The effect of writing exercises in classroom on the production of written sentences at undergraduate level by Saudi EFL learners: A case study of error analysis

Abstract This study aims to analyze the types, and frequency of errors in written sentences by undergraduate Saudi English as a Foreign Language (EFL) learners, following a quasi-experimental design. Data were collected from two groups: the control group (CG) and the experimental group (EG) through written sentences. The control group (CG) did not receive the treatment whereas the experimental group (EG) did to show the effect of writing exercises in the classroom on the production of written sentences. The experimental group (EG) received extra 3 hours of teaching and writing practice in a week to show the effect of classroom writing on the production of written sentences. The researcher assigned extra writing exercises in the classroom for the experimental group (EG) from the coursebook (Unlock 1, and 2) for 2 months. The age of the participants ranged from 18 to 25. The findings revealed that both groups committed overall errors with a significantly different frequency such as the control group with 199, whereas the experimental group with 116. Moreover, the control group committed errors in 9 categories out of 11 like “wrong word order”, “wrong use of preposition”, “wrong use of article”, “wrong com-adjectives”, “wrong subject verb-tense”, “faulty capitalization”, “punctuations”, “spelling errors”, and “wrong word choice”. In contrast, the experimental group produced errors only in 7 categories out of 11, and these are, “wrong use of preposition”, “wrong use of article”, “wrong subject verb-tense”, “faulty capitalization”, “punctuations”, “spelling errors”, and “wrong word choice”. The study has important implications for the teachers to modify their instructional strategies and focus on writing exercises in the classroom, and course designers to add more writing course content consisting of writing materials to provide a wide range of opportunities to the EFL learners and minimize their errors in writing skills.


Introduction
Writing is one of the complex tasks in any language due to its multifaceted aspects in conveying the intended message, especially in the context of Second Language Acquisition or Foreign Language Acquisition (SLA/FLA). Writing skills require a balanced approach in terms of grammar, choice of vocabulary, coherence, and clarity. In a non-native academic context, students are frequently tested in writing tasks by measuring their language performance level.
Writing in a non-native context is more challenging for students because this skill requires accurate grammatical and syntactic skills, appropriate semantic, and cultural suitability of the language. English writing is more challenging for Saudi students due to the different orthographic patterns of Arabic and English. It is generally assumed that Arabic students encounter numerous challenges in writing the English language due to the negative influence and transfer of L1 (Arabic) to that of English. Henceforth, they commit errors in writing pertaining to many aspects in terms of FLA.
Error Analysis (EA) has received a great deal of attention from scholars throughout the world to facilitate Second Language Acquisition/Learning. Dulay et al. (1982) argue that error analysis is an approach to analyze the errors committed by ESL/EFL students and facilitate Second Language Learning (L2). They assert that error analysis not only helps facilitate the L2 acquisition but assists the learners to learn the target language, as well as improve the teaching pedagogies on part of the teachers to promote their teaching strategies.
Error Analysis (EA) is defined by different scholars in different words. According to Crystal's (1999) definition, Error Analysis (EA) consists of language analysis that depicts the deviation from the standard patterns. Error Analysis (EA) assists language learners, teachers, and researchers to discern the level of L2 knowledge acquired by the students, revealing the major difficulties they encounter bringing a change in teachers' teaching strategies and finally, opening venues for researchers to fill the gap.
The current study aims to identify, categorize and analyze the types and frequencies of errors committed by undergraduate Saudi EFL students. The study aims to achieve three objectives. In phase one, the study focuses on categorizing the types and frequency of errors in written sentences produced by EFL students in Saudi Arabia. In phase two, the research aims to make a comparison in terms of errors produced by the L2 students-between the (CG) and (EG). Finally, the study will shed light on the errors and give possible reasons for their occurrence.
Thus, the study aims to answer the following research questions: (1) What are the types and frequencies of errors in written sentences produced by Saudi undergraduate students?
(2) What are the similarities and differences of errors' frequency between the control group (CG) and the experimental groups (EG))?

Review of literature
Before the emergence of Contrastive Analysis Hypothesis (CAH), errors in writing were considered failures on part of the L2 learners and immediate actions were required to eradicate them, but CAH explained the phenomenon that L1 and L2 have a positive or negative influence on each other. If they are closer in structure, so the influence is positive, but if the L1 and L2 have a greater contrast so the influence is to be negative. In this way, CAH focused on the teaching strategies and course materials to highlight the similarities and differences between the two languages (Fisiak, 1981). According to the emergence of "Generative Theory" in the field of SLA, errors are not to be eradicated immediately, as they indicate the progressive and developmental aspects of learning and provide essential feedback to the learners and teachers to modify the target structure (Lightbown & Spada, 2006;Keshavarz, 2015;& Ellis & Barkhuizen, 2005).
Writing is a very complex productive skill that requires cognitive processes and thinking which goes through different stages, for example, pre-writing, writing, and post writing. Valuable and successful writing very much requires many overlapping aspects such as punctuation, spelling, tense, organization, clarity, coherence and cohesion, capitalization, and content. Corder (1967) was the first to introduce the concept of errors by Foreign Language Learners in their initial stages of writing. Mistakes were regarded as failures in performance but errors are considered as failures of competence (Corder, 1982). Errors require advanced skills and knowledge and cannot be easily corrected by inexperienced students (Cetereisi & Bostanci, 2018;Keshavarz, 2015). To differentiate between an error and a mistake, it is advisable to understand their frequency and even go deeper to trace the possible reasons for them (Keshavarz, 2015). Corder (1982) gives five phases in order to better examine and make Error Analyses (EA) in the production of written sentences. First, obtain a language sample from L2 learners; next, identify the types of errors; then, categorize; explain, analyze; and finally, evaluate them.

Categorization of errors
Errors are categorized based on their distinguishing features in the field of L2 acquisition. Dulay et al. (1982) distinguish errors into six categories: disordering, alternative use of form, use of irregular rules, double marking of semantic features, omission of morphemes, and use of wrong word choices.
Errors are categorized into two main types based on their causes: interlingual; and intralingual (Brown, 2007). Interlingual errors are L1 based transfer errors such as morphological, grammatical and lexical transfer to the target language. While intralingual errors are those either caused by the insufficient knowledge of the L2 learners or adverse effects of the target language on the L2 users. Intralingual errors mainly consist of six types: wrong categorization; hypercorrection; hyperextension; wrong analogy; lack of rules; and overgeneralization (Keshavarz, 2015). The wrong categorization in learning the target language refers to the instances when the L2 learners apply L1 rules to the target language. Hypercorrection in L2 learning involves selfcorrection by the learner implying the words and phrases beyond the conventional style of the target language. Hyperextension involves the use of lexical items beyond the conventional structure of the target language. Wrong analogy consists of applying inappropriate structures such as verb and preposition in the target language. Lack of rules refers to insufficient knowledge concerning the structure of the target language. Overgeneralization consists of using a language structure that deviates from the conventional mechanism of the target language (Terzioğlu & Bensen Bostanci, 2020). Keshavarz (2012) differentiates errors into two categories: global and local. Global errors consist of wrong word order, pronouns, and prepositions. This category should be taken care of as it could cause comprehensibility issues. Local errors refer to misspelling, wrong order of prepositions and other minor issues in written Khansir and Ilkhani (2016) mentioned two main purposes of errors in written sentences by EFL learners at their early stages of learning. First, errors in writing at the initial stages of learning as an EFL are regarded as a natural process; by making errors, the students integrate learning processes; making errors in writing indicates an inevitable system in learning EFL. Secondly, L1 (mother tongue), (Hereafter L1) is not the only source of interference. Learners often make errors in writing that needs to be examined and observed in order to know the thinking process of the language learners. Errors are inevitable features of the learning gap on the part of the learner's language acquisition and provide insights into the teachers to take remedial measures to improve teaching material and teaching strategies (Khan & Sohail, 2021;Terzioğlu & Bensen Bostanci, 2020;& Jabeen et al., 2015).
Empirical studies over the decades have shown us how to better train students in the writing skills they need to be effective writers. Khalil (1989) argues that the Arab students have underused grammatical and lexical cohesive devices while they have overused the reiteration of the same lexical item as a cohesive device. Salehi et al. (2018) assert that L2 writers produce the highest frequency of errors in the category of "wrong word order". Fengjie et al. (2014) found out that the L2 learners' writing errors in terms of sentence effectiveness and mainly coherence suffer from six main problems. The inability of the students to separate words that are closely related until and unless necessary, with ambiguous references in the use of a pronoun, to put a modifier far from the word which it modifies or using dangling modifiers. Other mistakes were making unnecessary changes in tense, voice or mood of verb, making an unnecessary shift in numbers or persons and lastly, the use of different forms to express parallel ideas. The innumerable errors committed at the lexical and syntactic level negatively affected the overall writing coherence. Non-native and native students' were sometimes indistinguishable as both of them tended to suffer from the loss of cohesion in their writing works. Khansir (2012) reported on syntactic errors such as passive voice, tenses, indirect form, auxiliary verb, tag questions, prepositions, and relative pronouns in the writing of Indian and Iranian undergraduate university EFL learners. For this purpose, data were collected from two hundred students from two universities, namely Mysore University in India and Bushehr Islamic Azad University in Iran. EFL learners in India and Iran participated in this study. Indian students committed 3736 errors and their Iranian counterparts committed 2841 errors in their written assignments. The study investigated five types of errors like "punctuation", "spelling", "paragraph", "articles", and "conjunction". The findings revealed that the lack of classroom writing practices and L2 learning techniques were the main causes for making these errors by L2 learners. Aydin (2010) researched errors in writing assignments by undergraduate students at Balikesir University in Turkey. The data were collected from 39 ESL/EFL teachers. The findings demonstrated that explicit treatment of teaching writing rules to undergraduate classes in the first year of their study helped students in integrating content knowledge, proficiency skills, language form, and organization. The study revealed that practising writing assignments helped students in learning grammar and improved their vocabulary. Though the study indicated new findings with an effect on the effective filters that created the feelings of tardiness because writing is boring, and it takes time to complete each written assignment. Qinghua (2010) researched two sophomore EFL undergraduate classes in China. The students were aged 18 to 21 and each class consisted of 34 EFL learners. The researcher investigated a comparative study to demonstrate the difference between a Non-Portfolios Based Writing Assessment (NPBWA) and the experimental based class of Portfolios Based Writing Assessment (PBWA). The study consisted of writing assignments focusing on accuracy (Grammar), Complexity (content) and coherence (organization). The findings demonstrated that (PBWA) in the context of EFL learning helps the learners in some essential domains, especially coherence. The researcher suggested that (PBWA) is an effective strategy in ESL/EFL, particularly in the context of non-native speakers. This improves learning and consequently enhances the writing skills of undergraduate students.
The most comprehensive framework in analyzing the cohesive errors in the students' writing is the model of Halliday and Hasan (1976) that identified five types of cohesion: ellipses, reference, substitution, conjunction, and lexical choices. To write more cohesively, the students need cognitive competence and higher linguistic skills that may help them to write comprehensible texts that communicate certain ideas and information to others.
Al-Tamimi (2018) researched students' perceptions about errors in academic writing at Hadhramut University in Yemen. Data were collected from 60 participants based on their perceptions about the type of errors that the participants frequently made in academic writing. Data were also collected from four Department of English faculty members to demonstrate the frequent errors in writing assignments by the students at the undergraduate level. Data were triangulated through semi-structured interviews and questionnaires. The findings demonstrated that students in their first year at the undergraduate level made many errors in their writing assignments. Besides sentence structure, they made major errors in the expression of ideas, vocabulary, prepositions, punctuation, spelling, irregular verbs and use of articles. The findings revealed that these types of errors in academic writing occurred with different frequencies.
Pablo and Lasaten (2018) investigated qualitative research on the difficulties of undergraduate students learning writing academic essays in English as a Second Language. Data collected from Grade 11 students consisted of 227 essays including both public and private Senior High Schools in the Philippines. The researcher analyzed the data following the criteria suggested by Jacobs's (1981) Standardized Rubric in Evaluating Academic Essays elaborated by (Valdez, 2016). The findings demonstrated that English as a Foreign Language learners have many difficulties such as content (lack of variation of ideas), organization (lack of connectives), poor vocabulary and word choices, poor structure with weak language and grammar, lack of in-text citation about references and use of the first-person pronoun. The researchers suggested that the overall quality of their writing essays range from poor to fair. Moreover, the findings revealed that the writing of the students in private institutes is better than that of those in public in terms of quality. Abdulkareem (2013) investigated the problems in academic writing faced by Arab EFL learners. Data were collected from 85 students at the University of Technology Malaysia (UMT). The students who participated in the study were from different countries such as Saudi Arabia, Yemen, Iraq, Libya, Algeria, Sudan, Syria, and Palestine. The Snowballing technique was used for the selection of participants and data were collected through writing tasks and questionnaires. The participants were asked to write two paragraphs about their field of study. The assignments were checked and corrected by EFL teachers to find the errors committed by students. The results revealed that the errors committed by students substantially consisted of the choice of vocabulary and sentence structure. The number of error related ideas committed by all the students was few, regardless of nationality. The study made some recommendations for avoiding errors in written sentences by students of English as a Foreign Language based on these findings.
Al-Khairy (2013) investigated the types of errors in writing committed by Saudi undergraduate students at Taif University. The analysis of the data was carried out through standard deviation and means by descriptive statistics. The findings demonstrated that students at the undergraduate level in universities made major errors in their academic writing. The study revealed that students made some blunders at the sentence level in their writing assignments. The study made some recommendations regarding the improvement of the English language skills of the students in general and particularly in academic writing skills.
This study focuses on investigating the types, frequency, and reasons for the errors in written sentences produced by undergraduate university EFL learners in Saudi Arabia. The practice of writing portfolios in the field of Second Language Acquisition has been used over the years in order to reveal how the writing skills develop in response to the students' learning in the EFL context. Researchers have done several studies that focused primarily on reasons impacting the writing of non-native Saudi EFL undergraduate learners. For example, Al-Khasawneh and Huwari (2013) examined the sources and reasons of errors by non-native Saudi EFL learners and demonstrated that Saudi Students have less practice of writing assignments which makes them deficient in producing a good piece of error-free writing. Moreover, the research demonstrated that almost all the participants revealed that they do not practice writing at all or have limited exposure to the writing exercises in the classroom, which causes extensive errors in written sentences. The findings indicated that the educational system has a strong effect on the writing of the students. The curriculum focuses mostly on speaking, reading, and listening, but the less frequent aspect of writing exercises and skills makes it a challenge for EFL learners even to compose a simple paragraph. Institutions have a role in producing good writers but the teachers also play a crucial role in adapting strategies to motivate the students to improve their writing skills.
The previously available research conducted on Error Analyses (EA) in the EFL context revealed that the major types of errors committed by EFL learners consisted of wrong use of "tense", "preposition" and "articles" (Terzioğlu & Bensen Bostanci, 2020;Atmaca, 2016;Abushihab, 2014;Zheng & Park, 2013;& Kırkgöz, 2010). The most frequent and prevalent interlingual errors in the L2 context consist of "grammatical", "prepositional" and "lexical" errors (Kırkgöz, 2010). The practice of written assignments in classrooms, motivate L2 learners, reduce the frequencies of errors and give opportunities to the teachers to provide timely feedback (Pawlak, 2014;& Bensen, 2014). L2 learners of the Arabic language are ignorant of rules restrictions and consequently, overgeneralizations cause intralingual errors (Tahaineh, 2010). The inappropriate use of definite articles and their omission reflect the incompetency and inappropriate knowledge of the articles in the target language by the EFL learners (Elumalai, 2019). The poor training skills of the teachers, their lack of competencies in the target language, though they are native speakers consequently lead the learners to make excessive errors in the structure of the target language (IvyPanda, 2020).
Based on the empirical findings of the literature reviewed, students at the undergraduate level in their first year make unforced errors in written sentences At the sentence and paragraph levels, the majority of syntactic errors committed by students are fragmentation, spelling, verb tense, subject-verb agreement, conjunction, and gerund. The findings of the previous studies revealed that classroom writing practices and teachers' teaching strategies provide extensive training to students to reduce these errors and the students are assisted to identify and correct their own errors through the writing assignments given to them.

Research design
The current study follows a quasi-experimental design to show the effect of writing practices in the classroom on the production of written sentences of EFL Saudi students. In view of the research questions under investigation, the quasi-experimental design does not violate the validity and reliability of the study. Experimental design helps the researcher to show the cause and effect relationship between the control group and the experimental groups (Creswell, 2012;& Cook & Campbell, 1979). The study follows content analysis procedures to examine the errors committed by the students in writing. Therefore, qualitative description will complement to portray a broader picture in order to reveal the types of errors committed by the students and will point out the major difficulties that the students face in their writing. Quantitative analysis in terms of frequency measurement means score, frequency score and percentile distribution through tables to demonstrate an overall picture of the findings in the study.

Participants and sampling
A convenience sampling procedure was adopted. The samples in the study are not randomized, but rather follow a technique consisting of the control group (CG) and experimental group (EG). For a firsthand observer, action research is the best and systematic tool to dynamically plan, observe, analyze and gain insight into the practices for developmental purposes, Ary et al. (2010). The data were collected from 60 Saudi EFL male students at the English Language Centre at Taif University in 2019, Saudi Arabia. Every group consisted of 30 students, that is (CG) and (EG) to create homogeneity.
The learners studied the English language for almost 10 to 12 years as non-major courses at different institutions across the country. The age of the participants ranged from 18 to 25. The data were collected from two classes and they were from the Engineering and Medical sciences. These classes named "Intensive English for Academic purpose" were chosen based on the reason that they are considered proficient users and have a higher proficiency in the target language. They study English as a compulsory subject for 12 hours a week. The course material used for these classes consisted of e-books Unlock 1, and 2. The course materials in these books are designed by the Cambridge University Press based on the Common European Framework of Reference for Languages (CEFR) B1-B2 level.
The researcher adopted convenient sampling, a non-randomized sampling procedure for the collection of data from participants. To obtain written permission and convey the objectives of the study, he first approached about 120 participants who were enrolled in "Intensive English for Academic purpose" in different sections. A total of 66 participants agreed to participate including both the control group (CG) and experimental group (EG) in the study. However, 6 participants withdrew from the study during the intervention, and their data is not included in the study. A total of 60 participants willingly completed the study from both the control group (CG) and experimental group (EG) and their findings are given in the analysis part of this study.

Ethical consideration
Prior permission was obtained from the Research Committee of the English Language Centre (ELC) at Taif University in writing. The students were informed about the aims and objectives of the study before the collection of data and it was stated to them that they could withdraw any time from the study if they wished to. Moreover, they were informed that their identity would remain hidden from the public and the data would be solely used for research purposes and would be destroyed after the study is completed.

Instruments
The instrumentation in the current study is Document Analysis/Analytical descriptive approach (Paragraph writing). A topic was given to the participants to produce homogeneity in their approach. All the participants were asked to write a paragraph about their country or city. Moreover, they were asked to write at least 10 sentences, use the appropriate verb tenses, linking words and punctuation marks. They were all provided with the following keywords/phrases to create homogeneity in their writing. The worst traffic, have problems with flooding, popular, green areas, the most polluted, have problems if sea level goes up, tourist spot, expensive place, far from the workstation.

Procedures
Both the groups used to attend three hour English classes (180 minutes) in a day and four days a week as an intensive English for Academic purposes (IEAP) following the course material for L2-e-books Unlock 1, and 2. These books include teaching material on listening and speaking, reading, and writing. The researcher taught these sections for 12 hours each week (A and B). The control group (CG) did not take treatment, and on the other hand, the experimental group (EG) received extra 3 hours of teaching and writing practice in a week to show the effect of classroom writing on the production of written sentences. The researcher assigned extra writing exercises in the classroom for the experimental group (EG) from the coursebook (Unlock 1, and 2) for 2 months. This was done to investigate the effect on the written sentences as assumed earlier that writing practices in the classroom will diminish errors in written sentences (Bensen, 2014;Khansir, 2012;Qinghua, 2010;& Aydin, 2010).

Data collection and analysis
The researcher collected all the 60 written responses from the participants (A and B) on the same day. The researcher gave the participants 30 minutes to complete the assigned work and add their bio-data to the form for research purposes only. After 30 minutes, the written responses were collected from both groups; the control group (CG) and experimental group (EG), with the help of a co-teacher from the English Language Center. Corder's (1982) (EA) categorizations technique was employed to analyze and interpret the data. First, data were collected from the EFL learners. Secondly, data were categorized according to the taxonomy presented by (Keshavarz's, 2015) into four main categories. Finally, the data were transcribed, presented in the tables and the possible causes of the errors were explained.
A sampling technique was employed to assign the errors into four taxonomies based on the framework presented by (Keshavarz's, 2015). The first taxonomy consisted of "syntactic category", which includes, wrong word order, preposition, and article. The second category refers to "morphological taxonomy" that includes faulty comparative adjectives, wrong use of possessives', omission of plurals, and wrong use of the subject-verb agreement. The third taxonomy in the study employed was "orthographic" that includes, faulty capitalization, punctuations, and spelling errors. Finally, the category of "lexical taxonomy" includes the wrong word choice (Terzioğlu & Bensen Bostanci, 2020).
The data were carefully analyzed and coded in a separate sheet according to the types and frequencies of errors by three experienced English Language teachers. The coded data were further shown/displayed in the tables with standard frequency, mean scores and percentages, followed by descriptive analysis and discussion to enrich the findings.
The statistical analyses were carried out using NCSS software version 30 to demonstrate the t-tests, analysis-of-variance between the (CG) and (EG). A paired sample t-test was employed to show whether there was a significant difference between the errors committed by two groups-(CG) and (EG). The statistical analysis demonstrated the mean frequency, standard deviation, and standard error for variance analysis between the two groups.
In view of ethical consideration, the identity of the participants was kept concealed and the data were presented through codes. The data were coded based on the following 11 symbols assigned to the worksheet, that is, WO-Word Order; Prep-preposition error; Art-article; Adj-Wrong comadjectives; Prep-preposition error; OPls-Omission of plurals; SVTense-Wrong subject verb-tense; FCfaulty capitalization; Pun-punctuation; Sp-spelling error; and Wf-wrong word choice.

Findings and discussion
This section discusses the types and frequency of errors identified in the data. In addition, in this phase of the study, the findings are presented in the table to show the differences and similarities between the control group (CG) and experimental group (EG) in light of the identified categories. Table 1 illustrates the respective taxonomies of error along with frequency and percentages in both control (CG) and experimental groups. It shows the precise numeric values of each taxonomy and its overall distribution. Figure 1 is the visual interpretation of all of the four taxonomies of errors representing both the control and experimental group. Table 2 gives a detailed description of statistical data in both the groups separately and shows the respective mean and the percentage difference of each group.

Paired 1 Variable(CG)
Paired 2 Variable(EG) Paired Difference(CG)-(EG) Table 3 is the overall representation of both the control and experimental groups along with its T-test values and gives us important information about the upper and lower limits of the mean difference. Figure 2 represents the distribution of mean difference of categorization in both control and experimental groups.  The findings revealed that both groups of EFL learners (CG) and (EG), committed errors with an overall frequency of 315. The (CG) committed errors across all categories to 199 frequency compared to the (EG) with a total frequency of 116. The findings demonstrate that the treatment given to the (EG) concerning writing exercises in the classroom has a positive relationship with the production of written sentences and are consistent with the previous study that practising writing assignments in the classroom reduce the frequency of the errors (Bensen, 2014;Khansir, 2012;Qinghua, 2010;& Aydin, 2010).

Syntactic errors
In the category of "syntactic taxonomy", the learners made errors with an overall frequency of 33, 10%, which showed the third-largest frequency in the data in terms of errors. The second category "morphological taxonomy" accounted for an overall error frequency of 65, 2.6% in the data across both groups. This demonstrated to be the second-highest taxonomy in error production by the EFL learners. The third category, "orthographic taxonomy" is the largest category in error production with an overall frequency of 102, 32.3% across both groups: the control group and experimental. The least produced errors were revealed by the category in "lexical taxonomy" with an overall frequency of 5, 1.5%. The findings demonstrated that the learners in both groups committed errors in all four categories with some variations. The findings are consistent with the previous study that EFL learners in academic writing commit errors with various frequencies (Al-Tamimi, 2018).
The findings revealed that the experimental group (EG) committed errors significantly smaller in frequency compared to the control group (CG) which did not receive the treatment. The control group (CG) committed 1 error in the "wrong word order" but the experimental group did not commit an error in this category which could be the effect of treatment received by this particular group. The findings are dissimilar in the category of "wrong word order" with the previous study that EFL learners commit errors in writing with the highest frequency (Salehi et al., 2018). The following illustration exemplifies the error in this category. The learners produced the second-highest errors in "wrong use of preposition" in a syntactic category with a frequency of 12 errors by the control group (CG), whereas 7 errors were committed by the experimental group (EG). The overall error frequency in this group mounted to 19, 6% in the overall data. The findings indicated the effect of written exercises on the experimental group (EG) by committing errors reduced almost to half, which is 7 compared to the higher frequency of errors produced by the control group (CG), which are 12. Excerpt 2 indicated the misuse of prepositions "of" and "for" by the learners. . The learners have difficulties in choosing the correct prepositions which seems to be the negative interference of L1 on the target language. Therefore, a large number of interlingual prepositional errors occurred in the data from both groups. The EFL

Figure 3. Scatter Plot of Pairs
learners, finding it hard to choose a correct preposition also indicates the incomplete knowledge of the L2 users. Consequently, the interlocutors have difficulties picking up the right preposition for the language in use. Excerpt 3 demonstrates the overused application of unnecessary prepositions by the L2 learners. The findings regarding the prepositional aspects of errors are consistent with the previous study that L2 learners of the Arabic language commit errors due to ignorance of rule restrictions and consequently, overgeneralizations cause intralingual errors (Tahaineh, 2010). The excerpts exemplify the mistakes in this category.
Excerpt 2-CG: Taif is famous of *the flowers.
(Taif is famous for flowers).

Excerpt 3-EG:
Taif is one of the cities in the of* Saudi Arabia.
(Taif is one of the cities in Saudi Arabia).
The wrong use of articles, the third category in the "Syntactic taxonomy", indicated the highest errors in this group with an overall frequency of 33 (10.4%). The control group (CG) committed the highest number of errors in this group with a frequency of 21 compared to a significantly smaller number of errors made by the experimental group (EG) with a total frequency of 12. The findings revealed that the effect of treatment on the experimental group reduced the number of errors to a smaller frequency with 12, but the control group which did not receive the treatment made errors almost double the number of the experimental group with a frequency of 21. Excerpt 4 reveals the faulty expression of the L2 learners of misusing the articles due to the L1 interference. The findings of the errors in the article category are consistent with the previous study that inappropriate use of the definite articles and the omission of definite articles reflect the incompetency and inappropriate knowledge of the articles in the target language by the EFL learners (Elumalai, 2019). The following illustrations exemplify the errors made by the learners in this category.

Excerpt 4-CG:
Taif is the* famous city in Saudi Arabia.
(Taif is a famous city in Saudi Arabia).
Excerpt 5-EG: Taif is one of* most popular city for tourism.
(Taif is one of the most popular cities for tourism).

Morphological errors
The category of "wrong comparative adjectives", in the morphological taxonomy, received a smaller number of errors across the data. The control group (CG) produced 2 errors in this category compared to the experimental group (EG) which produced 0 errors in wrong comparative adjectives. The following sample reveals the incorrect use of superlative adjectives by the control group.
Excerpt 6-CG: Saudi Arabia is known as one of the hotte* countries in the world.
(Saudi Arabia is known as one of the hottest countries in the world).
Surprisingly, the second and third categories, the "wrong use of possessives" and "omission of plurals" in the morphological taxonomy demonstrate 0 error frequency. These two categories revealed the adequate knowledge of the L2 learners of the target language and the effect of positive interference due to their similarities in the structure of both the languages (Contrastive Analysis Hypothesis).
The last category "wrong subject verb-tense" in "morphological taxonomy" reveals the second-highest errors with an overall frequency of 65 (20.6%) by both groups: control group (CG) and experimental group (EG). However, the effect of treatment on the experimental group is indicated by their smaller frequency of overall errors with 24, compared to the control group which produced errors with almost a double frequency of 41 in the data. Learners' incompleteness of the basic knowledge regarding the errors in subject verb-tense is consistent with the previous study that attributed these errors of low proficiency level to their low motivation by various stakeholders. The findings agree with the previous study that reveals the poor training skills of the teachers, and their lack of competencies in the target language, though they are native speakers (IvyPanda, 2020). Little exposure to the target language out of class and inadequate and outdated instructional materials could also be the reasons responsible for these errors at this level. It also puts into question, the instructors' teaching strategies and marking criteria whether L2 learners receive enough attention or just pass on to the next class (Khan & Sohail, 2021). This question will be answered by future researchers due to the limitations of the current study. Of course, the interference of the mother tongue in the acquisition of L2 cannot be underestimated. Excerpt 8 indicates the L1 negative interferences due to the structure of the Arabic language compared to the English Subject-Verb-Object structure.
Excerpt 7-CG: There are* a restaurant.
(There is a restaurant).

Excerpt 8-CG: It is my city and I love is*.
(It is my city and I love it).

Orthographic errors
The errors in the "orthographic taxonomy", the third category, revealed the errors with the highest frequency across the data. In the category of "faulty capitalization", the control group (CG) committed the errors with a higher frequency that is 30 compared to the lower number of errors committed by the experimental group (EG) with 22 frequencies in the data. The learners' lack of knowledge about the target language and their inattentiveness to form made them produce sentences with faulty capitalization. Except 9 reveals the overgeneralization and misuse of rules employed by the learners in capitalization. The following sample sentences illustrate the errors in sentences regarding faulty capitalization.
Excerpt 9-CG: Taif is characterized by a moderate climate in the Summer*.
(Taif is characterized by a moderate climate in the summer).
Excerpt 10-EG: my* city name is Taif, it's is in Saudi Arabia in middle* east*.
(My city name is Taif, it is in Saudi Arabia in the Middle East).
The second category of "orthographic taxonomy" consisted of 'punctuation in which the L2 learners produced errors with an overall frequency of 36 (11.4%) by both groups. In this category, the control group (CG) produced a higher number of errors with a frequency of 28 compared to the lower frequency 8 produced by the experimental group (EG). In punctuation, the findings revealed that the insufficient knowledge of the learners about the target language made them susceptible to omitting comma and using incorrect punctuation in the production of their sentences. The following illustrations from the data exemplify these errors committed by both groups with varying frequencies.
Excerpt 11-CG: In Taif* there is a rose festival.
(In Taif, there is a rose festival).
Excerpt 12-EG: My city is a good city* My city has little traffic jams* (My city is a good city. My city has a few traffic jams).
The category of "spelling errors" in the "orthographic taxonomy" indicated the highest frequency of errors in the whole data with 102 (32.3%). Moreover, the findings demonstrated that the control group (CG) produced more errors in this category with a frequency of 60 which is significantly higher compared to the lower frequency of 42 errors in this category by the experimental group. The errors in the misuse of spelling by both groups indicate their incomplete knowledge of the target language.
The following examples from the data reveal the errors in this category by the learners.
(There are many flowers there).

Excerpt 14-EG:
The wather* in the Summer good.
(The weather in the summer is good).

Lexical errors
In the category of "wrong word choice", in the "lexical taxonomy", the learners in both groups produced the errors with a considerably overall lower frequency of 5 (1.5%) across the data. The control group (CG) indicated the errors in this category with a higher frequency, of 4 compared to the lowest frequency produced by the experimental group with 1 frequency. The misuse of "wrong word choice" in except 15 by the control group demonstrates their insufficient knowledge of the target language because the word "overlooking" has a contrary meaning to what the interlocutors intended for the target item "looking over". Moreover, in excerpt 16, the experimental group faced difficulties using the appropriate word in the first place such as they employed "lower" for "low level" and encountered problems placing the words in an appropriate order. Secondly, the experimental group, in excerpt 16, used pronouns and nouns together which violates the grammatical rules of the target language and creates confusion in the mind of the readers due to the incomplete knowledge of the L2 learners.
Excerpt 15-CG: it has many wonderful places. Over looking* high views.
(It has many wonderful places. Looking over high views).
Excerpt 16-EG: it is mustly mountins so taif have oxgen lower*, the taif good rols for streets and they people* have fun.
(It is mostly mountains, so Taif has a low level of oxygen, Taif has good roads and streets and the people have fun).

Implications of the study
The study reveals pertinent implications that both groups committed a significant number of errors in all the four taxonomies: "Syntactic taxonomy"; "Morphological taxonomy"; "Orthographic taxonomy"; and "Lexical taxonomy". The control group (CG) produced the t most number of errors with a 199 frequency compared to the lower frequency of errors produced by the experimental group (EG) with 116. The lower frequency of errors by the experimental group can be attributed to the effect of writing exercises by learners which reduced the errors to a greater extent compared to the higher frequency of errors made by the control group. Therefore, the adaptation of teaching strategies by employing writing exercises in the classroom could lessen the number of EFL learners' errors significantly. Secondly, the course content and syllabus require revision to incorporate writing exercises in the classroom which will help improve the L2 writing skills.
The most frequent errors produced by both groups with a higher frequency consisted of "spelling errors". The control group (CG) committed 60 errors in this category while the experimental group (EG) made 42 errors. The next highest category of errors by both groups in the data consisted of "wrong subject verb-tense", the control group committed 41 errors in this category compared to the smaller frequency of errors from the experimental group with a frequency of 24. The thirdhighest category of errors committed by both groups, that is, "faulty capitalization"-the control group and experimental group respectively produced errors in this category with 30, and 22, with an overall 52 error frequency. The medium range of error frequency occurred in the category of "punctuation" by both groups-control group and experimental group respectively, with 28, and 8 and overall 36 frequency. The "wrong use of the article" occurred with 21 and 12 error frequencies.
The overall errors by both groups amounted to 33, and the "wrong use of preposition" occurred with a 12 and 7 error frequency and the overall error frequency by both groups amounted to 19 The lowest frequency of errors occurred in both groups respectively in the categories of "wrong word choice", with a frequency of 4, and 1, overall errors 5; "wrong com-adjectives", with a frequency of 2, and 0, overall 2; and "wrong word order", with a frequency of 1, and 0, and overall 1. The findings revealed that both the groups produced "0" errors in two categories, which are, "wrong use of possessives" and 'omission of plurals'. The findings are consistent with previous studies that EFL learners commit frequent writing errors in "tense", "preposition", and "articles" (Abushihab, 2014;Atmaca, 2016;Kırkgöz, 2010;Terzioğlu & Bensen Bostanci, 2020;Zheng & Park, 2013).
The interlingual negative interference of L1 on the learners is indicated by the misuse of prepositions. The learners face difficulties to choose the appropriate prepositions between "of" and "for". The overgeneralization and ignorance of rules lead the learners to commit serious errors. Therefore, the explicit teaching strategies regarding the prepositional rules will be of immense use for the learners. The misuse of the definite article and its omission indicate the incomplete knowledge of the L2 learners which needs to be dealt with in the class by the instructors through writing exercises. The highest frequency of errors in "subject verb-tense" category can be attributed to the poor skills of the English language teachers, lack of exposure to the target language in and out of the class, and poor marking criteria which do not highlight the learners' errors, and finally, the demotivation factor which blocks the affective filter as L2 learners pass on from one class to another. The findings are similar to the previous hypotheses developed that practising writing essays and assignments in the classroom motivates l2 learners, reduces the frequencies of errors and gives opportunities to the teachers to provide timely feedback (Bensen, 2014).
Proper attention is required to be paid to the learners' excessive errors to prevent fossilization in the writing process. Coding of errors in written exercises in the classroom by teachers' feedback, peer work and group work will help and facilitate the learners understand them and minimize the errors (Pawlak, 2014). Explicit correction of errors may affect the developmental aspect of the learning process in the L2 context, but on the other hand, it helps the teachers plan ahead in view of the learners' abilities, design course materials, and provide feedback on the learners' shortcomings.

Conclusion
The findings of the study highlight the effect of writing exercises on the types and frequency of errors on the learners' written sentences at the undergraduate level at universities in Saudi Arabia following a quasi-experimental design. The learners committed all four types of errors with various frequencies. The control group produced errors with a higher frequency in individual categories as well as in overall data such as 199, and 116. The most frequent errors produced by the control group (CG) across the data are "spelling errors" with 60 frequency, "wrong subject verb-tense, with a 41 frequency, 'faulty capitalization' with 30 frequency, 'punctuation' with 28 frequency, 'wrong use of articles' with 21 frequency, and 'wrong use of preposition' with 12 frequency. On the other hand, the experimental group (EG) produced the most frequent errors in "spelling errors" with 42 frequency, 'wrong subject verb-tense, with 24 frequency, "faulty capitalization" with 22 frequency, "wrong use of articles" with 12 frequency, and "wrong use of preposition" with 7 frequency. Both groups, the control group (CG) and experimental group (EG), had errors on "wrong word choice" with a frequency of 4, and 1, respectively though the experimental group prevented committing errors in some categories such as "wrong word order" with frequency 1, and 0, and "wrong comadjectives" with a frequency 2, and 0 compared to the control group. Finally, both groups revealed the same behaviour in producing 0 errors in two categories such as "wrong use of possessives", with frequency 0, and 0, and the "omission of plurals" with frequency 0, and 0.
The findings indicated the experimental group (EG) produced a significantly smaller frequency of errors in all the categories. In some categories, the experimental group (EG) prevented making errors due to the effect of writing treatment on them, unlike the control group (CG). Moreover, the findings revealed interlingual negative L1 interference on the learners, intralingual overgeneralization, fossilization, lack of attentiveness, and incomplete knowledge of the learners' target language. The findings also revealed the inapt skills of the teachers and their classroom instructional methodologies, as well as inappropriate course materials. Certain questions remained unanswered due to the scope and limitations of the current study. These questions might be of interest to future researchers. For example, data from female learners, highest or lowest level learners, measuring the level of students' motivation on written sentences, level of learners' anxiety and effective filters on the production of written sentences could yield different results.

Funding
The author received no direct funding for this research.

Author details
Afzal Khan 1 E-mail: afzalenghu@gmail.com 1 English Language Skills Department (ELSD) Common First Year, Al Khaleej Training and Education-King Saud University, Riyadh, Saudi Arabia.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Citation information
Cite this article as: The effect of writing exercises in classroom on the production of written sentences at undergraduate level by Saudi EFL learners: A case study of error analysis, Afzal Khan, Cogent Education (2022), 9: 2122259.