place of articulation.

Figure8. Errors and problems by EFL learners

As demonstrated in the table, there are different reasons behind making such errors ranging from the absence in first language to different manner of articulation. The treatment lasted for 30 minutes class hours; one class hour – 90 minutes – a week for four weeks. After the four-week treatment was completed the post test was administered to all students. Figure 2 shows the data collection procedures.

Experimental Group Control Group

3.5. Data Analysis Figure9.

Figure9. Data Collection Procedures

3.6. Methods of Analyzing Data

After the pretest and post-test, the data collected was analyzed quantitatively via Statistical Package for Social Sciences (SPSS), which is a computer program that is used to analyze data in research studies in Social Sciences. In order to answer the research question (R.Q.), first, both the pre and the post-tests were scored and all test results were entered into SPSS to analyze the data, to do the results quantitative data have been analyzed through descriptive statistical methods (mean and standard deviation), and the researcher ran inferential statistics (t-test) and analysis of covariance (ANCOVA).

3.7. Summary

In this chapter the methodology of the research was described in detail. The setting and participants, the instruments and materials used in the training as well as the data collection procedures were explained. The next chapter will present the findings coming from the data analysis of the research.

CHAPTER IV: DATA ANALYSIS

4.0. Introduction

As it was mentioned before the aim of this study was to investigate the impact of contrastive based pronunciation teaching on listening comprehension of Iranian EFL learners. The research question addressed in this study was:

Does contrastive based pronunciation teaching have any effect on Iranian EFL learners’ listening comprehension?

In order to answer this question one experimental and one control group were formed at Iran Language Institute. Two groups were administered a pre-test. After the pre-test, the experimental groups received four-week contrastive based pronunciation training. Meanwhile, the control group continued their regular classes offered by the Iran Language Institute. After four weeks, all groups were administered another test as the post-test. The pre-test was administered to see the level of the participants in both the experimental and control group, and the post-test was administered to examine the improvement the participants have made at the end of four weeks. This section of the article is dedicated to answering the research question and testing the related hypothesis.

4.1. Data Analysis and findings

To fulfill the purpose of this study a pre-test, post-test design was used. The data collected from the participant‘s pretest and posttest were analyzed quantitatively to answer the question addressed in the study. To do so the results of quantitative data have been analyzed through descriptive statistical methods (mean and standard deviation), inferential statistics (t-test) and analysis of covariance (ANCOVA).

Before conducting an ANCOVA, its assumptions were examined.

1. Linear relationship between the dependent variable and the random variable which can be examined through the distribution graph between each pair of variables (Multicollinearity).

2. The homogeneity of variance assumption.

3. The homogeneity of regression slopes assumption.

The relationship between random independent variables and dependent variables is illustrated in graph 4-1.

Figure 10: Listening comprehension ability of control and experimental groups

As seen in the graph there is a linear relationship between random independent variables (pre-test) of listening comprehension ability and dependent variables (post-test) of listening comprehension ability since the regression lines are almost parallel which means the relationship between the variables in both groups is similar. The correlation coefficient between the dependent variables is illustrated in table 4-1.

Table 4-1: The correlation coefficient between the random variables of listening comprehension ability (pre-test) and the independent variables

Variable Listening comprehension ability P

Listening comprehension ability (pre-test) **0/688 0/000

P0/01**

The correlation results show that there is a significant and acceptable relationship between listening comprehension random independent variable (pretest) and listening comprehension’s dependent variable (posttest), therefore ANCOVA can be applied. Statistical index of listening random independent variables (pretest) and listening dependent variables (posttest) and the results of the T-test in both experimental and control group are presented in tables 4-2 and4-3.

Table 4-2: Statistical characteristics of the dependent variable (post-test) of listening comprehension ability in the control and experimental groups

Group Mean Std. Deviation

Exp. 22.0000 3.22817

Cont. 26.7000 2.75490

The figures shown in the table above indicate that there is a difference between the average mean of the control group and the experimental group in the dependent variable (post-test), and the difference is not meaningful ( F(1,38)=0/044 , p=0/835).

Table 4-3: Paired T-test to compare the mean difference between pre-test and post-test experimental and control groups

Group

Mean

Std. Deviation

Std. Error Mean

Mean Difference

T

Sig

Experimental

Pair 1

Posttest

21.2500

3.47737

.77756

-5.45000

-16.602

.000

Pretest

26.7000

2.75490

.61601

Control

Pair 1

Posttest

21.5500

3.08605

.69006

-.45000

-1.917

.070

The figures in table 4-3 show there is a significant difference between the averages of experimental group in the dependent variable (post-test). In other words, since there has been no instruction in the control group, no change has been evident; but experimental group there is a difference between pre-test and post-test.

Table 4-4: Mean and the rounded mean of listening comprehension ability dependent variable

Post-test Rounded mean

Source M SD M SE

Control 22.00 3.228 21.874a .264

Experimental 26.70 2.755 26.826a .264

The rounded means of listening comprehension ability dependent variable are presented in the above table which means the effect of random variables is statistically removed. The means show that the mean of experimental group is higher than the mean of the control group.

Table 4-5: The correlation coefficient of listening comprehension ability with removing mutual effect in the experimental and control groups.

Variable source df Sum of squares Mean Squares F p Eta

Listening Contrast 1 244.724 244.724

Comprehension 175.653 .000 .826<
br /> ability Error 38 51.549 1.393

As it can be seen the effectiveness of F ratio is statistically significant (Eta = 0.826, p= .000, F (1.38) 175.653 ) which shows there is a significant difference between the listening comprehension abilities of the experimental and control group. Therefore, it can be concluded that contrastive based pronunciation teaching has a positive impact on listening comprehension ability of Iranian EFL learners. In order to clarify the above points, the information is presented in graph 4-2. In this graph experimental and control variables are presented in the horizontal axis and the writing dependent variable posttest is presented in the vertical axis.

Figure 11: The post-test means of listening comprehension ability in the experimental and control groups.

The graph above shows that there is a significant difference between the listening comprehension ability of experimental and control groups. The graph clearly indicates that the participants who have been treated with contrastive based pronunciation teaching have a higher listening comprehension ability compared to the control group.

4.2. Summary

This chapter has presented information regarding the data analysis and the results. According to the statistical tests conducted by the researcher, there is a statistically significant increase in the experimental group after a four-week of contrastive based pronunciation teaching. The results of the ANCOVA test which was used to analyze the data driven from the participants clearly showed the nullification of the hypothesis and the experimental group outperformed the control group in terms of listening comprehension ability. Therefore, it can be concluded that the participants who were given the contrastive pronunciation teaching showed a greater improvement in their listening comprehension. According to the findings, the development that the experimental group has performed is significantly higher than the control group. Thus, the results suggest that contrastive based pronunciation training has an effect on the listening comprehension skills Iranian EFL students. The next chapter will discuss the results, limitations,