EFL Learners’ Preferences and Emotions about Oral Corrective Feedback at Secondary Education in Turkey: Are There Gender and Grade-level Differences?

DOI: http://dx.doi.org/10.21093 /ijeltal.v6i1.889 Oral corrective feedback has attracted the attention of EFL teachers and researchers in recent decades. The current study aims to investigate EFL learners’ preferences and emotions about oral corrective feedback with a specific focus on gender and grade level at secondary education in Turkey. Quantitative research methods were used to reveal the findings. 664 EFL learners participated in the study at two different state high schools in Ankara, Turkey. Mann-Whitney U and Kruskal-Wallis tests were used to investigate the OCF preferences of EFL high school learners. It was revealed that more than half of the high school EFL learners worried about making mistakes in the classroom; however, the majority of the learners agreed on the necessity of receiving OCF and they had positive feelings for being given immediate feedback by their teachers. Regardless of gender and grade level, EFL learners had a higher preference for ‘self-correction’. Females preferred ‘metalinguistic clues’ as the first choice and ‘peer correction’ as the last choice among the OCF types whereas males prefer ‘peer correction’ as the first, but ‘metalinguistic clues’ as the last choice. Among the grade levels, 10th-grade level learners have more concerns about making mistakes and these learners prefer ‘clarification request’ and ‘repetition of error’ a lot more than the 9thgrade level learners.


Introduction
Foreign language teachers have relied on the notion of 'Oral Corrective Feedback' (henceforth, OCF) for the last two decades as it serves as a powerful remedy for leading learners to notice the incorrect language form and then to correct it (Fadilah et al., 2017;Park, 2010;Russel, 2009;Zhang & Rasimi, 2014).
Indonesian Journal of English Language Teaching and Applied Linguistics, 6(1), 2021 The findings of many studies have revealed that the provision of OCF can help learners acquire the L2 target forms in classrooms (Ellis et al., 2006;Lyster, 2004;Lyster & Saito, 2010;Sheen, 2004;Yang,2016). Furthermore, many studies have suggested that supporting learners with various corrective feedback might help them acquire the correct form as they are actively participating in the learning process. Therefore, the lack of corrective feedback might lead to fossilized errors as learners would assume their statements are correct (Fadilah, et al., 2017).
Language teachers have a pivotal role in this process as they try to turn this erroneous speech into learning opportunities without demotivating learners, and most learners have positive attitudes towards receiving OCF (Amalia et al., 2019). During this collaborative process, language teachers might have some responsibilities. For instance, as Katayama (2007) suggests, they can conduct surveys to find out learners' expectations during the learning process. This might contribute to the effectiveness of OCF in language classrooms. Likewise, Ellis (2010) forwards an argument that individual factors, such as anxiety, beliefs, and selfefficacy about OCF play a pivotal role in the effectiveness of OCF. Sopin (2015) also points out that language teachers should care about learners' emotional state and understand their personality traits since the learners' expectations might affect their attitudes.
The motive for investigating preferences and emotions of language learners with regard to OCF types takes its roots from the above discussions. Likewise, the study of preferences for OCF types within the gender and grade level framework has some other motivations. As cited in Geçkin (2020), there is some research which investigates the correlation between gender and foreign language anxiety (Çağatay, 2015;Yih et al., 2017); however, there are still many unexplored areas regarding the role of gender in the field of foreign language learning (Sunderland, 2010). Furthermore, investigating the role of gender differences might help determine the use of OCF (Amalia, et al. 2019). The second framework, which is the relation between the preferences for OCF types and grade levels has been highly ignored.
Though teachers' perceptions regarding OCF types are out of the scope of this article, the findings might be accepted as food for thought for teacher educators and language instructors so that they can adjust their choices for OCF types in line with their students' preferences. Additionally, most OCF research has been carried out regarding learners at the tertiary level. (Amalia, et al. 2019;Geçgin, 2020;Hassan & Arslan, 2018;Shoaei & Kafipour, 2016;Ünsal-Şakiroğlu, 2020;Yang, 2016). However, high school learners also constitute a big majority of language learners and should not be ignored.
Taking its reasoning from the above-mentioned motivations, this study aims to answer the following research questions: a. Are there any differences among high school EFL learners in the preferences and emotions towards OCF? b. Are there any differences among high school EFL learners in the preferences and emotions towards OCF in terms of gender and grade level? c. How do high school EFL learners emotionally respond to the immediate feedback in class? d. What types of OCF do high-school EFL learners prefer in language classes? e. Are there any differences among high school EFL learners in preferences of OCF types in terms of gender and grade level?
The majority of OCF research is based on the classification of OCF types in the study of Lyster & Ranta (1997). Primary level four immersion classrooms are investigated and then corrective feedback is categorized into six types: repetition, elicitation, metalinguistic feedback, recast, explicit correction, and clarification requests. Lyster & Ranta (1997) forward that among OCF types, 'recast' is the most commonly used feedback type by the language teachers (%55). This finding, since then, has been the focus of many subsequent studies on corrective feedback. After 'recasts', elicitation, clarification requests, metalinguistic feedback, explicit correction, and repetition of errors are rated in terms of preferences. Despite its high-level preference, recast is found to be the least likely feedback type which results in learner uptake whereas elicitation results in more successful learner responses. The ineffectiveness of recasts is also highlighted by Zhao (2015). EFL Chinese learners have difficulty in noticing the corrections in recasts. EFL learners assume that recast is a repetition of their utterance and it might fail to warn the learners that they have generated an error word (Amalia et al., 2019;Elçin & Öztürk, 2016).
In their subsequent study, Ranta & Lyster (2007) divide corrective feedback into two categories: prompts and reformulations. Prompts consist of metalinguistic feedback, elicitation, clarification request, and repetition. They help the learners to repair their errors by providing necessary cues. Reformulations refer to the explicit feedback and recast, which do not lead to learner repair.
Concurrent with Lyster & Ranta (1997), Sheen (2004) studies four different communicative classroom settings where learner uptake, subsequent learner repair, and teacher feedback moves are investigated. It is argued that the effectiveness of recasts related to learner uptake is controversial. On the contrary, Ellis et al. (2001) investigate OCF types used by teachers in form-focused practices and report that 'recasts' are the main feedback type leading to the highest amount of learner uptake. Particularly, for grammatical errors, Lyster & Saito (2010) argue that recasts tend to be ambiguous. Therefore, learners are required to have a high degree of awareness to distinguish the mismatch between the wrong and the correct forms of the target language. In contrast, Lyster (2001) reports that negotiation of form (i.e. repetition, metalinguistic feedback, elicitation, and clarification request) can lead to a higher rate of grammatical repair. However, recast is the feedback type most commonly used to correct phonological errors since the correction of phonological errors does not require changing the word; therefore, it does not cause ambiguity (Lyster, 2001). Lochtman (2002) has carried out a study with German students undertaking form-focused instruction. It is found in the study that teachers use metalinguistic feedback and elicitations so as to guide their students to correct themselves. It is reported that recasts are significantly higher when the context of instruction moves from form-focused instruction into a meaning-Indonesian Journal of English Language Teaching and Applied Linguistics, 6(1), 2021 focused one. On the contrary, when the activity is more analytic, teacher guidance to selfcorrection occurs, encouraging students to correct the target linguistic form.
Besides the nature of the lesson, the type of error indicates the type of OCF used by the instructor. For instance, Yang (2016) has suggested that metalinguistic feedback has a good effect on pragmatic errors. Additionally, recasts are regarded as more effective for phonological errors than for lexical and grammatical errors. A notable finding from Yang's (2016) study is that learners at the intermediate level mostly favor clarification requests for phonological errors.
It is worth mentioning the prominent studies which signal a specific type of corrective feedback to be more successful depending on some variables. In some cases, these variables are individual differences including learner anxiety and proficiency level, in other cases, gender differences are investigated. The findings of the studies carried out among advanced learners display that repetition and elicitation OCF types are favoured by advanced learners (Kaivanpanah et al., 2015;Katayama, 2006). In a similar vein, Lin & Hedgcock (1996) have studied the impact of OCF types on student repair with a focus on learners' proficiency levels.
Four high proficient university students and four low proficient Chinese immigrants are selected as participants. The results display that the high proficient group notices the error correction and successfully repairs them while the low proficient immigrants have a lower error correction success rate. Yoshida (2008) reports that teachers choose corrective feedback in accordance with learners' proficiency levels and learning styles. Fadilah et al. (2017) have investigated freshman and sophomore Indonesian learners' corrective feedback preferences, including time, error types, corrective feedback strategies; and relationships between students' foreign language anxiety and corrective feedback preferences. The authors have found out that as the grade level goes up, learners attach more importance to the accuracy of expression. However, both groups have a tendency to prefer delayed feedback which is provided after they end their speech.
The first issue is about the gender differences in foreign language learning anxiety which is also reflected in their OCF preferences. Geçgin (2020) gives an example of the fear of negative evaluation and forwards that females get more upset when they do not understand what the language teacher is correcting. As regards the timing of OCF, female participants prefer delayed feedback while male participants prefer immediate feedback.
Gender differences are also investigated in preferences for OCF types. Amalia et al. (2019) have pointed out that neither male nor female participants prefer 'repetition' because emphasizing the erroneous part makes them feel uncomfortable. 'Request for clarification' is another OCF type neither gender prefers since it makes the teacher mean ambiguous. Unlike these common points between males and females, the authors also point out some differences. For example, males mostly prefer 'explicit correction' while females prefer 'recasts' and 'metalinguistic feedback'. 'Metalinguistic feedback' leads the participants to self-correct themselves benefiting from the teacher's clue. Geçgin (2020) also supports this view and forwards that females mostly prefer 'metalinguistic feedback'. However, males rate 'elicitation' as a more effective method in her study. Studies might yield controversial results in gender differences; for instance, Khorshidi & Rassaei (2013) point out that clarification request and repetition are the mostly preferred feedback types whereas explicit feedback is the least frequent feedback type among males and females.

Research Design
Quantitative research methods have been used to analyze High School EFL learners' preferences and emotions towards OCF in terms of gender and grade level. To achieve this goal, this study uses a descriptive research design to provide detailed information about the constructs to be explored (Cohen et al., 2011).

Participants
A total of 664 high school EFL learners from two different state high schools participated in this study. Participants were selected randomly. Though the high school education program consisted of four grades; learners in 9 th , 10 th , and 11 th grades participated. As the learners at the 12 th -grade level were training for the university entrance exam and they were much fewer than other grades in number, they did not participate in the study. Of the 664 participants, 361 (54,4 %) were females, and 303 (45,6 %) were males. Additionally, of the 664 students, 277 (41,7 %) students were in 9 th grade (the first year of high school education), 307 (46,2 %) of them in 10 th grade, and finally, 80 (12%) of them were at 11 th grade. Although participants had studied English in primary education, they started from A1 level at 9 th grade, which was the first year at high school (secondary) education as they did not get an intense English program formerly. When this study was conducted, learners from the 9 th -grade level were studying A1 level; learners at the 10 th -grade level, A2 level; and learners from the 11 th -grade level were studying B1 level course materials.

Instrument
The data were collected by means of a questionnaire used to investigate learners' preferences and emotions for OCF. The questionnaire was adapted from Fukuda (2004), Agudo (2013), and Katayama (2007), and consisted of three parts. The first part consisted of 4 statements about students' attitudes towards corrective feedback use in the classroom. The statements were presented with choices such as 'always, sometimes, never'. The second part investigated students' perceptions and emotions about OCF using a Likert scale on eight declarative statements. A 5-point Likert-type scale was utilized for each item ranging from 'strongly agree' to 'strongly disagree'. The 9 th statement, which was not presented in a Likerttype scale, investigated students' emotions about receiving immediate feedback during speaking activities. Students were allowed to choose more than one emotion. The last part of the questionnaire consisted of various cases displaying an example about a specific type of OCF. The statements were presented on 5-point Likert-type scale. All the content in the student questionnaire was translated into Turkish to avoid ambiguity and misunderstanding. Two experts (a professor and an instructor in the field of English Language Teaching) translated the questionnaires into Turkish. A pilot study was conducted on 137 learners and the results displayed that the questionnaire was reliable, with a Cronbach's Alpha value of .80.

Data Collection
The researcher contacted the school administration and language teachers at both high schools. Formal consents were taken from school administration and the language teachers were asked to collect data from students in class. As the instructions and statements were in learners' native language, they did not have any difficulty in understanding statements.

Data Analysis
Quantitative data were analyzed in SPSS 21. First of all, frequency analysis was employed to set the scene about the gender and grade level. Then comparison tests were conducted to compare the responses given in each part in terms of gender and grade level. In order to test the correlation between the responses in part 1 with gender and grade level, crosstabs Chisquare analysis were conducted. To analyze the data in the second and third parts of the questionnaire, descriptive statistics (e.g. standard deviation, mean scores) were used to present participants' responses. Normality test was employed in order to test the assumption of comparison tests, and it was observed in Kolmogorov-Smirnov test that the assumption of normality was not met in each subset of data for the grade level and gender. Therefore, Mann-Whitney-U test, which is a non-parametric test, was conducted to analyze the difference in preferences about OCF in terms of gender, and Kruskal Wallis test was conducted to explicate the different preferences about OCF in terms of grade level.

Findings
The aim of the study is to explore the preferences and emotions of high school EFL learners about OCF and to reveal whether there is a statistically significant difference in terms of gender and grade levels. The results are presented under each research question.

4.1.
Are there any differences among high school EFL learners in the preferences and emotions towards OCF?
The overall percentages of responses to the first part of the questionnaire is shown below.  Table 1 shows that 55 % of the participants prefer to be corrected by their teachers whenever they make mistakes while speaking. 41,7% of the participants prefer to be corrected by their teachers only when they make major mistakes, which means they want their teacher to tolerate their minor mistakes while speaking. It is also seen in the table that only 3,3% of the participants do not prefer to be corrected at all.
Likewise, 67,9% of the learners prefer their oral mistakes to be corrected in class while 32,1% of the learners prefer to be corrected by the teacher when they are alone after class. It is so remarkable that 80,7% of the learners prefer to be given immediate feedback during speaking activities while 19,3% of the learners prefer to be corrected after class. There is not a big difference in the percentages of the responses to the 4 th statement. 56,8% of the learners prefer to be corrected individually; however, 43,2% of the learners prefer to get group feedback for their oral mistakes.
The following table displays the descriptive statistics (mean, frequency, standard deviation) for the preferences and emotions of high school EFL learners towards OCF regardless of gender and grade levels. As shown in Table 2, the statement which received the highest mean score is the second one (4,2003), which shows high school EFL learners recognize the need and usefulness of OCF. This also signals the effectiveness of OCF in language classes. The statement with the second highest mean score is the first statement (4,1386), which indicates that learners benefit from OCF in class. The rest of the statements are about how high school EFL learners emotionally react to OCF. Nearly two-thirds of the high school EFL learners (60,3%) have negative feelings towards making oral mistakes. Furthermore, more than half of the participants (53%) have worries about making oral mistakes in class. 43,2% of the participants feel upset when they do not understand what the teacher is correcting. In this vein, it is important to note that the teacher should use the appropriate type of OCF in order not to lead the individual learner to negative feelings. For example, an explicit correction could be appropriate for these types of learners. Nearly one-third of learners (38,2%) state that oral mistakes make them doubt themselves. Additionally, one-third of high school EFL learners (33%) are not comfortable with their language teachers' tendency to correct each oral mistake in class. Finally, the sixth statement has obtained the lowest mean score, which indicates that 21,1 % of the participants are not content with the teachers' error correction in class. When observed from another angle, the last finding tells us that more than half of the participants (65,5%) favor receiving error correction from the teacher, which also shows us that learners acknowledge the authority figure of their language teachers.

Are there any differences among high school EFL learners in the preferences towards OCF in terms of gender and grade level?
In the table below, the comparison of the responses of male and female high school EFL learners is displayed. As Table 3 shows, there is not a statistically significant difference in EFL learners' preferences towards OCF in terms of gender.
The same data have been statistically analyzed to see if there is any significant difference in responses in terms of grade level. As shown in Table 4, responses to statements 2,3 and 4 differ in terms of grade level. Learners at the 10 th grade prefer to be corrected in class; given immediate feedback and corrected individually more than learners at the 9 th and 11 th grade levels.   According to Table 6, high school EFL learners differ only in items 4 and 5. A significant difference was found between the 9 th and 10 th grades and 10 th and 11 th grades in the 4 th and the 5 th items. In this vein, learners at the 10 th grade tend to worry about making oral mistakes in language class more than learners at the 9 th and 11 th grades. Likewise, learners at the 10 th grade have negative feelings about making oral mistakes because mistakes make the learners doubt themselves when compared to the other grades.

How do high school EFL learners emotionally respond to getting immediate feedback in class?
This study also aims to understand learners' emotional reactions to immediate feedback while speaking. The findings of this part are illustrated in Figure 1. Percentages in Figure 1 are obtained by dividing the given frequency by the total number of students (664). Since multiple markings are involved, the sum of percentages is not 100. Figure 1 displays that 'feeling satisfied' (57,98 %) is the top choice, followed by 'feeling sorry' (33,73%) in the second rank, and followed by 'feeling embarrassed' (31,93%) in the third rank. This figure indicates that although the top choice is a positive feeling and most learners feel satisfied when they were provided with immediate oral feedback, there are some learners who can be sensitive to immediate OCF and feel uncomfortable with it in the class. Table 7 indicates the preference of the OCF perceived by the high school EFL learners. Findings have been revealed through analyzing the third part of the questionnaire. As indicated in Table 7, learners preferred self-correction (4,2786) as the top choice, followed by recasts (4,1205), explicit correction (4,0060), metalinguistic clue (3,9307), elicitation (3,8434), clarification request (3,3102), repetition of error (2,9639), and peer correction (2,4578), as the lowest choice. This finding aligns with the finding in Table 1 which indicates high school EFL learners' perception of their teacher as an authority figure.

Are there any differences among high school EFL learners in preferences of OCF types in terms of gender and grade level?
In the third part of the questionnaire, the assumption of normality has been primarily taken into consideration in order to make comparisons according to grade level and gender variables. According to Kolmogorov-Smirnov test, there is no normal distribution. In this vein, Mann Whitney-U test, which is a non-parametric test, has been conducted for the gender variable. On the other hand, Kruskal-Wallis test, which is also a non-parametric test, has been used for grade-level variables since it consists of three groups. The findings of the tests are displayed below. preferences, and in the fifth-order in males' preferences. However, self-correction is observed as the top choice when the results are assessed generally regardless of any variables (shown in Table 7). This can be explained by the fact that the first three preferences of females are the last three preferences of males. Table 9 indicates the results of Kruskal-Wallis test, which has been conducted to investigate high school EFL learners' preferences for OCF types in terms of a grade-level variable. As illustrated in Table 9, learners in the 9 th grade preferred elicitation as the first choice, and repetition of error, as the last choice. Learners in the 10 th grade perceived repetition of error as the most preferred one, but elicitation, as the least preferred type of OCF. The first and the last choices of the 9 th and 10 th grades are surprisingly contradicting. As for the 11 th -grade learners, explicit correction is the first choice, and the metalinguistic clue is the last choice. When the data is analyzed through Kruskal-Wallis Test, it is observed that learners differ in items 5 and 8 in terms of grade level. In both items, there is a statistically significant difference between the 9 th and 10 th grades. The fifth type of OCF is clarification request, which learners in the 10 th grade preferred more than learners in the 9 th grade. Similarly, learners in the 10 th grade preferred repetition of error more than learners in the 9 th grade.

Discussion
In a foreign language learning context, FL teachers provide various types of OCF when learners generate erroneous utterances in communicative contexts. In the present study, most of the EFL learners (55%) preferred to be corrected any time they make mistakes during oral performances in the classroom. Only a small percent of EFL learners (3.3 %) do not prefer to be corrected at all. This finding supports Katayama (2007) suggesting that foreign language learners believe in the necessity of feedback.
The majority of EFL learners (80%) have positive attitudes towards receiving immediate OCF. This finding is supported by Kazemi, Araghi, and Davatgari's (2013) study exploring Iranian EFL learners' preference for immediate correction in the classroom. Similarly, Fidan (2015) reports that Turkish EFL learners prefer to be corrected by their language teachers using OCF. However, this finding contradicts the finding derived from Rahimi & Dastjerdi's (2012) research. They have investigated the effect of delayed and immediate feedback and have found that delayed feedback fosters learners' oral production. Likewise, Fadilah et al. (2017) state that learners are prudent to receive delayed feedback which occurs at the end of their speech.
As for the emotional reactions towards making errors, it is found that the majority of the learners have concerns about making mistakes during their oral performances. It is remarkable that despite these concerns about making errors in communicative contexts, learners have positive attitudes towards being corrected immediately in the classroom. This finding is congruent with that of Fadilah et al. (2017), displaying that learners believe they have benefited from oral corrective feedback regardless of their anxiety level.
When their feelings are investigated, it is found that the majority of the learners feel satisfied (57,98 %) if they are provided with immediate OCF. The second feeling following 'satisfied' is feeling 'sorry' (33,73 %). Similarly, in Agudo's study (2013), when learners' emotional reactions are asked, the top choice is 'feeling satisfied' and the second choice is 'feeling embarrassed' given immediate OCF. This finding signals the indispensable role of individual differences. Teachers should always consider there will be reserved learners who can feel sorry or embarrassed when being corrected instantly. Another conclusion might be, if learners are provided with the type of OCF addressing their preferences, their negative feelings can change into positive ones; and the process of giving feedback would be more constructive.
When the preferences for OCF types of high school EFL learners are considered regardless of their gender and grade level, three highly preferred OCF types are present. Learners prefer self-correction as the first choice, recasts as the second, and explicit correction as the third.
Self-correction was also promoted by many researchers (Chaudron, 1988;Kasper, 1985;van Lier, 1988). The top preference of EFL learners is supported in another study conducted by Yoshida (2008) in the Japanese context where learners mostly prefer self-correction. Comps (2003) yields that language teacher should motivate learners to correct themselves in EFL classes so that they can maintain self-correction skills outside the class, as well. The second choice of EFL learners in this study is recast, and it has been highlighted in many studies before. For instance, Lyster & Ranta (1997) accept as one of the favored corrective feedback types as the teacher provides the correct form by repeating all or only a part of the learner's speech. Although many researchers carry out studies promoting recasts in corrective feedback as they are effective for language acquisition (Mackey & Philp, 1998;Loewen & Philp, 2006;Morris, 2002), some research highlight the potential ambiguity of recasts (Lyster & Ranta, 1997;Philp, 2003). This ambiguity, as Chaudron (1988) explains, might stem from learners' misperception of teacher feedback such as confirmation of meaning rather than correction of erroneous utterances.
When the gender variable is considered, it is found that females worry about making mistakes more than males. This indicates that female learners have less tolerance for producing erroneous utterances as they might expect to perform well in communicative contexts. Amalia et al. (2019) support this finding revealing that female learners mostly express some concerns about making errors or being corrected in front of their peers. Males have more negative feelings towards the teacher's explicit correction in the class whereas females are more open to the teacher's feedback. Carvalho et al. (2014) support this finding claiming that females are more critical than males about the quality of OCF they receive. Hence, females regard teacher feedback as more dependable as it is provided with an authority figure.
Zacharias (2007) also forwards that regardless of gender, Indonesian students tend to rely more on teacher feedback as they consider teachers' correction more reliable and the cultural belief is that teachers are the source of knowledge. As Rollinson (2005) claims, feedback obtained from peers is regarded as less authoritarian and beneficial compared to the other sources (cited in Geçgin, 2020). On the other hand, this study reveals that males feel comfortable when they receive peer feedback. This might show that they could find their peers more encouraging and less judgmental. However, this finding contradicts with Hassan and Arslan's (2018) study where peer feedback is rejected as they believe that peers do not have the necessary knowledge to be able to correct the errors. It is revealed in this study that females mostly feel upset when they do not understand what the teacher is correcting. Very similarly, Geçgin (2020) has drawn the same conclusion for female learners in her study.
This study has also explored the difference between female and male learners in terms of their preferences for OCF type. Though Khorshidi & Rassaei (2013) have found no significant difference between OCF preferences in terms of gender, the findings of this study indicate some differences. In this study, the most preferred OCF type of female learners is 'metalinguistic clue'. This preference might indicate that female learners need explanations and clarifications such as metalinguistic clues. This finding is in line with Geçgin's study (2020) where female learners prefer 'metalinguistic clues' as it decreases the anxiety level and leads to self-correction, which is mostly face-saving. This assumption is also supported by Fitriana et al. (2016), who reveal that students preferred metalinguistic feedback as it helps learners activate their background knowledge and think deeper. Therefore, it can be concluded that female learners prefer to think critically under the teacher's guidance. The second choice for female learners is recast, which is the second least favourite preference for male learners. This view is also supported by Amalia et al. (2019) since their study reveals that the female learners assume that recasts make conversation fluent, and does not make students feel anxious but helps students be more confident in developing conversation skills. The least preferred OCF type of females is peer correction contrary to the males whose first choice is peer correction. It is significant to note that some researchers have investigated the EFL learners' OCF preferences and EFL teachers' OCF preferences and have found that teachers do not rely on peer feedback since it may cause the learner to feel humiliated. (Hassan & Arslan, 2018;Kaivanpanah, et al., 2015). Furthermore, female learners' first choice 'metalinguistic clue' is explored as male learners' last choice. This signals female learners' predisposition for metalinguistic clues, unlike male learners. One of the conclusions can be female learners in this study tend to have a concrete sequential thinking style as they prefer to learn by facts, explanations, specific information, and rules (Gregorc, 1985).
In addition to disparities in OCF types according to gender, the difference between grade levels is recognized in terms of preferences and emotions about OCF. Compared to the 9 th and 11 th -grade levels, learners at the 10 th -grade level have a higher tendency to be corrected in class; given immediate feedback, and corrected individually. This might indicate that these learners have more academic concerns. This finding is congruent with the finding revealed by Fadilah et al. (2017). They report that learners pay more attention to the accuracy of their speech as the level increases. This could be explained by the external factors affecting EFL learners' motivation. At the end of the 12 th grade, learners take the university entrance exam, which is a high-stake exam in Turkey and requires an intensive study program. That's why these learners are not included in this study. Similarly, learners in the 11 th grade are engrossed in this intensive study program aiming at achievement in the university entrance exam. To this end, their academic concerns are different from the other grades at school. The other grade level (9 th ) is the first year of high school education, so it can take some time for the learners to get used to the courses, EFL teachers' methodology to teach the language, and the school's approach to foreign language teaching. As 9 th -grade learners' primary concern is to adapt to these issues, it is likely that learners at the 10 th -grade level have the highest mean scores for immediate OCF as they might have higher academic expectations from language courses.
Learners in the 10 th grade prefer clarification requests and repetition of error a lot more than the learners at the 9 th -grade level. This difference might stem from the academic setting at school. As indicated before, learners at the 10 th -grade level the group of learners might have a high degree of academic concerns. Fadilah et al. (2017) support this finding by stating that learners pay more attention to the accuracy of their speech as the level increases. Furthermore, Katayama (2006) makes a different interpretation and reports learners' low preferences for the OCF types clarification requests and repetition of errors are due to their vagueness. Hence, learners do not understand the purpose of the teacher. In their study with tertiary level learners, Fadilah et al. (2017) have found out that freshmen do not think that repetition of errors is an effective type of OCF as it might be confusing as it is not always clear that the teacher is repeating the learner's utterance to highlight the problem or to identify the content; however, second-year learners believe the repetition of error helps them to rethink their own utterances. This also supports the finding which indicates the 10 th grade learners (the second year of high-school education) prefer repetition of error.
Some research investigating the preferences for OCF types in terms of proficiency level indicates that low-level learners tend to learn more about the rules and structures (Kaivanpanah et al., 2015;Philip, 2003) whereas high-level learners are able to recognize their errors and make self-correction when they are mediated with elicitation and other prompts (Lyster & Ranta, 2012). This is a finding supporting 10 th -grade learners' motivation to get clarification requests and repetition of errors in this study, which may also guide them to selfcorrect their errors.
The results of this study indicate that FL teachers should consider gender differences in their curriculum plans and teaching practices. Instructors at schools can carry out some surveys to explore the language learners' OCF preferences and personality traits so that instructors can have an insight into their learners' both attitudes and emotions towards OCF. Teachers can also use drama techniques in the classroom to introduce the OCF types in conversations before they actually provide OCF in real conversational situations. Likewise, language teachers can create a comfortable atmosphere where EFL learners feel responsible for their own learning. Likewise, teachers can utilize activities to decrease learners' worries and anxieties towards making mistakes and being corrected by the teacher in the classroom.

Conclusion
In a conclusion, this research revealed that there were certain differences between the high school EFL learners' preferences and emotions regarding OCF and OCF types in terms of gender and grade level. Receiving OCF was considered to be important for EFL learners, and the majority of the learners preferred to be corrected immediately. Moreover, learners preferred self-correction as the first choice. To this end, language instructors should be aware of individual differences in the class and give corrective feedback with different prompts that address learners' differences so that learners are encouraged to self-correct the errors.
The difference between the female and male EFL learners in terms of their preferences for OCF types was revealed. While female learners preferred metalinguistic clues as to the first choice; male learners preferred peer correction. Hence, teachers might be encouraged to provide feedback with explanations and concrete examples for female learners; and activating peer collaboration for male learners. As for the grade levels, language learners at the 10 th -grade level had more academic concerns compared to the 9 th and 11 th -grade levels. They tend to avoid making mistakes and if an erroneous utterance was generated, they preferred to be corrected immediately and individually.