The Effect of Explicit Feedback in Oral Performance in Use of Past Tense

The correction of grammar was said not to be facilitating in the language learning process by some researchers, including Krashen (1982, 1985) because it might affect learners negatively. However, the feedback has been studied in terms of its influence on language acquisition, and many studies point the positive effect of feedback on language learning (Ellis, Loewen, & Erlam, 2006; Khanlarzadeh & Nemati, 2016; Russel & Spada, 2006). Concordantly, this study examined the effect of corrective feedback on performance in speaking tasks targeting the use of simple past tense. Since many studies on corrective feedback (CF) focused on feedback on writing, especially in regard to grammatical errors, the present study aimed to contribute to the existing prose with the focus of oral performance and speaking. The participants were first-year students at a private university in Istanbul. Since intact classrooms were used to create samples, a quasi-experimental design was applied. A pre-test was applied to identify their current competence of the target topic in terms of oral production. The control and experimental groups were assigned randomly. The experimental group received explicit feedback, while the control group was not exposed to any kind of feedback. Verbal feedback was provided with regard to grammatical error correction. Pretest and post-test results of control and experimental groups which were examined indicated that the experimental group receiving feedback for six weeks outperformed the control group. Based on this finding, corrective feedback can be considered as a facilitating tool for speaking activities with a grammar focus.


LITERATURE REVIEW
Corrective feedback (CF) has been a conspicuous notion in L2 learning and teaching as it has a vital and facilitating role in students' perception of their learning process and teachers' attitude towards error correction. In parallel to this, many SLA theories consisting of Schmidt's Noticing Hypothesis (1990), Swain's Output Hypothesis (1995) and Long's Interaction Hypothesis (1996) have provided a strong theoretical background for the use of CF in language teaching. CF has been associated with Schmidt's Noticing Hypothesis (1990) in which it is suggested that learners can make a comparison of actual utterances and their current production with the help of noticing, and CF may play a facilitating role for noticing (Kim, 2004). As for CF, Swain (1995) puts forward that modified output results from feedback which is crucial in language learning, and Long (1996) addresses in Interaction Hypothesis that negative feedback in negotiated interaction may promote second language learning by letting them know their problematic utterances. Ellis (2006) defines corrective feedback as any response given to learners' erroneous production with the intention of correction of learner error. Similarly defined by Russell and Spada (2006), CF includes evidence to learner error of a language form. Since the use of CF allows learners to notice their errors and targeted forms, it has been categorized by researchers according to the degree of its explicitness. As Ellis, et al. (2006) initiated, explicit CF refers to apparent attention to the error, whereas implicit one does not have an explicit indication of error. With the very much alike definitions, another classification of CF, which includes direct and indirect corrective feedback, has been done by Bitchener and Knoch (2009). Elicitation, metalinguistic feedback (explicit correction) and clues, and didactic recasts are explicit corrective feedback types. Implicit ones, on the other hand, are more conversational recasts such as clarification requests, confirmation checks, and reformulation of erroneous utterance (Lyster, et al., 2013).
A number of studies on corrective feedback have investigated the types of CF. As one of the current studies, Ajabshir's study (2014) explores the effect of explicitness and implicitness of CF on pragmatic development of language learners. It is argued that experimental groups which have received implicit and explicit feedback do better in terms of the subcomponents they use and their hesitation duration. Moreover, the group receiving explicit feedback shows better performance than the implicit feedback group in only one component, which can support the idea that explicit feedback helps learners to raise their consciousness level and enhances interlanguage development. Similarly, findings of research on corrective feedback carried out by Gitsaki and Althabaiti (2010) suggest that explicit corrective feedback can enhance learners' awareness of their own mistakes and lead to successful uptake. Nonetheless, it adds that repetition and metalinguistic clues are the most successful feedback types in leading to successful uptake. Therefore, it is suggested that error types and related feedback types should be researched more for classroom applications at different levels and various contexts.

Research on Corrective Feedback in Oral Performance
There has been a rapid change from a functional process of the language to the communicative process (Savignon, 2018). This change also affects error perception. Consequently, as the communicative instruction gained importance in the classrooms, corrective feedback became the subject of discussion, which leads to the discussion on corrective feedback in oral performance. As one of the recent studies, Sarandi's (2017) research explores the effect of mixed oral CF (a combination of recasts and prompts) on L2 learners' oral performance with first-year ELT students in a Turkish university. After completion of five oral tasks, elicited imitation (EI) test and narrative task are used to measure the participants' oral performance accuracy. The findings of the study indicate that mixed oral CF improves learners' oral performance with partially learned structures.
To investigate common error types and corrective feedback frequencies in oral production in EFL classrooms, Jabbari, and Fazilatfar (2012) conduct research with the use of audio-recordings of elementary and high intermediate classes and observations of teachers' preferences for corrective feedback types. According to the interpretation of the data, grammatical error correction (50.5%) is the most common one. However, phonological (26%) and lexical errors (22%) had a lower rank error type. Furthermore, it is also observed that there is a common tendency for instructors to use recasts (50.5%) as the main corrective feedback form. Similarly, Öztürk's (2016) research highlights the corrective feedback preferences, and it is found that there is again a tendency for teachers to use recast as the primary corrective feedback type. However, it has also been observed that there is a significant difference between experienced and novice teachers in terms of corrective feedback usage. According to the data, experienced teachers prefer recasting relatively more while novice teachers use clarification requests frequently. Correspondingly, Kırkgöz, Ağçam , and Babanoğlu (2015) study the corrective feedback types used in EFL context at primary schools in Turkey. It is marked that the teacher used all types of feedback in the classroom, and an explicit correction was the most used type in EFL classrooms. In addition to these findings, the study has found that clarification request, elicitation, metalinguistic feedback , and paralinguistic feedback had more significant and successful results for the students' uptake whereas explicit correction produced less self-repair. Farrokhi and Chehrazad (2012), on the other hand, investigate recast and explicit types of feedback in terms of oral performance and argue that there is no significant difference between them. In their research, no treatment is provided to the control group, whereas the experimental group receives CF in the form of recast and experimental group 2 receives delayed explicit and metalinguistic feedback. As the results indicate, there is a significant difference between the control group and the experimental groups. Experimental groups outperform the control group in terms of correct oral production of the target form. It can be concluded from the data that CF has a positive effect on oral production. However, in the research, no significant data that shows the difference in the effectiveness of recast and explicit feedback has been obtained.

Research on Corrective Feedback with regard to Grammatical Structures
Corrective feedback has been a prominent part of language learning and teaching. Thus, the literature on CF has varied in different features. Some of the studies on CF has focused on the grammar component of language learning. For instance, Ellis, Loewen, and Erlam (2006) probe into the effects of both explicit and implicit feedback on learners' grammatical competence in their study by focusing on regular past tense '-ed' as target structure. The data analysis indicates that the experimental group receiving corrective feedback outperform the other two groups. Furthermore, it is suggested that corrective feedback has an effect on implicit knowledge through the data gained from students' performance on grammatical and ungrammatical sentences.
The research carried out by Khanlarzadeh and Nemati (2016) focuses on the effects of corrective feedback on grammatical accuracy in writing tasks. According to its results, the experimental group which receives corrective feedback gets higher scores than the control group. Besides, there is a significant difference between the pre-and post-test results of the experimental group. Nonetheless, the control group does not perform better in the post-test than they do in the pre-test. Hence, it is claimed that learners benefit from CF in terms of realizing their own grammatical mistakes and correcting them in writing tasks.
Another study is conducted by Daneshvar and Rahimi (2013) to examine the effectiveness of direct, focused feedback and recast on grammatical competence. The post-test results of the research show that the group which receives recasts as feedback displays lower performance than the other two groups. However, results of a delayed post-test which is given to the students to see the long-term effects of different feedback types indicate that the group which is given direct, focused feedback gets lower scores than the other experimental group receiving recast. Thus, it is asserted that recast can be used as a powerful feedback method since its effects last longer, and it may foster implicit learning. As for the effect of different CF types on grammatical structures, Hosseini (2015), similarly, investigates the effectiveness of explicit and implicit feedback types for the use of definite/indefinite articles. Two experimental groups (one is given explicit CF, and the one is provided implicit CF) and one control group are determined.The analysis of the results of pre-and post-tests points that the explicit feedback group showed better performance in use of indefinite article than the implicit feedback group while there is not a significant difference for the use of definite article between three groups in the study.
As seen from the given current discussion of CF in different focal points, there have been various results and findings; correspondingly, conflicting ideas were developed on the effect of CF on language learning and teaching. It is not possible to draw a particular conclusion to this argumentation.

RESEARCH PROBLEM
There have been a number of studies on corrective feedback. However, most of the studies, especially in terms of grammatical accuracy, have been carried out on writing performance or written feedback. As the number of studies on CF regarding oral production has been limited, the present study addresses a gap by searching an answer to the question of whether there is an effect of corrective feedback on oral performance in speaking tasks with the simple past tense target. Based on this research question, it is hypothesized that explicit feedback would affect language learners' oral performance positively in regard to the simple past tense use.

Research Design
The present study aimed to evaluate the effectiveness of corrective feedback on performance in speaking tasks with the simple past target. While doing so, it employed a quasi-experimental design which resembles true experimental design with manipulation of independent variable but lacks random assignment since having a random assignment is difficult and sometimes impossible in an educational setting. Thus the participants of the present study were not randomly assigned to the experimental and control groups, and the treatment was administered to only one of these two groups. The oral performance of the experimental group and control group was measured via recordings and analysis of those recordings before and after the treatment was applied. As for the treatment, the experimental group was provided with explicit feedback on a particular grammatical structure, Past Simple Tense, during the implementation period whereas no kind of feedback was given to the control group in order to explore whether explicit feedback holds any influence on oral performance.

Setting and Participants
The participants were freshman students at a private university in Istanbul. 28 students participated in the research. The control group consisted of 14 students, and the experimental group consisted of 14 students. All the students had an A2 level of English. They were assigned to their classes by a placement test, including grammar, vocabulary, listening, reading, writing, and speaking sections at the beginning of the term. Since the placement exam assesses not only the receptive skills but also the productive skills, the students are assumed to attain an equal level in language skills. The students had 12 hours of Main Course English class in a week.

Data Collection Tools
The study has two groups: one control group and one experimental group. The control group and the experimental group had 14 students. A pre-test was applied to both groups at the beginning. The groups were randomly divided into two groups. They were given a speaking prompt, which required them to talk about their last summer holiday. Their oral performances were recorded. The recordings were transcribed, and the mistakes regarding past simple were highlighted. For the following four weeks, both the control group and the experimental group were given four different speaking prompts for which they had to use the past simple. The experimental group received explicit corrective feedback for their grammatical mistakes, whereas the control group did not get any feedback. These sessions were not recorded. After four weeks, a post-test was implemented. The same process with the pre-test was applied. In other words, the participants were given a speaking prompt, which required them to talk about their last summer holiday again. Their oral performances were recorded as well. The recordings were transcribed, and the mistakes regarding Past Simple Tense were highlighted.

Analysis of the Data
To score learners' pre and post-test performances, their use of the simple past tense during the oral tasks applied in tests was marked either as correct or incorrect. To maintain the evaluation standardization, a strict scoring process which required the participants to use the past simple form of the verbs without any single mistake, was applied. A paired samples t-test was applied to both experimental and control groups separately in order to compare their pre and post-test results and examine whether the use of feedback led to any improvement or not.

RESULTS
A pre-test and a post-test were applied to both the control and the experimental group. In the transcriptions of these tests, the mistakes regarding past simple were highlighted. While analyzing the data, the number of these mistakes has been taken into consideration. The number of mistakes can be seen in more detail in Table 1. When looked at the number of the mistakes, it is seen that the number of the mistakes of the experimental group, which received explicit corrective feedback, has declined; whereas the number of the mistakes of the control group, which did not receive explicit corrective feedback remains similar.
After the number of the mistakes of each group in the study was identified, SPSS was used to get the results of quantitative data in the study. A paired samples t-test was applied to see if there is a statistical significance between pre-test and post-test results of control and experimental groups and the overall significance of the treatment. The descriptive statistics of the paired samples t-test are shown in Table 2. In Table 2, the means for the control group pre-test and post-test scores are 1,57 and 1,64, respectively. The means for the experimental group pre-test and post-test scores are 2,86 and 0,29, respectively. The standard deviation values for the control group pre-test and post-test are 0,76 and 0,63, respectively. The standard deviation values for the experimental group pre-test and post-test are 1,66 and 0,47, respectively. The number of participants in all the conditions (N) is 14.
The results of the paired samples t-test, which was applied to compare the number of mistakes in control and experimental groups, can be seen in Table 3. There was a statistically significant difference in the scores for the experimental group (M=2,57, SD= 1,6); t(13)=6, p = ,0. However, there was not a statistical significance in the scores for the control group (M=-,07, SD=0,6); t(13)=-0,43, p = 0,67.
The research problem of the present study was to identify the effect of explicit feedback on speaking through the use of simple past tense. The results of the paired samples t-test indicate that the experimental group which received explicit feedback for their simple past tense use showed better performance in the post-test recordings, and there seem to be a statistically significant results group. However, there is not a significant difference between pre-test and post-test results of the control group, which may be simply explained by the fact that they did not get any type of feedback through the treatment period. According to the overall results, when the two groups are compared, it can be seen that giving explicit feedback to one of the groups of the participants affected their grammar accuracy positively whereas the other group did not show any noticeable improvement.

DISCUSSION
Explicit corrective feedback is preferred to be used while assessing students' writings or written exams in terms of grammatical accuracy. Teachers may hesitate to give corrective feedback during oral tasks in order not to interrupt the student or impede the flow of the conversation. However, the results have indicated that the experimental group that received explicit feedback showed better performance in the post-test than they did in the pre-test. However, there is not a statistical significance between the pre and post-test results of the control group. Thus, it can be concluded that learners can benefit from explicit corrective feedback during oral grammar tasks. These findings align with the results of Farrokhi and Chehrazad's study (2012) on the effect of different types of CF on oral production. The results indicated CF has a positive effect on oral skills regardless of its implicitness/explicitness. As to grammatical accuracy, the present study showed that students' use of a specific grammatical structure improves when they are provided with explicit CF. As stated in the findings of Ellis, Loewen, and Erlam's study (2006), explicit corrective feedback can help learners to detect their mistakes and raise their awareness about the use of the target structure. This argument show similarities with a recent study in which Khanlarzadeh and Nemati (2016) found out that CF leads to self-correction in the use of grammatical structures. Furthermore; explicit corrective feedback can also facilitate learners' monitoring skills and shows them on which specific points they need more practice on.
It can be concluded that the results obtained from the study support the hypothesis which claims that explicit feedback may facilitate language learners' oral performance in terms of simple past tense use. However, the results cannot be generalized and associated with corrective feedback since there are some other factors which could not be eliminated. The first factor is the fact that the students took 12 hours of English classes in a week apart from the sessions they had for the study. Hence, their regular English classes might have played a role in their oral grammar task performances. Besides, since the post-test was immediately given to the students, the results only reflect the short term effects of the explicit corrective feedback. Hence, a delayed post-test appliance may help us to study the long term effects in depth. Lastly; since the sample is quite small, more studies should be conducted to see whether the results would be similar or not.

CONCLUSION
This study aimed to investigate the potential effects of explicit corrective feedback on oral production with the simple past tense target. Data were collected through recordings, and these recordings were transcribed, and the mistakes were highlighted. After the analysis of the findings, it was revealed that explicit feedback has a significant effect on oral production of the past tense. Concordantly, the null hypothesis was rejected. Although the collected data was in small numbers, the results correspond with those of Russell and Spada (2006). Similarly, in their research, it is argued that with corrective feedback, learners notice their errors and monitor their output accordingly. This study addresses the necessary research gap in the field in terms of the effect of explicit feedback for a particular grammatical structure on oral production. It is suggested that teachers can use corrective feedback to facilitate speaking skills in the classroom environment. However, this study suggests that the findings are limited because of the fact that participants' use of the strategy of 'avoidance' was of likely during the experiment. Therefore, the key findings of this study need to be considered within this possibility. Besides, further research is required as the duration of the experiment was short.