Independent project in Educational Linguistics
The effectiveness of oral corrective feedback in experimental and quasi- experimental studies:
A systematic literature review
Author: Attila Czaholi Supervisor: Ibolya Maricic Examiner: Marie Källkvist Term: Spring term 2021 Subject: English V Level: Advanced Course code: 4ENÄ2E
Abstract
By exploring previous research, this systematic literature review aims to shed light on the effectiveness of oral corrective feedback practices by teachers to students on oral proficiency with the focus on recasts and prompts. This systematic literature review also intends to shed light on the effectiveness of recasts and prompts compared with each other. This study reports on the results of 13 experimental or quasi-experimental studies that measured the effectiveness of prompts and recasts on language acquisition and the effectiveness of prompts and recasts compared with each other. In those studies, outcome measures such as grammaticality judgement tasks, oral production tasks, and written production tasks were utilized. The target structures of the included studies were different morpho-syntactical and phonological structures. The accounted findings of this study show that recasts are effective at increasing students’ oral proficiency in seven of thirteen studies and not effective in six of 13 studies. The reported results also
demonstrate that prompts are effective at improving students’ oral proficiency in eight of ten studies and ineffective in two of ten studies. In eight of ten studies in which the effectiveness of prompts and recasts are compared with each other, prompts are more effective than recasts. However, recasts are more effective than prompts in two of ten studies. The reported findings indicate that the effectiveness of corrective feedback in general on language acquisition is uncertain and that prompts might be effective. In addition, the accounted results suggest that the effectiveness of recasts on improving students’ oral proficiency is doubtful. Moreover, the reported findings of this study also suggest the higher degree of efficacy of prompts over recasts.
Key words
Additional language education in ESL/EFL context, effectiveness of oral corrective feedback, experimental/quasi-experimental studies, learners, prompts, recasts, teachers.
Special thanks to
Many thanks to Ibolya Maricic, senior lecturer in English linguistics at Linnaeus University, for her valuable supervision.
Table of Contents
1. Introduction_________________________________________________1 2. Aim and research questions____________________________________3 3. Background and literature review_______________________________3 3.1 Relevance for the English syllabus in upper-secondary level_________________ 4 3.2 Key concepts_______________________________________________________4 3.2.1 Additional language education_____________________________________ 4 3.2.2 Corrective feedback____________________________________________ 5 3.2.3 Long’s Interactional hypothesis___________________________________ 7 3.2.4 Negative evidence_____________________________________________ 8 4. Method_____________________________________________________ 9 4.1 Systematic literature review____________________________________________9 4.2 Data collection procedures_____________________________________________9 4.3 Data description_____________________________________________________12 4.4 Data analysis procedures______________________________________________13 4.5 Reliability and validity_______________________________________________ 16 4.6 Ethical considerations________________________________________________16 4.7 Limitations________________________________________________________ 17 5. Results_____________________________________________________ 18 5.1 Effectiveness of recasts on language acquisition___________________________18 5.2 Effectiveness of prompts on language acquisition__________________________22 5.3 The effectiveness of recasts and prompts compared with each
other_____________________________________________________________ 25 6. Discussion__________________________________________________ 30 6.1 Contexts, methods and participants in the included studies___________________31 6.2 Results discussion___________________________________________________32 6.3 Method discussion___________________________________________________34 6.4 Pedagogical implications______________________________________________35 7. Conclusion__________________________________________________ 36 8. References__________________________________________________ 37 Appendices__________________________________________________I Appendix 1 Coding of the included articles___________________________________I Appendix 2 Questions utilized to scrutinize articles____________________________VI Appendix 3 Studies included in the systematic literature review_________________VII Appendix 4 The effectiveness of recasts on language acquisition__________________XIII Appendix 5 The effectiveness of prompts on language acquisition_________________XXXIV Appendix 6 Effectiveness of recasts and prompts compared with each other_________LIII
1 Introduction
By receiving oral corrective feedback, students may commence using language structures correctly. Thus, oral corrective feedback could increase the students’ oral proficiency. By reporting on how oral corrective feedback could increase the students’
oral proficiency, this systematic literature review could aid teachers and student teachers to enhance their teaching practice. This in turn may enable teachers and student teachers to facilitate their students’ language development.
Feedback from teachers to students in schools occurs in different shapes and forms and with different purposes. Feedback could occur in oral or in written form. The purpose of feedback could be to confirm the student that their utterances were correct. Such types of feedbacks are positive feedbacks (Leeman 2007: 112). One example could be Good job John after John utilized the past tense verb form correctly in sentences such as I kicked the ball yesterday.
Yet, feedback can also have a different purpose. Feedback can make the students aware of the incorrectness in their output by signalling the students’ incorrectness in different ways. Such feedback is called negative or corrective feedbacks (ibid). One example of corrective feedback could be the teacher’s utterance of kicked after a student used the present tense verb form instead of the past tense verb form in his or her utterance I kick the ball yesterday. Such corrective feedback is an example of recasts. Recasts entail that the correct answer is provided for the students without telling them explicitly that their utterances were incorrect (Lyster & Ranta 1997:165).
Another example of corrective feedback is metalinguistic feedback. One example of metalinguistic feedback could be the teacher telling the student You should use the past tense verb form after hearing the student’s incorrect use of the present tense verb form instead of the correct past tense verb form in the sentence I kick the ball yesterday.
Metalinguistic feedback is about providing comments on students’ incorrect utterances without telling them the correct answer (Lyster & Ranta 1997: 166).
Recasts and metalinguistic feedback are two, but not the only, corrective feedback practices teachers might use to facilitate their students’ language development. The use of different corrective feedback practices can make students utilize grammatical or
phonological structures correctly. This in turn could increase the students’ oral proficiency. Therefore, the investigation of the potential effectiveness of corrective feedback and the effectiveness of different corrective feedback practices compared with each other could be important for teachers and student teachers. This study can provide useful information about the possible effectiveness of oral corrective feedback on language acquisition. This study could also potentially provide hints to teachers and student teachers about the degree of effectiveness of different oral corrective feedback practices. By doing so, this study could give suggestions to teachers and student teachers about which corrective feedback practices to use to increase the likelihood of language acquisition among students. Therefore, the study has the potential to inform teachers and student teachers on how they can increase their students’ language proficiency through corrective feedback.
Systematic literature reviews about the effectiveness of oral corrective feedback and the effectiveness of different oral corrective feedback practices in experimental/quasi- experimental studies were made in the past. Lyster & Saito (2010), Lyster et al. (2012) and Oksana’s (2015) systematic literature reviews are examples of such studies. Those studies report on the general effectiveness of oral corrective feedback on language acquisition. The accounted findings of those studies show that participants that received oral corrective feedback generally attained higher results than participants that did not receive any oral corrective feedback in different outcome measures (Lyster & Saito 2010: 280, 283; Lyster et al. 2012: 19; Oksana 2015: 63). In addition, the reported findings of those studies also indicate that prompts as corrective feedback practices are more effective than recasts at enhancing students’ language acquisition (Lyster & Saito 2010: 283; Lyster et al. 2012: 20; Oksana 2015: 64). The essence of recasts, prompts and other corrective feedback practices are explained thoroughly in the key concepts section.
Yet, systematic literature reviews such as those mentioned above report on experimental or quasi-experimental studies conducted in the time range between 1990-2010. No systematic literature review about the effectiveness of oral corrective feedback and the effectiveness of different oral corrective feedback practices in experimental settings based on studies from 2010 and onwards are conducted. This gap justifies the focus of this systematic literature review on studies from 2010 onwards. This systematic
literature review will focus on giving an account of results of experimental/quasi-
experimental studies conducted from 2010 onwards about the efficacy of oral corrective feedback on language acquisition. This study will also focus on reporting on the
efficacy of different corrective feedback practices on language development. The reason for focusing on experimental and quasi-experimental studies is because studies with an experimental or quasi-experimental design are regarded as studies producing results with high evidence value (Petticrew & Roberts 2006: 59).
2 Aim and research questions
Based on previous research on experimental and quasi-experimental studies from 2010 onwards, the purpose of this systematic literature review is to shed light on the potential effectiveness of oral corrective feedback from teachers to students on language
acquisition. Another purpose of this systematic literature review is to shed light on the effectiveness of different corrective feedback practices by teachers to enhance the students’ language development.
Based on the selected research literature, the research questions to be answered are the following:
How effective is oral corrective feedback in additional language education (second and foreign language acquisition)?
What differences are there, if any, between the effectiveness of different oral corrective feedback practices?
3 Background and literature review
The following sections are about the importance of oral corrective feedback by teachers in additional language education, different ways of giving oral corrective feedback, and concepts/theories related to corrective feedback. The focus is on oral corrective
feedback from teachers to students, because the purpose of this study is investigating how teachers’ corrective feedback on students’ errors can promote language
development among students. The reason for only focusing on oral corrective feedback is the limited scope of this study and the lack of experimental studies written from 2010 onwards about the efficacy of written corrective feedback.
3.1 Relevance for the English syllabus at upper-secondary school level
The study is connected to the English syllabus at upper-secondary school level in Sweden. The Swedish National Agency for Education, Skolverket, states that the students at upper-secondary school level are expected to express themselves clearly and fluently in oral productions as one of its knowledge requirements (Skolverket 2011a:
55-57). Oral corrective feedback can facilitate student proficiency in the target
language. By focusing on how oral corrective feedback might improve the students’ oral fluency, this systematic literature review might aid educators to improve their teaching practice and their students’ abilities to produce oral output with increased fluency and accuracy.
Learning to utilize word order or inflection structures correctly, i.e., in accordance with the Standard English variety, can increase the students’ oral proficiency. Standard English can be defined as a variety of English taught in formal school settings with its own set of regularities related to language aspects such as syntax, morphology, lexis, or phonology. Examples of Standard English varieties are, for instance, American
Standard English or Standard British English (Trudgill 1999: 119).
Thus, this study with its focus on the potential effectiveness of oral corrective feedback and the potential variation among the degree of efficacy of the different oral corrective feedback practices can provide useful knowledge about the efficacy of oral corrective feedback for teachers/student teachers. This study can inform teachers and student teachers about which oral corrective feedback practices are more effective to enhance students’ oral proficiency. Therefore, this study can aid educators to use the more effective corrective feedback practices to increase the probability of their students attaining the above-mentioned knowledge requirements.
3.2 Key concepts
3.2.1 Additional language education
Additional language education entails both second language education and foreign language education. There is a distinction between second language education and
foreign language education. Second language education means that the language
education occurs in a country where the target language is widely spoken in a society in general. Students utilize the target language both in school and outside school. On the other hand, foreign language education entails that language education occurs in a country where the target language is not widely spoken. In such cases, students’ use of the target language may be limited to school setting. The similarity between second and foreign language education is that both second and foreign language education entails learning a language which is not the students’ native language (Hammarberg 2013: 28).
3.2.2 Corrective feedback
Corrective feedback entails providing feedback on students’ errors. It is distinguished from the concept of feedback which entails a reaction, either positive or negative, to students’ output and positive feedback, i.e., a positive reinforcement given following correct language use (Leeman 2007: 112). By receiving corrective feedback, students can increase their declarative and procedural knowledge. Declarative knowledge is about the students understanding the rules of linguistic features. Procedural knowledge means students utilizing linguistic features in their own oral and written productions unconsciously (Bialystok 1981: 34).
Corrective feedback can be provided in the following ways according to Lyster & Ranta (1997: 165-167).
Explicit correctness: means that students are told explicitly that their utterances were incorrect, and a correct reformulation is provided by the teacher. For example: Student:
I was very excited under the lesson. Teacher: No John, I want you to say I was very excited during the lesson.
Recasts: means that the teacher provides a correct answer to a student’s incorrect utterance without stating explicitly that the utterance was incorrect. For example:
Student: I was very excited under the lesson. Teacher: I was very excited during the lesson.
Clarification requests: students are asked to clarify what they mean without telling them explicitly that they have made an incorrect utterance. For example: Student: Why does
he fly to Korea yesterday? Teacher: Pardon? Student. Why did he fly to Korea yesterday?
Metalinguistic feedback: comments or questions related to the accuracy of the production of the student. For example: Student: I worked yesterday and visit my
grandmother. Teacher: You should use the past tense verb form of the verb visit because the two verbs are in the same sentence. Student: I worked yesterday and visited my grandmother.
Elicitation: the student is encouraged to come up with the correct answer and express the correct answer with the aid of the teacher. For example: Student: Once upon a time, there lives a poor girl named Cinderella. Teacher: Once upon a time, there. Student:
Once upon a time, there lived a poor girl named Cinderella.
Repetition: the teacher repeats the students’ utterances, including the errors made by the students to provoke the correct answer. Student: Eric work many hours last month.
Teacher: Eric work many hours last month? Student: Eric worked many hours last month.
The latter four types of corrective feedback are also called prompts. Prompts are feedback practices that students receive which promote the students to produce revised/self-corrected utterances (Lyster & Saito 2010: 270).
Prompts are examples of scaffolded feedback. Scaffolded feedback entails that the students are provided with scaffolding in form of hints from their teachers or other individuals to solve their linguistic problems (Ellis 2009: 12). The concept of
scaffolding is connected to the Sociocultural Theory of language acquisition. According to the Sociocultural Theory, scaffolding is a collaborative process that enables the students to accomplish tasks that lie beyond their current levels with the aid of more competent individuals such as teachers (Vygotsky 1978: 86). Scaffolded feedback is adhered to each student’s individual knowledge and needs and is provided to each student differently based on their needs and knowledge (Aljaafreh & Lantolf 1994:
472).
Prompts are also dynamic feedback practices. Dynamic feedback is characterized by the following: It should be provided incrementally depending on the students’ knowledge and it should be contingent on students’ errors (provided when help is needed and withdrawn when the student is able to work independently). In addition, it should be negotiated via hints or other cues to promote the students to come up with the correct answer to their errors (ibid: 470). Consequently, explicit correction and recasts are non- dynamic feedback practices (ibid).
All of the corrective feedback practices can be provided in an incidental/unfocused manner or intensive/focused manner. Providing corrective feedback in an
incidental/unfocused manner entails that the provided corrective feedback is adjusted to address different kind of errors produced by each student. In such cases, the focus of the corrective feedback is not limited to a particular structure or a limited number of
structures during the lesson. However, intensive/focused corrective feedback entails that the provision of corrective feedback is concentrated on a particular structure or a limited number of structures and is provided frequently during lessons. In such cases, the
provision of corrective feedback on structures which are not the focal points of the lesson is neglected (Ellis 2006: 20).
3.2.3 Long’s Interaction Hypothesis
In this study, Long’s Interaction Hypothesis is utilized, since the focus of the study is to investigate how oral interaction between teachers and students could enhance language acquisition among students. Long’s Interaction Hypothesis means that students
participate in interaction with native or non-native speakers. In such interactions, negotiation of meaning occurs. Such interactions could, for instance, be informal conversations or language learning games/tasks rich in meaning negotiation. Meaning negotiation in interaction enables language acquisition for the students, because the students notice the incorrectness in their utterances during interaction with other language users. By doing so, students may commence utilizing linguistic features accurately (Long 1981: 275).
By participating in interactions that contain meaning negotiation, the students can be provided with comprehensible input that could facilitate second/foreign language acquisition. According to Krashen’s Input Hypothesis, language learners can develop further in their language development if they are provided with comprehensible input (Krashen 1985: 2).
In addition, students can notice the incorrectness in their utterances while participating in interaction and providing output. Students could receive feedback about their incorrect utterances from teachers. This in turn could lead to students modifying their output towards a more accurate output and enhance their language development in the process according to Swain’s Output Hypothesis (Swain & Lapkin 1995: 372).
3.2.4 Negative Evidence
By participating in interactions with meaning negotiations, negative evidence can occur.
Negative evidence means feedback on the students’ output which demonstrates or suggests the difference between the students’ incorrect use of a structure and the correct use of the same structure. Thus, negative evidence during interactions can increase the acquisition of the students’ syntax, morphology, or vocabulary in their additional language education (Long 1996: 414).
Since negative evidence can facilitate students to notice the gap between their incorrect utterances and the correct/target-like utterances, negative evidence can cause self- correction (Schmidt 1990: 148). Negative evidence can also lead to uptake among students. Uptake means that the students commence utilizing structures correctly in their production after receiving corrective feedback from teachers about their inaccurate utterances (Lyster & Ranta 1997: 49).
To summarize, teachers providing corrective feedback on their students’ incorrect output in meaning negotiation during interactions can enhance the students’ language development in additional language education. Corrective feedback can be given in many ways and can provide learners with comprehensible input in interactions
containing meaning negotiation. Corrective feedback can entice students to reformulate their incorrect utterances and produce correct utterances instead by making the students
notice the incorrectness in their output. Because of this, the students may start using linguistic features correctly after the provision of corrective feedback.
4 Method
4.1 Systematic literature review
This study was conducted as a systematic literature review. The purpose of a systematic literature review is to review and report on available studies about a particular topic with the aim of answering the proposed research questions (Denscombe 2014: 132). The process of conducting a systematic literature review includes the following steps:
Identify an area of interest and select keywords, determine inclusion/exclusion criteria, perform a search in relevant databases, choose relevant studies, read abstracts of such studies and thereafter, read entire articles. Afterwards, an evaluation of the articles should be conducted to determine which of the articles will be included in the
systematic literature review (Eriksson Barajas et al. 2013: 83). After that, the researcher reaches a conclusion based on the analysis of the included studies (Denscombe 2014:
133). The systematic literature review is a good option for researchers who intend to review and report on results of previous studies about well-researched and described topics (Denscombe 2014: 141).
While conducting systematic literature reviews, researchers have to determine and limit the scope of the study. Researchers are expected to demonstrate transparency while searching for studies to avoid bias, make judgements about the inclusion/exclusion of studies and present a list of the utilized sources found in the study. In addition, the researchers have to compile, analyse and draw conclusions based on the analysis of previous research (Denscombe 2014: 135-140).
It is also imperative that researchers give reasons for why studies were included or excluded and how they were analyzed (Eriksson Barajas et al. 2013: 49).
4.2 Data collection procedures
After determining the area of interest, the researcher consulted with a librarian of the Linnaeus University. During the consultation, OneSearch as search engine was
recommended to find relevant articles. Via OneSearch, I searched for scientific peer- reviewed articles in databases such as ERIC or Cambridge Core.
The keywords that were utilized during the search process were the following: explicit correction, recast, metalinguistic feedback, elicitation, repetition, clarification request, teacher and language acquisition. That was because this study focused on the
effectiveness of explicit correction, recast, metalinguistic feedback, elicitation, repetition, and clarification request as corrective feedback practices. The reason for focusing on the efficacy of those corrective feedback practices was because they were utilized in many studies such as in Lyster & Ranta (1997), Lyster (2004), Ellis (2009) or Long (2006). By using frequently utilized corrective feedback practices as keywords, the probability of finding relevant articles to report on the effectiveness of oral corrective feedback and the efficacy of different corrective feedback practices could increase. The reason for utilizing teacher and language acquisition as keywords was because this study was concerned with corrective feedback by teachers to facilitate students’ language acquisition.
Only peer-reviewed research articles written from 2010 onwards were included in the study. The reason for that was because existing systematic literature reviews based on experimental or quasi-experimental studies about the effectiveness of oral corrective feedback were either conducted before 2010 or consisted of studies made before 2010.
By performing a systematic literature review containing experimental and quasi- experimental studies from 2010 onwards, a research gap could be addressed.
In addition, results from more recent, up-to-date experimental and quasi-experimental studies about the efficacy of oral corrective feedback by teachers may be provided for educators to aid their language teaching. Only studies with experimental or quasi- experimental designs were included in the systematic literature review, since studies with experimental design provide results with high evidence value compared with studies with non-experimental design (Petticrew & Roberts 2006: 59).
Furthermore, only studies about the effectiveness of oral corrective feedback were included in this study because of the lack of experimental or quasi-experimental studies written after 2010 about the efficacy of written corrective feedback. Peer-feedback or perceptions/preferences about corrective feedback were not included in this study either since the focus of this study was on the efficacy of teachers’ oral corrective feedback.
The initial search on OneSearch yielded 52 results/studies. The abstracts of the 52 studies were read to determine which studies will be selected for further investigation based on the inclusion/exclusion criteria mentioned above. After reading the abstracts of 52 articles, 15 articles were chosen for further scrutiny. The included 15 articles were coded using a coding scheme in accordance with Eriksson Barajas et al. (2013: 82).
Therefore, the utilized database/databases where the included studies were found, keywords, number of results/views of the studies and the number of citations of the studies were presented. In addition, the selection criteria to this systematic literature review were also presented in the coding scheme. For more information about the 15 included articles in the coding scheme, see Appendix 1 below.
Thereafter, the 15 articles were scrutinized based on questions found in Appendix 2 below. In order for an article to be included in this study in the material section and be analysed and presented in the results section, all questions in Appendix 2 had to be answered with yes after the scrutiny. The questions in Appendix 2 were adapted from questions utilized by SBU, The Swedish Agency for Health Technology and
Assessment of Social Services (SBU 1997). The reason for utilizing those questions was that such questions were deemed applicable to scrutinize experimental and quasi-
experimental studies. The studies that passed the scrutiny were included in this study for analysis and compilation. After the scrutiny, 13 articles were included in this systematic literature review since they corresponded with the inclusion criteria found in Appendix 2 in addition to being scientific peer-reviewed articles. One article, (Zhai et al. 2018) was excluded because it did not have a control group. Experimental and quasi- experimental studies must contain control groups (Eriksson Barajas et al. (2013: 88) (Denscombe 2014:68).
In addition, because of the limited scope of this study, only studies that measured the effectiveness of recasts and/or one or more corrective feedback practices designated as prompts were included. Consequently, only the effectiveness of prompts and recasts on language acquisition and the efficacy of prompts and recasts compared with each other were accounted for in this study, because prompts and recasts were found to be the most frequently researched corrective feedback practices. Thus, Yucel & Granena (2016) was
also excluded, because it measured the efficacy of explicit and implicit corrective feedback practices.
4.3 Data description
While conducting a systematic literature review, information about the
purpose/perspective and method of the included studies should be presented (Eriksson Barajas et al. 2013: 82). Therefore, the following information about the included studies were presented in Appendix 3 below: the author/publication year/title of the study, purpose of the study, number and level of participants, number of control
groups/intervention groups, and the number of participants in each group. Furthermore, the treatments received by the intervention groups, target structure/skills investigated, outcome pre-test/post-test measures and context/country where the studies were conducted were also presented in Appendix 3.
Experimental and quasi-experimental studies
Studies that were analysed and reported on in this systematic literature review were either experimental or quasi-experimental studies. The purpose of experimental and quasi-experimental studies is to establish whether there is a relationship between different variables. The aim of both experimental and quasi-experimental studies is to investigate and identify cause-effect relationships between independent variables such as corrective feedback practices and dependent factors such as language acquisition.
Independent factors are factors that influence the dependent factors. Dependent factors are factors that change because of the independent factors (Denscombe 2014: 67, 71).
Certain conditions must be fulfilled for a study to be considered as an experimental study. One such condition is that the researcher has to identify factors to include or exclude from the study to determine the potential effects of the chosen factors and decrease the effects of unwanted factors (Eriksson Barajas et al. 2013: 87).
Another condition is that an experimental study has to have at least one intervention group and a control group. An intervention group is a group of participants that received a particular treatment whose effects were measured. The control group is a group of participants who did not receive a particular treatment. The intervention groups and the control groups should be as similar as possible to decrease the risk of the influence of
unwanted factors on the outcome. In experimental studies, the selection and division of participants to intervention groups and control groups occur randomly. Frequently, the participants of the control groups only perform pre-tests and post-tests. Contrary to the control groups, the intervention groups also perform treatment tasks in treatment sessions in addition to performing pre-tests and post-tests. Some experimental studies also include a comparison group or treatment-only group. Comparison/treatment-only groups are participants who perform pre-tests and post-tests but also treatment tasks without receiving a particular treatment (Denscombe 2014: 68). The duration of the treatment sessions in the included studies was 3-8 weeks depending on each study.
Experimental studies also have to contain pre-tests and post-tests. Pre-tests are tests that all participants perform before the treatment has been applied to the intervention group.
Post-tests are tests that all participants do after the treatment has been applied. Both pre- test and post-test measure certain skills or knowledge among the participants with the aim of investigating the effect of certain interventions. By comparing the intervention group’s pre-test and post-test results with each other and with the control group’s pre- test and post-test results, the effects of the treatment may be determined. High post-test score of an intervention group compared with their pre-test score and with a control group’s post-test score indicate the effectiveness of a treatment (Denscombe 2014: 67).
Quasi-experimental studies are similar to experimental studies apart from one aspect.
Unlike in experimental studies, the selection and division of participants to intervention groups and control groups do not occur in a randomised manner in quasi-experimental studies. The reason for that could be that it is not possible to select and divide
participants into intervention groups and control groups randomly (Eriksson Barajas et al. 2013: 89).
4.4 Data analysis procedures
Firstly, the reporting on the effectiveness of different corrective feedback practices on language acquisition was done study by study with focus on one corrective feedback practice or cluster of practices at a time. For example, the potential statistically significant increase/decrease between the mean pre-test/post-test scores of an
intervention group receiving recasts in a study was searched for. That was followed by a
search for the potential statistically significant increase/decrease between the mean pre- test/post-test scores of the control group in the same study. Thereafter, search was conducted about whether the potential difference between the recast intervention group’s and the control group’s mean post-test scores was statistically significant (SS).
A statistically significant increase in the recast group’s mean post-test score compared with their mean pre-test score and the control group’s post-test score together with no increase or no statistically significant increase in the control group’s mean post-test score compared with their mean pre-test score indicated the efficacy of recasts. No increase or no statistically significant increase in the recast group’s mean post-test score compared with their mean pre-test score implied no or doubtful effectiveness for
recasts. Similarly, a statistically significant increase in both the mean post-test scores of the recast group and the control group compared with their mean pre-test scores also indicated no or doubtful efficacy for recasts. ANOVA measures were used in all of the included studies to determine whether the potential increase between the pre-test and post-test scores was statistically significant, i.e., was not caused by chance. ANOVA entails different statistical calculations to use to measure the likelihood of whether the potential differences within groups and between groups regarding mean score were caused by chance (Box et al. 2005: 133). For statistical significance, P value should not be higher than 5, i.e., P = .005. A P value of 5 entails that the probability of chance influencing the data is less than 5 percent (Box et al. 2005: 137). Since ANOVA measures were utilized in the included studies, this study was able to report on findings in which statistical measures were utilized to measure the likelihood of the degree of effectiveness of corrective feedback on language acquisition.
Thereafter, the same procedure was repeated in other studies in which the effectiveness of recasts was investigated. This made it possible to give an account of the efficacy of recasts on language acquisition. Afterwards, the same procedure was utilized in relation to the effectiveness of other corrective feedback practices.
Searching for the potential effectiveness of each corrective feedback practice/cluster of practices found in the included studies enabled the researcher to report on the general effectiveness of oral corrective feedback on additional language education. By doing so, the first research question might be answered.
Secondly, the potential statistically significant increase/decrease between the mean pre- test/post-test scores of an intervention group receiving a corrective feedback practice in a study was searched for. Thereafter, the potential statistically significant
increase/decrease between the mean pre-test/post-test scores of another intervention group receiving another corrective feedback practice in the same study was searched for. Afterwards, information was searched for about whether there is a statistically significant difference between different intervention groups’ post-test scores. The reason for that was to give an account of which corrective feedback practices are more effective at increasing language acquisition. For example, a recast group attained a higher, statistically significant mean post-test score in a study compared with their mean pre-test score and the mean post-test score of an elicitation group in the same study.
That indicated the effectiveness of recasts over elicitation in that study. The search for the mean pre-test/post-test scores of different intervention groups receiving different corrective feedback practices was also done study by study.
The above-mentioned data analysis procedure enabled the researcher to report on the efficacy of different corrective feedback practices on language acquisition compared with each other. By doing so, the second research question could be answered.
The intervention and control groups’ pre-test/post-test scores were the results of how accurately they utilized language features in different outcome measures before and after a treatment/corrective feedback practice was applied.
Lastly, the effectiveness of corrective feedback and the effectiveness of different corrective feedback practices on language acquisition were related to theories and concepts mentioned in section 3 to provide potential explanations for the accounted results of this study.
4.5 Reliability and validity
The results reported in this study were based on peer-reviewed articles which yielded a certain reliability to the reported results of this study. That is because peer-reviewed research articles were scrutinized by experts within a certain research field before such articles were published (San Diego State University 2020). In most of the included studies, inter-rater reliability was used. Inter-rater reliability means that two or more raters analyzed the same data to investigate if, and if so, how their data analysis results correspond with each other (Thornberg & Jönsson 2015: 185). In all of the studies in which inter-rater reliability was measured, inter-rater reliability was above 80 %. The inclusion of studies with inter-rater reliability to a systematic literature review increases the reliability of the systematic literature review itself (Eriksson Barajas et al. 2013:
103-104).
Internal validity is about whether a study actually carried out what it intended to carry out (Skolverket 2011b: 92). This study aimed to report on the efficacy of oral corrective feedback by teachers and the efficacy of different corrective feedback practices on language acquisition compared with each other. This study gave an account of the mean pre-test/post-test results of the control groups and intervention groups receiving
different corrective feedback practices in the included studies. Thus, the two research questions could be answered which in turn yielded internal validity for this study.
External validity is about whether the results of a study can be generalized to contexts outside the study (Eriksson Barajas et al. 2013: 100). This systematic literature review was not able to generalize its reported findings because of the limited number of included studies and linguistic features whose utilizations were measured in the
included studies. Yet, this systematic literature review could indicate tendencies related to the effectiveness of oral corrective feedback in general and prompts and recasts in particular.
4.6 Ethical considerations
This conducted systematic literature review adhered to ethical rules that a research is required to follow according to Vetenskapsrådet, the Swedish Research Council (Vetenskapsrådet 2017). Being transparent and sincere while reporting about the
methods, results, and purposes of the study, being fair while judging others research and
only using the research data in the context of the conducted research (Vetenskapsrådet 2017: 10) were examples of ethical rules that were followed. In addition, accurate referencing to the included articles was utilized. The included articles of this study also adhered to the ethical requirements of conducting research, since the included studies provided, for example, truthful and transparent descriptions of their data collection/data analysis procedures and results in addition to utilizing accurate citations to other
sources.
4.7 Limitations
This study had some limitations. One limitation was that intensive/focused corrective feedback was utilized from teachers to learners in all studies except for Rahimi & Zang (2013). Intensive/focused corrective feedback entails that extensive feedback was provided to correct students’ errors related to a single structure or a limited number of structures during the entire lesson (Ellis 2006: 20). This deviated from many
instructional contexts in additional language education in which incidental/unfocused corrective feedback is utilized. Incidental/unfocused corrective feedback is adjusted to each student’s errors and is not limited to only addressing errors related to a single or a limited number of structures (ibid.). However, this study could provide knowledge for teachers and student teachers about which corrective feedback practices are more effective at facilitating students’ language acquisition. Thus, educators could utilize the reported findings of this study while providing corrective feedback regardless of whether the feedback is given in an intensive/focused manner or incidental/unfocused manner. By doing so, their teaching practice can improve which in turn can increase the likelihood of enhancing their students’ language proficiency as well.
Another limitation of this study was that it consisted of studies in which corrective feedback was provided either on students’ morpho-syntactic or phonological errors.
Therefore, the effectiveness of corrective feedback to enhance other areas of language acquisition such as semantic development was not discussed. Yet, corrective feedback practices could probably be utilized to enhance students’ semantic knowledge too, because the use of corrective feedback is not limited to only correcting students’
morpho-syntactical or phonological errors. Corrective feedback is about providing feedback on learners’ errors in general (Leeman 2007: 112). Thus, the reported results of this study could also be relevant to correct and enhance other areas of language
acquisition too. Teachers and student teachers could choose the more effective corrective feedback practices to correct their students’ semantic errors as well as morpho-syntactical and phonological errors. By doing so, they might enhance not only their students’ morpho-syntactic or phonological knowledge but their semantic
knowledge too.
5 Results
In this section, information is reported on the basis of the included studies on the effectiveness of recasts and prompts on language acquisition with the focus on one corrective feedback practice at a time. This is done by presenting information about the potential statistically significant increase (SSI) between the intervention groups’ mean pre-test and post-test scores, and the control groups’ mean pre-test and post-test scores.
Furthermore, data is also provided about the possible statistically noteworthy increase in intervention groups’ mean post-test scores in relation to the control groups’ mean post- test scores. The effectiveness of recasts and prompts are presented in their own
subsections.
Thereafter, information is shown about the effectiveness of recasts and prompts in comparison with each other. This is done by providing data about the potential
statistically noteworthy improvement between the recast groups’ average pre-test/post- test results, and the prompts groups’ mean pre-test/post-test scores. Additionally, the potential statistically significant differences between the prompts groups’ post-test results and the recast groups’ post-test results are also presented. The effectiveness of recasts and prompts compared with each other is also presented in its own subsection.
5.1 Effectiveness of recasts on language acquisition
In this section, information is shown about the target structures whose utilization was investigated, outcome measures utilized to estimate the efficacy of recasts on language acquisition, and the potential statistically significant increase (SSI) between the recast groups’ mean post-test score and their mean pre-test score. The potential statistically significant increase between the control groups’ mean post-test score compared with their mean pre-test score and the recast groups’ mean post-test score compared with the
control groups’ mean post-test score are also shown below. Furthermore, data is presented about the potential effectiveness of recasts study by study.
Table 1 below demonstrates studies in which recasts were found effective at enhancing students’ language development. For more details about the exact numbers of pre- test/post-test scores, see Appendix 4 below.
Table 1.Studies in which recasts were found effective
Study Target
structure
Outcome measure
SSI between the recast group’s mean pre-test/post- test scores
SSI between the control group’s mean pre-test/post- test scores
SSI for the recast group’s mean post- test score over the control group’s post-test score
Recasts are effective
Rassaei, (2017)
English indefinite and definite articles
Oral production task,
grammaticality judgement task
Yes P<.001
No increase P = .099
Yes P<.001
Yes
Goo (2012) “That” trace filter
Grammaticality judgement test, written
production task Yes P<.001
No P = .099
Yes P<.001
Yes
Révész, Andrea (2012)
Past progressive tense
Grammaticality judgement task, written
production task Yes P<.001
No increase P = .099
Yes P<.001
Yes
Rahimi &
Zang (2013)
Use of morpho- syntactic structures such as subject-verb agreement, superlative adjective forms or linking words
Oral production task,
grammaticality judgement task
Yes P<.001
No increase P = 0.99
No Yes
Van De Guchte et al.
(2015)
Dative case after a preposition of place
Written accuracy task, oral accuracy/
fluency tasks
Yes P<.001
No P = 0.99
Yes P<.001
Yes
Comparative adjective forms
Yes P<.001
No increase P = 0.99
Yes P<.001
Yes
Fatemi &
Harati (2014)
Subject-verb agreement
Oral production task
Yes P<.001
No P = .051
Yes
P<.001
Yes
Rassaei (2019)
Wh questions Oral production task
Yes P<.001
No P = .052
Yes P<.001
Yes
Written production task
Yes P<.001
No P = .051
Yes P<.001
Yes
As Table 1 shows, seven studies (Rassaei 2017; Goo 2012; Révész 2012; Rahimi &
Zang 2013; Van de Guchte et al. 2015; Fatemi & Harati 2014; Rassaei 2019) found the effectiveness of recasts on language acquisition. Such studies showed that the recast
groups attained higher, statistically significant mean post-test scores compared with their mean pre-test scores and the control/comparison groups’ mean post-test scores.
The control groups/comparison groups either did not achieve higher mean post-test scores in comparison with their pre-test results or attained higher, but not statistically significant post-test scores in relation to their pre-test results. The outcome measures in those studies were grammaticality judgement tasks (Rassaei 2017; Goo 2012;Révész 2012; Rahimi & Zang 2013), written production tasks (Goo 2012; Révész 2012; Van de Guchte et al.2015; Rassaei 2019) and oral production tasks (Rassaei 2017; Van de Guchte et al. 2015; Rahimi & Zang 2013; Fatemi & Harati 2014; Rassaei 2019).
The utilized linguistic structures were English indefinite and definite articles (Rassaei 2017), “that” trace filter/redundance of the use of “that” (Goo 2012), past progressive tense (Révész 2012), and the use of different morphosyntactic structures (Rahimi &
Zang 2013). In addition, the dative case after preposition of place, comparative
adjective forms (Van de Guchte et al. 2015), subject-verb agreement (Fatemi & Harati 2014), and wh questions (Rassaei 2019) as linguistic features were also used.
However, recasts were not effective at increasing language acquisition in six of 13 studies. Table 2 below shows studies in which recasts were not found effective.
Table 2.Studies in which recasts were not found effective Study Target
structure
Outcome measure
SSI between the recast group’s mean pre-test/post- test scores
SSI between the control group’s mean pre-test/post- test scores
SSI for the recast group’s mean post- test score over the control group’s mean post-test score
Recasts are effective
Rassaei, (2014)
Wh questions
Grammaticality judgement task, oral production task
No increase P = .099
No increase P = .099
No P = .099
No
Kartchava
& Ammar (2014)
Past tense verb form
Oral production task
No increase P = .099
No increase P = .099
No P = .099
No
Questions in the past
Oral production task
Yes P<.001
No increase P = .099
Yes P<.001
Yes
Naeimi et al. (2018)
[w] sound, lax vowel [i], consonant clusters with added vowel
Oral production task
No P = .050
No increase P = .099
No P = .050
No
Karimi &
Esfandiari (2016).
Oral stress pattern
Oral production task
No increase P = .099
No P = .099
No P = .099
No
Yang, &
Lyster (2010)
Irregular past tense verb form
Oral production task
No P = .082
No P = .082
No P = 0.99
No
Irregular past tense verb form
Written production task
No P = .0.75
No increase P = .099
No P = .0.75
No
Regular past tense verb form
No P = .045
No increase P = .099
No P = .045
No
Regular past tense verb form
No P = .088
No increase P = .099
No P = .084
No
Li et al.
(2016)
Regular verbs
Gramaticality judgement task
No P = .050
No P = .047
No P = .098
No
Oral production task
No P = .099
No P = .099
No P = .099
No
Irregular verbs
No P = .050
No P = .050
No P = .099
No
No P = .099
No P = .099
No P = .099
No
Table 2 shows that six of 13 studies (Rassaei 2014; Kartchava & Ammar 2014; Naeimi et al. 2018; Karimi & Esfandiari 2016; Yang & Lyster 2010; Li et al. 2016)
demonstrated no effect of recasts on language acquistion. That was because the recast groups did not reach higher, statistically noteworthy post-test scores compared with their pre-test results in grammaticality judgement tasks (Rassaei 2014; Li et al. 2016), or
written production tasks (Yang & Lyster 2010). Moreover, the recast groups did not gain higher post-test results in oral production tasks either (Rassaei 2014; Kartchava &
Ammar 2014; Naeimi et al. 2018, Karimi & Esfandieri 2016; Li et al. 2016; Yang &
Lyster 2010).
The structures used and which were evaluated in those studies were morpho-syntactic structures such as wh questions (Rassaei 2014), past tense verb form, questions in the past (Kartchava & Ammar 2014), regular and irregular verbs (Li et al. 2016), and irregular and regular past tense verb forms (Yang & Lyster 2010). In addition, the utilization of phonological linguistic structures such as [w] sound, lax vowel [i], consonant clusters with added vowel (Naeimi et al. 2018), and oral stress pattern (Karimi & Esfandieri 2016) were also estimated.
The accounted results of this systematic literature review found that recasts are effective in seven studies while six studies show that they are not effective. Thus, the reported findings regarding the effectiveness of recasts are mixed. This indicates that the efficacy of recasts on language acquisition is uncertain.
5.2 Effectiveness of prompts on language acquisition
In this section, data is shown regarding the linguistic structures whose use was measured, result indicators to estimate the effectiveness of prompts on language acquisition, and the potential statistically significant increase between the prompts groups’ average post-test scores and their average pre-test scores. The potential
statistically significant increase between the control groups’ mean post-test results and their mean pre-test results and the prompts groups’ mean post-test results compared with the control groups’ mean post-test results are presented too below. Additionally, data is shown about the potential efficacy of prompts study by study.
Table 3 below shows eight studies which demonstrate the effectiveness of prompts on language development. To read the exact numbers of pre-test/post-test scores, see Appendix 5 below.
Table 3. Studies in which prompts were found effective
Study Target
structure
Outcome measure SSI between the prompts group’s mean pre- test/post-test scores
SSI between the control group’s mean pre- test/post-test scores
SSI for the prompts group’s post-test scores over the control group’s post-test scores
Prompts are effective
Rassaei (2014) Wh questions
Grammaticality judgement task, oral production task
Yes P<.001
No P = .099
Yes P<.001
Yes
Goo (2012) “That” trace filter
Grammaticality judgement test, written production task
Yes P<.001
No P = .099
Yes P<.001
Yes
Kartchava &
Ammar (2014)
Past tense verb form
Oral production task
Yes P<.001
No increase P = .099
Yes P<.001
Yes
Questions in the past
Oral production task
Yes P<.001
No increase P = .099
Yes P<.001
Yes
Van De Guchte et al (2015)
Dative case after a preposition of place
Written accuracy task, oral
accuracy/ fluency tasks
Yes P<.001
No P = .099
Yes P<.001
Yes
Comparative adjective forms
Yes P<.001
No increase P = 0.99
Yes P<.001
Yes
Yang & Lyster (2010)
Irregular past tense verb form
Oral production task
Yes P<.001
No P = .082
Yes P<.001
Yes
Irregular past tense form
Written production task
Yes P<.001
No increase P = .099
Yes P<.001
Yes
Regular past tense verb form
Yes P<.001
No increase P = .099
Yes P<.001
Yes
Regular past tense verb form
Yes P<.001
No increase P = .099
Yes P<.001
Yes
Naeimi et al.
(2018)
[w] sound, lax vowel [i], consonant clusters with added vowel
Oral production task
Yes P<.001
No increase P = .099
Yes P<.001
Yes
Li et al. (2016) Regular verbs
Gramaticality judgement task
Yes P<.001
No P = .047
Yes P<.001
Yes
Oral production task
Yes P<.001
No P = .099
Yes P<.001
Yes
Irregular verbs
Yes P<.001
No P = .050
Yes P<.001
Yes
Yes P<.001
No P = .099
Yes P<.001
Yes
Rassaei (2019) Wh questions
Oral production task
Yes P<.001
No P = .052
Yes P<.001
Yes
Written production task
Yes P<.001
No P = .051
Yes P<.001
Yes
Table 3 shows that eight studies (Rassaei 2014; Goo 2012; Kartchava & Ammar 2014;
Van de Guchte et al. 2015; Yang, Lyster 2010; Naeimi et al. 2018; Li et al. 2016;
Rassaei 2019) found the efficacy of prompts on language development. In those studies, the prompts groups reached higher, statistically significant average post-test scores compared with their average pre-test results and the control groups’ average post-test results. However, the control groups either reached higher, but not statistically
noteworthy mean post-test results compared with their mean pre-test results or did not attain higher average post-test scores at all. The result indicators in those studies were grammaticality judgement tasks (Rassaei 2014; Goo 2012; Li et al. 2016), or written production tasks (Goo 2012; Van de Guchte et al. 2015; Yang & Lyster 2010; Rassaei 2019). Other studies applied oral production tasks as result indicators (Rassaei 2014;
Kartchava & Ammar 2014; Van de Guchte et al. 2015; Yang & Lyster 2010; Naeimi et al. 2018; Li et al. 2016; Rassaei 2019).
In studies found in Table 3, the use of wh questions (Rassaei 2014; 2019), “that” trace filter (Goo 2012), past tense verb forms, questions in the past (Kartchava & Ammar 2014), comparative adjective forms, and dative case after preposition of place (Van de Guchte et al. 2015) were evaluated. Furthermore, the learners’ use of irregular and regular past tense verb form (Yang & Lyster 2010), [w] sound, lax vowel [i], consonant clusters with added vowel (Naeimi et al. 2018), and regular and irregular verbs (Li et al.
2016) were also measured.
Nonetheless, prompts were ineffective at improving language proficiency in two studies as seen in Table 4 below.
