English Writer 13 | - Freelance Writer
Nov 18, 2017 | #1
The Research Problem
The article reported on research, which was done to investigate and compare sustained versus immediate effects of different corrective feedback types. The feedback types were recasts and metalinguistic feedback. This was done with advanced and intermediate learners of English, who already had knowledge of the proper target form.
The feedback was delivered through synchronous computer-mediated communication (SCMC). The problem addressed by the study is that the precise effects of recasts, and metalinguistic feedback was not established with regard to sustained versus immediate effects delivered through written SCMC with regard to advanced or intermediate learners of English as a second language.
Research Procedures
The study involved 23 volunteers from an undergraduate English translation and grammar course designed for first-year students at Malmo University College in Sweden. The average age of the participants was 24 years old. On average, the subjects had 11 years of English education prior to the study. The majority of participants had Swedish as a native language. However, there were five subjects who had indigenous languages of Spanish, Bosnian, and Arabic. Only one of the subjects studied the majority of their English in a country other than Sweden. The subjects passed English A, which indicates that they were either intermediate or advanced with regard to their English proficiency.There were nine interlocutors who spoke English as their native language from the University Of Pennsylvania Graduate School Of Education who provided feedback to the participants through computer-mediated communications (CMC) which were written. It should be noted that the article uses the full wording of synchronous computer-mediated communication before simply referring to it by its abbreviated form, which is SCMC. However, CMC is used in the article without explanation. This is a breach of standard protocol for scholarly works.
The target form which was used for the study was an English zero article which included abstract non-count nouns. This was found in previous studies to be a challenge for native speakers of Swedish attempting to learn English. However, the errors which occur do not generally lead to a communication breakdown.
The study involved separating groups of students into two conditions of feedback and a single control group. The materials for the experiment consisted of two writing activities, which were computer mediated and collaborative. Participants were paired with the English-speaking interlocutors. There was also a computer delivered judgment regarding acceptability for the pretest, posttest, and delayed posttest. This was done to establish a measurement of the learning which had occurred.
The study was done as a traditional experiment which involved two types of treatment and a control group. The treatment types were recasts and metalinguistic feedback. The dependent variable was the effect of the treatments on the development of L2 grammar by learners of English, who were intermediate or advanced. More specifically, the study used written CMC to deliver the feedback.
A study was done in Japan, which investigated English language proficiency among University students in Japan, which also sought to evaluate the effectiveness of CMC feedback regarding language learning. In relation to the study previously discussed, this study had a significantly larger number of participants. There were 231 students who participated in the study. This is nearly 10 times as many as the study by Sauro. Furthermore, the study by Murphy included learners who were both at higher or lower levels of English proficiency. By having more subjects and testing both the low and high proficiency level students, the study by Murphy can be understood as being exceptional to that of Sauro. However, there was no control group used by Murphy, so in this way the study by Sauro is superior.
A study was recently done exploring the effect that two types of feedback associated with correction had on students learning in a synchronous learning environment. There were 31 subjects who were learning English as a 2nd language. The types of feedback they received during computer-mediated communication tasks were either metalinguistic information, which was explicit or recasts, which were implicit. This was done after they had uttered a sentence which was in error regarding the use of past tense in English. The study used two experimental groups which each received a specific type of feedback. There was also a control group. In this sense, the experiment was of the same basic structure as that done by Sauro. This study lends support to the use of this type of research design for the investigation of the effectiveness of computer mediated corrective feedback with regard to developing L2 grammar as was done by Sauro.
Another study which investigated the effect of feedback on computer-mediated correction of second language learners also investigated metalinguistic versus recast types of feedback. There were 30 subjects who participated in the experiment. Like the study by Sauro, the subjects were split into three groups with two groups receiving feedback of one type been studied and a control group. However, this study was different from the one conducted by Sauro in that the students were between 15 and 17 years old. This is younger than the participants in the study by pasted. This study lends further support to the general question in structure of the study by Sauro.
Research Findings
The study by Sauro discovered that metalinguistic feedback was slightly superior to recasts for the gains in the pretest and immediate posttests for the familiar items which were studied. However, the difference was not statistically significant. This means that the study indicates there is no advantage of using metalinguistic feedback or recasts in relation to each other. The study also discovered that both the recasts and metalinguistic feedbacks were effective methods of increasing the target form knowledge of familiar items. Nevertheless, there was not a significant difference between the two types of feedback.
The study had a number of limitations, one of which was the small sample size. The determination of the sample size is an important factor for any experiment. Experiments generally use samples of the population in order to make inferences regarding a population. This means it is vital that the sample size used in the experiment has enough statistical power in order to determine if it is possible to reject the null hypothesis. If the null hypothesis is false, improperly failing to reject it leads to committing a type II error.
In the case of the study done by Sauro, no statistically significant difference was found between the metalinguistic feedback, and the recasts. However, it is possible that a significantly larger sample size such as the one used by Murphy 2007, using 231 subjects would have led to an important finding. At the same time, it is not a common or proper practice to go back and project what would have happened if more subjects were used through extrapolation of the smaller sample's data.
The article by Sauro would have been more compelling if they had engaged in a Post-hoc power analysis. Power analyses are often done Post-hoc when there is a failure to reject the null hypothesis. The power analysis can be done before or after a study. If it is done before a study, is usually used to determine a sufficient sample size for measuring the effect of interest. A Post-hoc power analysis will indicate the power in the study according to the expected effect size and sample size.
The article does a relatively good job of describing its limitations. For example, it notes that there may have been a significant problem with the sensitivity for the acceptability judgment test. This is because the study used the Blackboard learning system interface which is limited with regard to the types of acceptability judgment items, which can be used. This means that participants in the study were presented with several variations of a sentence and then asked to select the one they thought was most appropriate. Providing the subjects with both unacceptable and acceptable versions of a sentence may have resulted in oversimplification of the decision-making process involved. In other words, the null hypothesis was not rejected because the test was too easy and almost any type of feedback would have allowed better performance.
Another possible problem which is addressed in the limitations section of the paper is that the zero article being used may have been nearly impossible for most learners to generalize. There was an improvement of both treatment groups with regard to the control group only for familiar items. This may be because most of the students engaged in learning, which was item-based instead of rule-based. In other words, the study was too short and only provided a limited amount of feedback to the participants. This may not have been sufficient for them to recall or notice lexical items in the zero article.
A final limitation of the article which is mentioned by the author is that there is overlap between English articles and those in the Swedish system. This could result in confusion with the target form instructions or the conscious raising activities with regard to the article systems of English and Swedish. This has been noted by other studies.
References
Blake, R. J. (2009). The use of technology for second language distance learning. The Modern Language Journal, 93(s1), 822-835.
Li, S. (2010). The Effectiveness of Corrective Feedback in SLA: A Metaâ€Analysis. Language Learning, 60(2), 309-365.
Loewen, S., & Erlam, R. (2006). Corrective feedback in the chat room: An experimental study. Computer Assisted Language Learning, 19(1), 1-14.
Miller, P. C., & Pan, W. (2012). Recasts in the L2 classroom: A meta-analytic review. International Journal of Educational Research.
Murphy, P. (2007). Reading comprehension exercises online: The effects of feedback, proficiency and interaction. Language Learning & Technology, 11(3), 107-129.
Oswald, F. L., & Plonsky, L. (2010). Meta-analysis in second language research: Choices and challenges. Annual Review of Applied Linguistics, 30(1), 85-110.
Plonsky, L., & Gass, S. (2011). Quantitative research methods, study quality, and outcomes: The case of interaction research. Language Learning, 61(2), 325-366.
Razagifard, P., & Rahimpour, M. (2010). The Effect of Computer-Mediated Corrective Feedback on the Development of Second Language Learners' Grammar. Instructional Technology & Distance Learning, 7(5).
Sauro, S. (2009). Computer-mediated corrective feedback and the development of L2 grammar. Language Learning & Technology, 13(1), 96-120.
Yamada, M. (2009). The role of social presence in learner-centered communicative language learning using synchronous computer-mediated communication: Experimental study. Computers & Education, 52(4), 820-833.