This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
The aim of this chapter is to discuss the findings and to look at potential clarifications for the outcomes. This chapter is divided into four sections. The first section will focus on the results of the groups' pre-tests. The second section will discuss the post-tests' results. Then the D post-tests' results will be considered in the third section. The fourth section will look at the effect of test types on the student's results. In the end, the conclusion will sum up the present research.
The first research hypothesis predicted that CG, RG, and MG, students' performances on the pre-tests in both types of tests (fill-in-the-blank and cloze) were comparable and there were no significant differences among them. In fact, the results support this hypothesis. To review the results, there was no significant difference found among the CG, RG, and MG results on the fill-in-the-blank test (F (2, 46) =.596, p > 0.05) and cloze test (F (2, 46) =.168, p > 0.05). An inspection of the pre-test fill-in-the-blank and cloze tests' scores proposes that all of the students initially face the same difficulty which could be attributed to their limited knowledge of the English articles. This was evident in their inability to produce the correct answers properly which may cause them to get low marks on the pre-tests.
This section discusses the results relevant to the second research hypothesis that is the RG on the post-test in the two types of tests would outperform the CG, whereas the MG students would produce answers at a higher level than RG and CG students would. In fact, the findings confirm this hypothesis and are discussed in this section. To review the results, as reported in Table 11 and Table 23, the Post Hoc contrasts for the post-tests showed that RG on both post-tests performances were more accurate in supplying the target articles than that of the CG (p > 0.05). However, the students in MG significantly outperformed those in RG and CG (p > 0.05).
Regarding the RG students, the results of this study partially support what has been discovered by some researchers that recasts do facilitate L2 learning (e.g. Han, 2002; Ishida, 2004; Sheen, 2008). For instance, as mentioned earlier in the discussion of theoretical issues relating to CF, the results of Han's study showed that the participants who received recasts outperformed the CG in both post-test and D post-test. Yet, the results of the present study indicated that RG only outperformed the CG on just the post-test of each type of test which demonstrates that recasts are effective in helping L2 learners to have a better control over using English articles in the short-term.
In the present study, the corrective metalinguistic caused more gains in comparison with corrective recast. This finding is specifically in line with what Lyster and Ranta (1997), Muranoi (2000), and Ellis et al. (2006) have found that the metalinguistic feedback is more effective than recast in L2 learning. Moreover, this finding can support what has been claimed by some researchers such as Gagne et al. (1993) and Ellis (2005) that is the declarative knowledge in the form of instruction and corrective feedback has played a very important role in triggering this result because as mentioned previously when learners have some difficulty in the use of the L2, they need to know and understand the information in order to remember and retrieve it at a later time when the need arises. Moreover, declarative knowledge could be learned and modified quickly not like procedural knowledge which can only be acquired slowly. However, it could be argued that if this study used communication skill in which students are divided into groups in order to interact with each other, the procedural knowledge in the form of feedback would be more efficient than declarative knowledge in the form of metalinguistic in helping students to supply target-like English articles as suggested by Jiamu (2001).
It can be suggested that there are two possible reasons that may account for the effectiveness of metalinguistic over recast in the present study. First, since this study was carried out with relatively quite large numbers of students 20 students in each experimental group, CF in the form of explicit is preferable than recasts as suggested by Lyster and Ranta (1997) who found it in thier immersion context with relatively large numbers of young learners (in classes of 24-30 students). In fact, the similarities between the present study and Lyster and Ranta's study can be attributed to the number of students. This may suggest that the efficiency of recast could be linked to the small size of class as in Han's who conducted the study on eight students. Nevertheless, since the present study has not looked at the effectiveness of recast over a smaller number of students, the researcher may not be able to argue for or against Lyster and Ranta's claim. In fact, supposing that recast in more efficient in smaller group of students may affect its usability for every class size.
The second reason could be the linguistic target. For example, the effectiveness of metalinguistic over recast in both this study and Muranoi's could be attributed to the similarity of the linguistic target (the English articles). In fact, Muranoi's results suggest that in providing the metalinguistic feedback to L2 learning classes, learning complex rules can be facilitated. Moreover, if the effectiveness of metalinguistic over recast is due to the assumption that as metalinguistic is longer than recast, it might have been better attended to and perceived as overtly CF then this supports Ellis claim but also contradicts (Philp, 2003; Loewen and Philp, 2006; Han and Kim, 2008) claim who found that the shorter the CF, the more likely it is to be noticed.
In fact, it may be difficult to be certain whether to argue for or against the applicability of metalinguistic over recast in teaching all types of linguistic targets. This uncertainty is related to the linguistic target used in the present study. In fact, the researcher acknowledges that if this study had looked at two or more linguistic targets, we would understand better if the efficiency of metalinguistic feedback could be linked to the linguistic target.
On the other hand, the results of the present study contradict what has been found by kim and mathes (2001) who have found that there is no significant between the recast and metalinguistic results on the post-test. However, as mentioned previously, the students in Kim and Mathes's showed preference to metalinguistic feedback.
The third research hypothesis expected that the MG students would advance to higher levels of supplying target-like English articles than CG and RG students on the D post-test. Nevertheless, contrary to expectations, the results of this research did not support this research hypothesis. In fact, when the researcher did a one-way ANOVA on the results of D post-tests in both types of tests, it showed that the students' results in the experimental groups were similar to each other and to the CG and significantly lower than that of the post-test. To review the results, as reported in Table 12, the three groups' Fill-in-the-blank D post-tests performances were quite similar and no significant difference was shown (F (2, 46) =2.025, p > 0.05). On the other hand, Table 24 indicated that the three groups' cloze D post-tests results were comparable and that the difference in mean scores was not statistically significant (F (2, 46) =2.176, p > 0.05).
With regards to the RG students performances presented in this research, there is a contradiction between them and what has been found by Han (2002) who claimed that recast is not only effective in helping L2 learners to acquire English articles in the short-term but also in the long term. In fact, Han's study was conducted in laboratory setting which could be different than classroom setting in which recasts may pass unnoticed which in turn may not help students to overcome their errors as claimed by (e.g. Ammar and Spada, 2006) and (Loewen and Philp, 2006). Moreover, Ellis (2006)argues against the findings from laboratory studies since they involve one-on-one interactions not like classrooms in which the teacher interacts with the whole class.
Regarding the MG results, surprisingly their results did not improve from their results on the post-test. In fact, it was found that their performance was not different from that on the pre-test or from the RG and CG D post-test. Although the post-test results of this study contradict the results of Kim and Mathes (2001), it provides a strong support to what has been found by them that there is no significant different between MG and RG on the D post-test. However, this is on the contrary to what has been found by other researchers, who argue that the advantages of the metalinguistic feedback become more evident as time passed (Muranoi, 2000; Ellis et al., 2006). In fact, the contradiction here could be attributed to a number of reasons. First, the D post-test in the present study was carried out six weeks after the feedback session, while the participants in other studies such as Ellis's study were D post-tested only two weeks after the feedback sessions and after five weeks in Muranoi's study. Both of these are significantly shorter than our six weeks interval.
The linguistic target issue used in this study will be discussed once again in terms of its difficulty. It could be argued that this contradiction between the present study and others may be a result of the participants' mastery of English language, especially English articles due to two reasons: the duration of their studying English and their article system. For example, as mentioned in Ellis study, the participants' average length of studying English was 7 years, in contrast, our participants' was only 3 years. Therefore, their mastery of English articles could be more accurate than ours'. Regarding the article system, as mentioned before, many L2 learners of English, especially Arabic speakers, face many difficulties in using articles, as claimed by some researchers who found that Arabic speakers find English articles one of the most difficult areas in the English language. Some researchers such as Sarko (2008) argued that Arabic students perform worse than other L2 learners such as French who have a more similar article system to that of English. Therefore, it may be plausible to predict that if this study conducted on other L2 learners who have similar article system to that of English, the results of the D post-test would be different and students' performances would be much accurate.
This research also investigated the effects of test types on students' results, because the instruments used to measure the students' performance with articles were two different tests. In order to determine if test type had any influence on students' performances with English articles, gain scores obtained by the three groups were submitted to a Paired Samples T Test ( See Tables 25, 26, and 27). These findings revealed that there was no significant difference between the groups' results in the two types of tests since both of these types are similar to each other. In fact, these results strongly support the fourth hypothesis which is students' performance would not differ between fill-in-the-blank tests and cloze tests. The researcher acknowledges that if she chose different types of tests, the students' performances would be different between the types of tests.
The choice of the topic was greatly determined by the fact that very little attention had been given to CF in the Saudi classroom. Therefore, it was an opportunity for the researcher to examine the role of certain types of CF which are recast and metalinguistic in helping students to develop their using of English articles. The researcher carried out the study on third-year intermediate female school students in Saudi Arabia. The research focused on four main hypotheses:
CG, RG, and MG students' performances on the pre-tests in both types of tests (fill-in-the-blank and cloze) will be comparable and there will be no significant differences among them.
the RG on the post-test in the two types of tests will outperform the CG, whereas the MG students will outperform both the CG and RG.
the MG students will advance to higher levels of supplying target-like English articles than CG and RG students on the D post-test.
students' performance will not differ between fill-in-the-blank tests and cloze tests.
In order to investigate these research hypotheses, the study followed a pre-test, post-test and D post-test design and was carried out over 6 weeks with 50 students 20 participants for the RG and 20 participants for MG and 9 controls, aged between 14 and 15 years. A fill-in-the-blank test and a cloze test were administered.
The results of this study supported the first hypothesis which proposed that the CG, RG, and MG students' performances on the pre-tests in both types of tests (fill-in-the-blank and cloze) would be comparable and there was no significant differences among them. Furthermore, the results of this study supported the second hypothesis which proposed that RG on the post-test in the two types of tests would outperform the CG, whereas the MG students would outperform both the CG and RG. On the other hand, the MG students would advance to higher levels of supplying target-like English articles than CG and RG students on the D post-test. However, in the D post-test, the results were not compatible with the third study hypothesis which suggested that the MG students would advance to higher levels of supplying target-like English articles than CG and RG students on the D post-test. Regarding the fourth hypothesis, students' performance did not differ between fill-in-the-blank tests and cloze tests.
All in all, both recast and metalinguistic feedback facilitate the acquisition of English articles in the short-term. However, the effectiveness of metalinguistic on the students' performance is more accurate than recast in the short-term. Regarding the long-term, there were no advantages of metalinguistic feedback over recast.
In spite of the enormous efforts made to tackle any potential problems during designing this study, as in all classroom studies, there were inevitable limitations. The researcher intended to conduct the D post-test after two months of the post-test in order to see the long-term effects of CF; however, administering two post-tests with a six-week interval was dictated by scheduling constraints as mentioned previously.
Another limitation is that this study looked at one linguistic target and they were tested in their written skill. However, if the study looked at two or more linguistic targets and they were orally, we would understand better if the efficiency of metalinguistic and recast feedback could be linked to the linguistic target. Furthermore, this study was conducted on females only. In fact, it would have been interesting to compare the performance of females with males. However, since the study was conducted in Saudi Arabia, there was no admission for female researchers to male schools.
Several suggestions can be made for future research. In all the studies that I have come across, the researchers have designed their tests as pre-test, post-test, and D post-test in order to see the effectiveness of the CF; however, it could be argued that if they design more than two post-tests, they could get a deep insight into the role of CF properly. Moreover, further research needs to be conducted to survey students and teachers' attitudes towards the efficiency of metalinguistic feedback in comparison with recast feedback. Further research needs to be performed to investigate students in their speaking classes and to record their natural performances before and after being given metalinguistic and recast feedback in order to see their accuracy of oral production. This might be useful to see to what extent the CF is helpful and effective.