ writing. It is believed that identifying students’ written tasks is an effective tool to explore the difficulties involved in learning language. This helps teachers’ awareness of the serious problems which occur in students’ writing and allows them to pay closer attention to their errors. The aim of the present research study is to pinpoint writing errors in English abstracts written by Thai undergraduate students. Forty abstracts of research projects were collected and analysed. The error analysis was conducted at the sentence level, word level, and mechanics aspect. The five most frequent error types ranking from the most frequent to least frequent were word choice, preposition, sentence construction, singular or plural forms and quotation marks. The findings of the present study have shed light on the students’ writing ability and give an insight into what the problems students face when writing their abstracts. Also, the errors found in the abstracts in the present study have pedagogical implications concerning English language learning, particularly with writing courses. The findings will be helpful for teachers to develop teaching materials to assist their students from committing errors when writing English abstracts and to improve academic writing skills.


INTRODUCTION
In has been observed that developing English writing skills is challenging for non-native English speakers and teaching such skills are difficult on part of the teachers (Heaton, 1990;Richards & Renandya, 2002). As a result, teachers have to pay careful attention to students' writing. This requires more time, energy, experience, and expertise in order to help students to gradually develop their writing skills.
Teaching writing is demanding because there are many different factors involved, including organization, grammatical structures, vocabulary and also appropriate methods of assessing students' progress. Specifically, the teaching of writing in English is a serious problem in teaching and learning in an EFL context. In L2 contexts, students' language knowledge is generally insufficient for developing students' English writing ability (Altinmakas & Bayyurt, 2019;Watcharapunyawong & Usaha, 2013). Similarly, some research studies have confirmed that writing in English is one of the major problems for non-native speakers. Even L2 students at a high standard of proficiency in English face difficulties in academic writing. Jogthong (2001) asserted that in countries where English is not the mother tongue, most learners struggle when writing academic texts even if Published by Australian International Academic Centre PTY.LTD. Copyright (c) the author(s). This is an open access article under CC BY license (https://creativecommons.org/licenses/by/4.0/) http://dx.doi.org/10.7575/aiac.alls.v.11n.4p.13 they can communicate well. These problems with writing are caused by the complex nature of the writing skill which needs considerable time and effort to solve. In fact, there are several factors affecting ESL/EFL students in writing such as attitudes, writing strategies, educational background, prior experiences, and linguistic knowledge. All these factors influence non-native learners when writing in English (Thongrin, 2000). Lee (2005) also supports the view that writing in a second or foreign language is difficult for many ESL/EFL students at all levels because of both learning styles and learners' attitudes. Littlewood (1997) emphasizes that the writing skill is an important skill, but that it frustrates most EFL learners. He also notes that to become proficient in writing, L2 learners need both information about composing skills and comprehensible access to specific types of linguistic knowledge. The grammatical structure of the L1 is also another factor affecting L2 writing. For example, Sukasam, Kantho, and Pennee, (2014) found that Thai students in high school level made a lot of errors in writing of which the past tense was the most serious problem for them. These researchers believe that the characteristics of the mother tongue were the causes of these problems. This is because in the Thai language there is no tense system indicates the time of an event as in the English language. The concept of tense and time is different from English language. That is, "there are no verbs or marker words to indicate time systematically in Thai language system" (Arakkitsakul, 2019, p. 261). This difference greatly affects learning target language and its acquisition.
There have been several studies on identifying writing problems or errors in students' writing. As we know, errors in writing are unavoidable while students are trying to develop their writing skills. In Thailand, these studies found that Thai students at all levels of education face writing problems (Chuenchaichon, 2015;Watcharapunyawong & Usaha, 2013). Grammatical problems are one of the main causes of errors found in the writing of Thai students. For example, Promsupa, Varasarin and Brudhiprabha (2017) reported that Thai students had problems in applying English grammar to their writing tasks at the word level such as the use of nouns, pronouns, prepositions, singular/plurals, and articles. Generally, the Thai and English languages are completely different at all levels such as pronunciation, words, grammar, punctuation, as well as cultural aspects (Thep-Ackrapong, 2015). These differences account for some of the difficulties students encounter in the writing of English.
As Gass and Selinker (2001) pointed out, errors can act as red flags or warning signals reflecting the learners' lack of knowledge of the target language. Therefore, in order to give productive feedback to students, teachers should be fully aware of their students' writing errors (Ferris, 2002). In writing, learners can easily make errors because they are using only their individual language knowledge in writing a text without the aid of any other sources. For these reasons, students' errors have been investigated with the aim of finding out the most serious problems encountered by learners. The use of error analysis is an effective approach to the investigation and of errors in students' texts. In other words, it is a study of the errors produced by second language learners by focusing on the process of identifying, analyzing, and categorizing the language produced by learners compared to the target language (Brown, 2000, Richard & Schmidt, 2002. As Ferris (2002) summarizes, by analyzing errors, the analysists could have a clear picture on how students produce their texts linguistically and see the progress of students' writing abilities. In short, it shows the evidence of students' language usage and indicates students' stages of language learning.
Error analysis is invaluable for the teaching and learning process. It assists teachers and course developers in recognizing the errors made by students, why they make such errors, and how these errors could be eliminated. Knowing the causes of errors allows teachers to directly solve those trouble-spots in the language learning of their students by preparing proper teaching materials and effective techniques to remedy students' errors (Darus & Subramaniam, 2009). For the learners, error analysis is evidence of the areas of difficulty in their writing (K1rkgöz, 2010;Seitova, 2016;Sompong, 2014). It is clear that the results from error analysis provide a factual information which will be very useful for creating suitable teaching materials to assist students in avoiding certain errors in the future.
Several research studies have been carried out to identify the errors in written texts regardless of genres and levels. These research studies (e.g. Darus & Subramanian, 2009;Erkaya, 2012;Watcharapunyawong & Usaha, 2013;Sermsook, Liamnimitr, & Pochakorn, 2017;Seitova, 2016;Khansir, 2016;Promsupa, Varasarin & Brudhiprabha, 2017) showed interesting results. For example, with a corpus of 72 essays, Darus and Subramaniam (2009) found that most errors made by Malaysian students were grammatical problems including singular/plural form, verb tenses, word choice, prepositions, subject-verb agreement and word order. Thus, students had serious problems in applying the grammar rules of English. Sermsook, Liamnimitr and Pochakorn (2017) analyzed written tasks in English produced by Thai university students, and found that punctuation was the major problem. Also, Klimova (2013) conducted an error analysis and she reported that Czech students struggled with word order, objectivity of abstracts, articles, subject-verb agreement, tenses, spelling, rhetorical structures, punctuation and capital letters when writing in English.
Olasehinde (2002) noted, it is inevitable that learners learning a target language will make errors. This is because becoming proficient in the target language is a long and gradual processes. Although there are some research studies which have analyzed students' errors, little research has been done to identify errors in undergraduate students' research abstracts, particularly those written by EFL learners. Therefore, the present study has been conducted to identify the errors found in students' research abstracts, firstly, with the first aim of examining the types of errors made by students and, secondly, to explore the frequency of their writing errors in a corpus of their research project abstracts.
With the ultimate goal of helping students to improve their academic writing, the researcher analyzed students' research abstracts to obtain in depth details of students' problems in writing research abstracts. This will lead to the improvement of teaching materials in the future, especially in the Thai EFL context. Broadly considered, it can be said that the findings of research studies on error analysis are useful and can be applied to pedagogy directly. This is because once the teachers know what the most common errors are, they can design their teaching materials to match to these problems appropriately (Sihombing, Nissa, & Estrelita (2015)).

Description of the Corpus
The corpus of the present study was 40 research project abstracts written in English by 4 th year students in "English for International Communication (EIC)" program. The abstract sections of their research projects were selected for the analysis in the present study. This is because abstract is important, and it presents a summary of the research project as whole. It captures the interest and attention of the reader and draws them in (Sharp, 2012). The research areas of students' projects were about language learning, including the four skills (listening, reading, speaking, writing), translation, vocabulary, and pronunciation. The number of the abstracts was quite limited since the program was a new program and only the availability and accessibility abstracts were included in the study. Each abstract contains approximately 250 words. The selected abstracts were written during the academic years 2015-2018. The abstracts were part of research project reports written by Thai students in fulfilment of the requirement of the Independent Studies course offered by Humanities Department, in the Faculty of Sciences and Liberal Arts, at Rajamangala University of Technology Isan, Thailand. The aim of the "Independent Studies" course was to encourage students to thoroughly investigate the interest topic using systematic research methods, as well as writing research paper and presenting the research findings. Each research project was carried out by two students, except for three projects which were conducted by only a single student. There was one supervisor for each research project and students had to present their work the examination committee for their comments or suggestions. Each research project followed the conventional style of a written research paper, which included an abstract, an introduction, a methodology, a results, a discussion, a conclusion, and also final sections with references, an appendix, and bio-data. However, only abstract section was used in the present study.

EIC Students' Background
English for International Communication (EIC) is a program provided by the English Department of the Faculty of Sciences and Liberal Arts, at Rajamangala University of Technology Isan, Nakhon Ratchasima. The focus group of the present study was 4 th EIC students, who got at least 500 scores of the TOEIC-like test. EIC students have to pass five English writing or writing-related courses in order to graduate, namely Paragraph Writing, Essay Writing, Argumentative Writing, Report Writing, and Independent Studies. All these courses emphasize the importance of writing skills and critical thinking. To pass these courses, students depend heavily on their performance in writing. They have to submit many written assignments. However, other related courses have already played a part in developing students writing skills in English such as Academic Reading, Critical Reading, and a Research Seminar. Based on a bunch of these courses taken, it was presumably that these students would have enough writing background. Yet, in spite of these courses, the students' writing ability is still far from satisfactory and they have many problems with their writing in English abilities, so they still require considerable help from their teachers. Most of the students commit errors in all aspects of language. As such identifying students' writing errors would be a direct and practical way to look closely into students' written abilities. Results reflected from the analysis would provide deep insights into students' writing problems. With the intention that helping the students to improve their writing skills is an urgent task not only for teachers but also for curriculum developers. Similar to other research studies, the present study attempts to help Thai EFL undergraduate students improve their writing skills by analyzing the problems they encounter in their writing tasks.

Data Analysis
After the project abstracts were collected, each of them was labeled as abstract number 1 to 40 to make it easier to access. Grounded on the previous research studies' findings (e.g. Darus & Subramaniam, 2009;Iamsiu, 2014;Seitova, 2016), certain errors were found frequently and be serious problems for EFL students. Therefore, the analysis classified errors into three main levels, namely sentence, word, and mechanics and each level of error was divided into sub-category as shown in Table 1. Each abstract was manually analyzed sentence by sentence in order to detect errors.
Based on the results of a preliminary study before conducting the main research, it was found that all these types of errors in Table 1 were found frequently. Also, in the literature, the same types of errors have been focused on (e.g. Darus & Subramaniam, 2009;Erkaya, 2012;Iamsiu, 2014;Seitova, 2016). The errors identified from these types could be a practical guide for the teaching of writing. The following procedure for the analysis was adopted. All the errors found were recorded and numbered. The record of error frequencies was separated into two groups. The first group was the overall frequency of occurrence. The second group was the frequency of occurrence in a particular abstract. To illustrate the second group, if a certain error was found several times in a single abstract, it was counted as having a frequency of one. The last step was to rank the frequencies of each error type from the highest to the lowest. Errors from each abstract were identified by the researcher. Also, the analysis was also verified by a Ph.D. candidate and reviewed by a native English speaker.

RESULTS
The frequencies of errors shown in Table 2 aim to answer the research objectives mentioned above. Five types of errors that were found frequently in the abstracts were word choice, prepositions, sentence problems, singular/plural forms, and quotation marks.

Five Most Frequent Errors and Examples of Errors
The frequencies of each type of errors are delineated in Table 2. Only the five most frequent error types are shown in detail. To make it easier to access the location of the error occurrence, each excerpt of all examples was labelled with the abstract number from which it was taken, and the areas of focus are underlined.

Word choice
Word choice in the present study referred to the lexical error, in which a wrong lexical item is used instead of the correct one. This made the sentence grammatical and semantic wrong. That is, the sentence sounds different, weird or does not make any sense (Ander &Yıldırım, 2010). Error of word choice were found in 55 places in 26 abstracts. Word choice was the most frequent error in the corpus showing that EIC students had problems in employing correct or suitable words as shown in the examples. Example 1 below shows the wrong use of word form and Example 2 shows an error of participial usage.

Example 1
The sampling was 33 Electrical Engineering students.

(Abstract 2)
In this Example, "sample" should have been used in place of "sampling". This indicates that the students lack appropriate knowledge of vocabulary forms. Also, it was found that students sometimes used a noun instead of a verb, for example, "analysis" vs "analyze". This reveals that they do not know how to use the correct parts of speech.
Example 2 The most frequent strategies employed was translation by using a loan word plus explanation (27.95%), following by …. (Abstract 5) The word "following" should be replaced by "followed". These students are unable to use the present and past participle correctly.

Prepositions
The use of prepositions was the second most frequent problem in the present study. There were 19 abstracts which contained this type of error (47.5 percent). In the present study, the prepositions which were most frequently used incorrectly in the abstracts included "in, on, with, of, over, for, from, and into" as demonstrated in Examples 3, 4, and 5.
Example 3 On overall, reading comprehension abilities of the stu-dents…. (Abstract 3) In this Example, the preposition "on" should not be used as "Overall" can be used to begin the sentence without any preposition.
Example 4 The results revealed that all of the students had the problems to …whereas making use of capitalization and punctuations was quite easy to them. (Abstract 2) In this excerpt, to make a sentence sounds naturally and correctly, the preposition "to" should be replaced by "for".
Example 5 The purposes of this study were to investigate….and to figure out the factors that contributed in English speaking problems. (Abstract 27) In this Example, "in" should be replaced by "to".

Sentence construction
Sentence construction in the present study refers to the kind of errors which usually impede understanding (i.e. incomplete sentence, confusing, ambiguous, unintelligible). In other words, this type of error focuses on sentence boundary problems covering both syntactic and semantic problems. The error on sentence construction was ranked third in Table 2. Fifteen abstracts (37.5 percent) contained this type of error. Most of the errors found were confusing or unintelligible sentences (Example 6) and sometimes it was an incomplete sentence (Example 7).

Example 6
As for finding the main ideas in each positions can be totally summarized that finding the main ideas at the beginning of the paragraph is the highest score (72.22%) and finding implied main ideas of the paragraph is the lowest score (35.55%). (Abstract 26) Example 6 was ungrammatical and confusing. To revise this excerpt, quotation marks should be used to separate the specific term used to classify the main idea. By doing so, it would be easier for readers to understand the excerpt. Moreover, the suffix "s" should be deleted from the noun "position". To avoid a run-on error, a punctuation mark like a comma is needed. Also, the past simple tense should be applied in this passage. This is because the function of this text was to present the findings of the research study. As suggested by Weissberg and Buker (1990), to announce the results of a study whether in the abstract or the main text, the simple past is preferable in nearly all fields. Example 7 Lastly, translation by a more general words (0.97%).

(Abstract 11)
In this example a complete sentence is not used, only a part of sentence. Therefore, to make it complete, a main verb should be used.

Singular/Plural forms
It was found that the misuse of singular/plural forms was another problem for the students when writing an abstract. As shown in Table 2, there were 20 incorrect uses of singular/plural forms in 14 abstracts. In addition to this problem, there was also failure to construct the correct forms of the singular or plural, which were classified into two groups. The first group was for the accidental mistakes, while the second group result of a lack of knowledge. Examples 8 and 9 demonstrates each of these problems.
Example 8 The sample of this study was 29 grade 5 students at…..

(Abstract 21)
From this excerpt, it can be seen that the authors of this abstract used subject verb agreement correctly. That is, "sample" was a singular noun in the sentence and the verb "was" was in agreement with it. However, this sentence was classified as an error because "sample" should be in the plural form since more than one sample was used in the study. This problem might simply be the result of carelessness. Presumably, if the authors had added the suffix "s" to the word "sample", they would have used "were" for the main verb. By reviewing the other sentences in the abstract to confirm the authors' usage of pluralization, it was found that all the sentences were used correctly.
Example 9 The purposes of this study were: 1) to find the ….; 2) to analyze each strategies used in…(Abstract 17) The noun "strategies" used in the excerpt should be in the singular form. This might be the result of the students' use of "each", which requires a singular noun. On further investigation, it was found that these authors keep committed errors systematically in their use of "each". Table 2 shows that the errors on quotation marks occurred 102 times in 11 abstracts. According to Oshima & Hogue (2006), quotation marks are used in three ways: "to enclose direct quotations, to enclose unusual words, and to enclose titles of short works" (p. 288). The problems found in the present study were that the EIC students omitted to use quotation marks when they referred to unusual words/expressions (Example 10) and titles of works (Example 11). These problems occurred 102 times in the abstracts which were from those studies that mentioned the strategies used for translation. Therefore, the terms for translation strategies were mentioned several times, but students omitted to place them in quotation marks. This resulted in a high level of errors.

Example 10
The most frequently technique employed was generic to specific translation (17.83%), followed by over translation (15.28%),….(Abstract 24) Example 11 This research was an analysis of the translation strategies of nonequivalence at word level in the story of Winniethe-Pooh written by …. (Abstract 25) In the analysis, the errors on the use of quotation marks were found most frequently when students referred to the results of the studies. They presented information both in texts and number forms. However, the EIC students did not highlight the technical terms or definitions of key terms used in their studies by placing them in quotation marks. A lack of quotation marks makes the abstracts difficult to read and sometimes the reader may be confused.
The findings reveal the problems of EIC students in their writing of English. Most of them have problems with using the correct words, followed by using the correct prepositions. This led to the problem on sentence structure, which ranked number three. It showed that 37.5 percent of the abstracts contained sentence construction errors. When students used the wrong words, the meaning of their sentences was confusing or unintelligible. It can be seen from this that the major problem was at the word level (word choice and prepositions) which then affected the sentence level. The mechanics part was less problematic in the EIC students' abstracts. The results of this case study should encourage teachers to pay closer attention to the problems of students writing in English.

DISCUSSION
The findings show that the most frequent errors were at word level with word choice ranked number one in the list. This problem comprised 65 percent in the corpus, which is relatively high in comparison with previous studies (Darus & Subramaniam, 2009;Erkaya, 2012;Klimova, 201;Watcharapunyawong & Usaha, 2013;Seitova, 2016). For example, the percentage of word choice appearing in Erkaya (2012), Klimova (2013), and Seitova (2016) showed 45%, 35%, and 13.7% respectively. Other research studies (Darus & Subramaniam, 2009;Erkaya, 2012;Watcharapunyawong & Usaha, 2013;Seitova, 2016) also found that errors on word choice occurred with a high frequency. As shown in Erkaya's (2012) study, word choice was a serious problem in 17 English essays produced by Turkish students. This caused comprehension difficulties for their readers. Erkaya (2012) believed that the lack of adequate vocabulary was the basic cause of the problem. Watcharapunyawong and Usaha (2013) and Klimova (2013) explained the problem with word choice usage derived from mother tongue interference which was the key factor affecting the errors in students' writing. This is supported by Klimova (2013) stated that the word choice problem in students' writing resulted from the mother tongue (Czech) which is more flexible than English, and this affected their writing in English to a certain extent. This explanation might also apply to the present study since there is no verbal form in the Thai language. Thus, it seems likely that the Thai language is one of the causes of the errors found in the abstracts.
In their study (Darus & Subramaniam, 2009), word choice was ranked number 3 in 15 types of errors, accounting for 10.5 percent of the total. They noted that word choice was a problem which needed to be remedied for Malaysian students. Word choice was an error which occurred the most frequent in a research article corpus in a study conducted by Salehi and Bahrami (2018). It accounted for 36.2 percent of the total number of errors in their study. They believed that the reasons behind the relatively high figure were the transference of L1 and the similarities or differences between the first language, which was Persian and the English language. The limited vocabulary of the students in this study was also another main cause of this problem. This is in line with the results of the present study. For example, the word "respectively" was used incorrectly, appearing in 13 abstracts (see Example 12 below). That means that 50 percent of the abstracts contained a word choice error. Such a large number implies a serious grammatical problem for the students which derives from their limited knowledge of English vocabulary. Also, the cultural and linguistic differences to some extent might influence to the errors made by these students. Therefore, they need to be taught how to use such words correctly.
The example shows that students do not know how to use the adverb "respectively". In fact, according to definition in a dictionary, this adverb is used when two or more items with each relating to something previously mentioned are used in the same order as first mentioned. However, in Example 12, all the techniques were listed with percentage use given in brackets immediately afterwards. Therefore, the word "respectively" should not be used.

Example 12
The most frequently technique employed was generic to specific translation (17.83%), followed by over translation (15.28%), synonyms translation (12.10%), euphemism (11.46%), doublet translation (8.91%), simile (7.00%), under translation (7.00%), mistranslation (7.00%), loan word (4.45%), hyperbole (3.18%), cultural substitute translation (2.54%), idiomatic translation (1.91%), loan word with classifier (1.27%), respectively. (Abstract 24) The use of the word "following" was another word choice error found frequently in the corpus. This word should be replaced by "followed" as shown in Example 13. The same incorrect use of this particle occurred in 6 abstracts, accounting for a total of 15 percent. This can be inferred form this that students do not understand how to construct a complex sentence by using a reducing adjective clause. This shows clearly that students need more guidance on sentence construction to enable them to form appropriate verbal structure.

Example 13
The study revealed that the most frequently used strategies are translation by omission (33.86%), following by … (Abstract 16) The incorrect use of prepositions was another problem that occurred frequently in the abstracts. Similarly, Promsupa, et al. (2017) also noted that prepositions were found to be a serious problem in their research study. Compared to the findings of Al-Khasawneh's (2014) study, the error of using prepositions incorrectly was ranked second in the list, showing a total of 12.4 percent in the whole corpus. In the present study, preposition errors were strikingly different in a number of occurrences, accounting for 40 percent in all, although the students were given plenty of time to writing their abstracts. This is in contrast to Al-khasawneh's (2014) study in which students were only allowed a limited time to write. Thus, it is clear that the students in the present study lack knowledge about the use of prepositions. The forms of the prepositions that students had problems with included on, of, in, into, and with.
The third in the rank of errors committed by students occurred in sentence construction as exemplified in Example 14. Similarly, Iamsiu's (2014) encountered the same problem. There were 102 occurrences or 36.17 percent which revealed this type of error. She believed that this error was the result of mother tongue interference. As in Iamsiu's (2014) study, the students who wrote the abstracts were native Thais. They had studied English as a foreign language which meant that their exposure to the English language depended heavily on classroom instruction. They needed to put more time and effort into the improvement of their English language skills, especially writing. With a limited knowledge of vocabulary and little writing experience, it was likely that their mother tongue interfered with their compositions.

Example 14
The finding of the study showed that an analysis of the translation strategies of non-equivalence at word level in the children's novel Farmer By, were tallied and calculated in term of percentage (P = n/N× 100) there were seven translation strategies of Mona Baker's (1992) are applied in translating the children 's novel,"Farmer Boy". (Abstract 23) This sentence is difficult to interpret. The sentence is a run-on sentence. A period or comma should be used to separate sentence into two parts.
The error in the use of the definite article "the" is also found in the present study, accounting for 20% as exemplified in Examples 15 and 16.
Example 15 The 79  The problem of using articles was also found in previous studies (Al-Khasawneh, 2014;Watcharapunyawong & Usaha, 2013;Sermsook, et al., 2017;Alhaisoni, Gaudel, & Al-Zuoud, 2017). The findings of these studies showed a high degree of incorrect usage of the article. It was ranked number 1 in two studies (Al-Khasawneh, 2014;Watcharapunyawong & Usaha, 2013). The authors asserted that mother tongue interference was the key factor affecting this problem. However, the articles "a" and "an" were not a serious problem for students in the present study, unlike Sermsook et al. (2017), who found that both definite and indefinite articles occurred frequently in their corpus. For example, students omitted to use "a" before a singular and countable noun and also they found the use of "the" in the expression "I go to the bed". The authors confirmed that this error was strongly influenced by the students' first language which was Thai. This is due to the fact that there is no article in the Thai language. Therefore, it is not surprising when students get confused when they have to use them in English. Because of these problems, article usage should be an urgent issue for Thai students to consider and remedy for. In particular, the English major students or EIC students in the present study should have been trained intensively how to use articles correctly.
The results on subject-verb agreement were strikingly different from those of previous research studies (Al-Khasawneh, 2014; Darus & Subramaniam, 2009;Seitova, 2016). These studies revealed a comparatively high degree errors in subject-verb agreement. In the present study, it was ranked as the lowest. Thus, although the Thai language has no subject-verb agreement, this pattern in English grammar was not a serious problem for the EIC students. They demonstrated that they were aware of how to apply this grammar rule. We can infer from this that the Thai language did not interfere with the EIC students' use of subject-verb agreement in the English language.

PEDAGOGICAL BENEFITS
With the aims of examining error types and their frequencies, the most five frequent errors types were 'word choice', 'prepositions', 'sentence construction', 'singular/plural forms', and 'quotation marks'. The findings have pedagogical benefits and reflect the teaching and learning processes. If errors committed by L2 learners are identified systematically, this will show a clear picture the process by which student learn and acquire the language (Corder, 1981). The findings from the present study reflect the students' writing problems, which serve as useful guidelines for developing English language teaching materials to assisting EIC or EFL students to write more correctly in English. It is essential to make students aware of their errors so that they will keep them in mind and avoid committing those same errors again when writing in the future. They should learn to use their own strategies to write without committing those errors that have been identified. English language learning in EFL contexts will never be fully achieved if students have to depend only on classroom instruction.
Students' written tasks which show grammatical errors or mechanics problems, confusing structures or styles are viewed as poor writing. That is, errors in writing reflect both the quality of the writing and the author (Johnson, Wilson, & Roscoe, 2017). Therefore, students should be aware that using appropriate and correct words is an essential basis for success in advanced academic writing. As noted by Henry & Roseberry (2007), for skillful language learners, lexical knowledge plays a more significant role than grammar in reaching native-like fluency. Thus, word choice should be especially emphasized in any course of second or foreign language teaching.

SUGGESTIONS FOR FURTHER RESEARCH
Although the present study provides valuable information concerning the errors in students' research project abstracts, some issues relevant to error analysis are also important and need to be researched in the future. Other aspects of writing processes such as rechecking, editing, revising drafts and checking sources that students have consulted while writing are also of interest for future research studies. Furthermore, a long-term study of students' improvement after their errors have been identified and pointed out to them would be useful topic for research in order to follow students' subsequent progress in their writing. Lastly, it is suggested that with a larger number of abstracts, results might have been different and more generalizable.