Cumulative Percent
Valid Strongly Agree 10 10.0 10.0 10.0
Agree 11 11.0 11.0 21.0
Undecided 16 16.0 16.0 37.0
Disagree 32 32.0 32.0 69.0
Strongly Disagree 31 31.0 31.0 100.0
Total 100 100.0 100.0

This table shows the responses’ frequencies for the twentyfifth question. According to this table, the adding frequency number of Agree and Strongly Agree is 21 and the adding frequency number of Disagree and Strongly Disagree is 63. Therefore the perceptions about this question are mostly on negative side and the distance is significantly different.
Table 4.36. Q26 Frequency
Frequency Percent Valid Percent Cumulative Percent
Valid Strongly Agree 9 9.0 9.0 9.0
Agree 13 13.0 13.0 22.0
Undecided 22 22.0 22.0 44.0
Disagree 28 28.0 28.0 72.0
Strongly Disagree 28 28.0 28.0 100.0
Total 100 100.0 100.0

This table shows the responses’ frequencies for the twentysixth question. According to this table, the adding frequency number of Agree and Strongly Agree is 22 and the adding frequency number of Disagree and Strongly Disagree is 56. Therefore the perceptions about this question are mostly on negative side and the distance is significantly different.
Table 4.37. Q27 Frequency
Frequency Percent Valid Percent Cumulative Percent
Valid Strongly Agree 9 9.0 9.0 9.0
Agree 12 12.0 12.0 21.0
Undecided 23 23.0 23.0 44.0
Disagree 29 29.0 29.0 73.0
Strongly Disagree 27 27.0 27.0 100.0
Total 100 100.0 100.0

This table shows the responses’ frequencies for the twentyseventh question. According to this table, the adding frequency number of Agree and Strongly Agree is 21 and the adding frequency number of Disagree and Strongly Disagree is 56. Therefore the perceptions about this question are mostly on negative side and the distance is significantly different.
Table 4.38. Q28 Frequency
Frequency Percent Valid Percent Cumulative Percent
Valid Strongly Agree 9 9.0 9.0 9.0
Agree 13 13.0 13.0 22.0
Undecided 21 21.0 21.0 43.0
Disagree 29 29.0 29.0 72.0
Strongly Disagree 28 28.0 28.0 100.0
Total 100 100.0 100.0

This table shows the responses’ frequencies for the twentyeight question. According to this table, the adding frequency number of Agree and Strongly Agree is 22 and the adding frequency number of Disagree and Strongly Disagree is 57. Therefore the perceptions about this question are mostly on negative side and the distance is significantly different.
Table 4.39. Q29 Frequency
Frequency Percent Valid Percent Cumulative Percent
Valid Strongly Agree 9 9.0 9.0 9.0
Agree 12 12.0 12.0 21.0
Undecided 23 23.0 23.0 44.0
Disagree 29 29.0 29.0 73.0
Strongly Disagree 27 27.0 27.0 100.0
Total 100 100.0 100.0

This table shows the responses’ frequencies for the twentyninth question. According to this table, the adding frequency number of Agree and Strongly Agree is 21 and the adding frequency number of Disagree and Strongly Disagree is 56. Therefore the perceptions about this question are mostly on negative side and the distance is significantly different.
Table 4.40. Q30 Frequency
Frequency Percent Valid Percent Cumulative Percent
Valid Strongly Agree 9 9.0 9.0 9.0
Agree 14 14.0 14.0 23.0
Undecided 19 19.0 19.0 42.0
Disagree 30 30.0 30.0 72.0
Strongly Disagree 28 28.0 28.0 100.0
Total 100 100.0 100.0

This table shows the responses’ frequencies for the thirtieth question. According to this table, the adding frequency number of Agree and Strongly Agree is 26 and the adding frequency number of Disagree and Strongly Disagree is 51. Therefore the perceptions about this question are mostly on negative side and the distance is significantly different.
4.11. Analysis of the data related to the Research First and Second Hypotheses
According to the first research question which is What are the teachers’ perceptions about the effects of Recast on Elementary learners’ pronunciation? , the first hypothesis is formulated: Teachers’ perceptions about the effects of Recast on Elementary learners’ pronunciation are not positively significant. Through investigating the descriptive statistics of the second ten questions and their frequency data, the null hypothesis is rejected. Therefore Teachers’ perceptions about the effects of Recast on Elementary learners’ pronunciation are positively significant and according to the statistics analysis they have positive perceptions about it and they consider it very important.
According to the second research question which is What are the teachers’ perceptions about the effects of Metalinguistic Feedback on Elementary learners’ pronunciation? , the second hypothesis is formulated: Teachers’ perceptions about the effects of Metalinguistic Feedback on Elementary learners’ pronunciation are not positively significant. Through investigating the descriptive statistics of the third ten questions and their frequency data, the null hypothesis is rejected. Therefore Teachers’ perceptions about the effects of Metalinguistic Feedback on Elementary learners’ pronunciation are positively significant and according to the statistics analysis they have negative perceptions about it and they don’t consider it very important.
4.12. Friedman test
The Tests for Several Related Samples procedure compares the distributions of two or more variables. The Friedman test is a nonparametric alternative to the one-way ANOVA with repeated measures. It means that if the data are normal, ANOVA is used for scores mean comparison but otherwise if any group of data is not normally distributed, the nonparametric alternative to ANOVA which is Friedman test is used. It is used to test for differences between groups when the dependent variable being measured is ordinal. The Tests for Several Related Samples procedure compares the distributions of two or more variables. Friedman tests whether k related samples have been drawn from the same population. The output shows the mean rank for each variable in the Friedman Ranks table and the number of valid cases, chi-square, degrees of freedom, and probability in the Test Statistics table.

Friedman Test

Table 4.41.Friedman Test Ranks
Mean Rank
first 10 questions 1.97
second 10 questions 1.37
third 10 questions 2.67

Test Statisticsa
N 100
Chi-square 86.583
Df 2
Asymp. Sig. .000
a. Friedman Test

As indicated in this table the mean rank of the second ten questions is orderly less than the first and the third ten questions, which really means the answers of the second ten questions led to options 1 and 2 and the answers of third ten questions led to options 4 and 5. In Friedman test, the null hypothesis states that: there is no significant difference among the three groups’ scores means. Since the Sig Value is .000 and it’s less than .05, the null hypothesis is rejected and there are significant differences among the three groups’ scores means.
4.13. Wilcoxon Signed Ranks Test

Because the Recast data was not normal, the researcher could not use the ordinary 2-sample t-test. Instead, the researcher used a nonparametric test. The Wilcoxon signed-rank is the nonparametric test equivalent to t-test. Regarding the obtained results, with due attention to rank and compare the data, the researcher applied Wilcoxon Signed Ranks Test. It is used to compare two sets of scores that come from the same participants. The two things the researchers are interested in the output are the Z value and the associated significance levels, presented as Asymp. Sig.(2-tailed).

Table 4.42.Wilcoxon Signed Ranks Test
N Mean Rank Sum of Ranks
third 10 questions – second 10 questions Negative Ranks 8a 22.94 183.50
Positive Ranks 89b 51.34 4569.50
Ties 3c
Total 100
a. third 10 questions < second 10 questions
b. third 10 questions > second 10 questions
c. third 10 questions = second 10 questions

Test Statisticsb
third 10 question – second 10 question
Z -7.894a
Asymp. Sig. (2-tailed) .000
a. Based on negative ranks.
b. Wilcoxon Signed Ranks Test

Regarding the above table , while Mean2 (mean of the questions 21-30) is more than Mean1 (mean of the questions 11-20), Numbers, Mean Rank and Sum of Ranks reach the maximum.
In this ground, the positive ranks which indicate Mean2 is more than Mean1 and equal to 89 are more than negative ranks which indicate Mean2 is less than Mean1 and equal to 8. If the significance level is equal or less than .05 then the researcher can conclude that the difference between the two scores is statistically significant. In this example the Sig. value is .000 (which really means less than .05). Therefore the researcher can conclude that the two sets of scores are significantly different.
4.14. Analysis of the data related to the third Research Hypothesis
Regarding the obtained results, with due attention to rank and compare the data ,the researcher applied Wilcoxon Signed Ranks Test. In this rank test the mean of teachers’ perceptions about the effects of Recast were compared with the mean of teachers’ perceptions about the effects of Metalinguistic Feedback. The results indicated that mean2 (mean of teachers’ perceptions about the effects of Recast) was more than mean1 (mean of teachers’ perceptions about the effects of Metalinguistic Feedback).
According to the third research question which was: Is there any significant difference between teachers’ perceptions about the effects of Recast and Metalinguistic Feedback on Elementary learners’ pronunciation?, the third hypothesis was formulated: There is no significant difference between teachers’ perceptions about the effects of Recast and Metalinguistic Feedback on Elementary learners’ pronunciation.
In this test the Sig. value is .000 (which really means less than .0005). Therefore the hypothesis was rejected and the researcher concluded that the two sets of scores are significantly different and there is significant difference between teachers’ perceptions about the effects of Recast and Metalinguistic Feedback on Elementary learners’ pronunciation.

Chapter 5

Conclusion, Implication and

Recommendation

5.1. Introduction
This chapter considers the results of the study. Each one of the research questions is elaborated and according to calculated numbers the result’s explanation is provided. To support the researcher’s own claims, reference to the literature is presented. Other elements which are provided in this chapter are theoretical and pedagogical implication.
5.2. Discussion and Conclusion
The current study investigated the teachers’ perceptions about the effects of two types of corrective feedback on the pronunciation error correction The descriptive statistics of Recast and Metalinguistic Feedback effects according to the teachers’ perceptions were provided through multiple analysis of SPSS such as investigating the means, medians, modes, standard deviation, variance, Friedman test and Wilcoxon signed-rank to see what are the teachers’ perceptions about the effects of Recast on Elementary learners’ pronunciation and to see whether