English language (such as grammar and mechanics).

Based on the achieved results, the mean difference between Group 1(experimental group) and Group 2 (control group) was not significant in the pre-test since the significance value was above 0.05 (p0.05). As a result, the two groups were homogeneous in their abilities in writing accurate sentences.

After making sure about subjects’ homogeneity, 14 students were assigned to the experimental group and the rest (n=14) were recruited for the control group. Then, a writing test was conducted for the experimental group and the students with lower marks were named writer and the ones with the higher marks were called helper. Next, one helper and one writer made a pair to supplement each other and followed the steps of treatment. The control group had its routine way of writing by the teacher, giving the topics; and the students writing individually and getting feedback from the instructor.

At the end of the term, a post-test writing was conducted. Based on the results, the mean difference between Group 1 and Group 2 was significant. It means that Group 1 outperformed significantly compared with the participants in Group 2 since the significance value (i.e., 0.001) was less than 0.05. Therefore, comparing the results of pretest and posttest for both groups shows that experimental group outperforms the control group in writing accuracy.

The reason for such a result can be utilizing pair work in experimental group besides teacher feedback. Thus, the alternative hypothesis of this study is confirmed, meaning that using pair work significantly affects writing performance of the EFL students.

One good way for visually examining the distribution of scores from two groups is to use a boxplot. Boxplots provide more information than bar plots and are able to be a quick visual check on the assumptions of normality and homogeneity of variances that are necessary for using a parametric t-test. In the independent sample T-tests, scores from two different groups on the same measure were compared.

Table 4.1 shows raw data of control and experimental groups in the pretest. After the produced error-free T-units were divided into the total number of T-units, the accuracy of writings for the students in both groups was achieved.

Table 4.1. Raw scores of control and experimental groups for pretest

Pre- control

Pre-experimental

Error-free T-units

The total number of T-units

Accuracy

Error-free T-units

The total number of

T-units

Accuracy

4

14

0.285714

2

15

0.133333333

5

17

0.294118

4

18

0.222222222

5

21

0.238095

5

20

0.25

6

21

0.285714

5

20

0.25

4

19

0.210526

5

18

0.277777778

5

20

0.25

5

19

0.263157895

5

18

0.277778

4

17

0.235294118

6

19

0.315789

6

17

0.352941176

7

21

0.333333

6

19

0.315789474

6

23

0.26087

7

21

0.333333333

6

17

0.352941

7

16

0.4375

6

16

0.375

5

16

0.3125

6

21

0.285714

8

15

0.533333333

5

19

0.263158

8

20

0.4

Table 4.2. Independent samples test to compare the accuracy of experimental and control groups in the pre-test (Italic=statistically significant difference)

Group

No

Mean

Std

Sig.

1

14

0.2878

0.04477

0.49

2

14

0.3084

0.10060

Notes: 1= Experimental Group , 2= Control Group

As shown in Table 4.2, the mean score of experimental group is achieved to be 0.2878; while, the mean score of the control group is 0.3084. Regarding the standard deviation value which is 0.04477 for the experimental group and the standard deviation of 0.10060 for control group, it is observed that the mean difference between Group 1 and Group 2 is not significant in the pre-test since the significance value is above the cut-off value (i.e. p 0.05).

Therefore, the two groups are homogeneous in their abilities and proficiency level in writing accurate sentences.

Based on the means of the participants in the experimental and control groups, Figure 4.1 shows the performances of EFL students in the pre-test, regarding the accuracy of their writings.

Notes: 1=Experimental Group, 2= Control Group

Figure 4.1 Mean differences of performance of the experimental and control groups in the pre-test regarding the accuracy

Table 4.3 shows raw scores of control and experimental groups for posttest

Table 4.3. Raw scores of control and experimental groups for posttest

Post control

Post experimental

Error-free T-units

The total number of T-units

Accuracy

Error-free T-units

The total number of T-units

Accuracy

5

15

0.333333333

6

18

0.333333333

5

18

0.277777778

8

19

0.421052632

4

20

0.2

7

19

0.368421053

5

22

0.227272727

5

20

0.25

4

20

0.2

6

18

0.333333333

4

20

0.2

6

20

0.3

4

19

0.210526316

8

18

0.444444444

6

19

0.315789474

7

20

0.35

4

20

0.2

7

18

0.388888889

5

22

0.227272727

8

21

0.380952381

7

18

0.388888889

7

15

0.466666667

7

17

0.411764706

6

17

0.352941176

6

20

0.3

8

18

0.444444444

6

17

0.352941176

6

19

0.315789474

Comparing the number of error-free T-units in the control and experimental groups after probing Table 4.1 and Table 4.3, showed that the number of produced error-free T-units in the experimental group is higher than the control group. As a results, the students in the experimental group produced more accurate sentences than the students in the control group.

Table 4.4 shows independent samples test to compare the accuracy of experimental and control groups in the post-test.

Table 4.4. Independent samples test to compare the accuracy of experimental and control groups in the post-test (Italic=statistically significant difference)

Group

N

Mean

Std

Sig.

1

14

0.3679

0.0614

0.001

2

14

0.2747

0.0757

Notes: 1= Experimental Group 2= Control Group

As shown in Table 4.4, the mean score of experimental group is achieved to be 0.3679; while, the mean score of control group is 0.2747. Regarding the standard deviation value which is 0.0614 for the experimental group and the standard deviation of 0.0757 for control group, it is observed that the mean difference between Group 1 and Group 2 is significant in the post-test with Group 1, implying that Group 1 has been performing significantly better than the participants in Group 2 since the significance value (i.e., 0.001) is less than the cut-off value (i.e. p 0.05).

Based on the means of the participants in the two groups, Figure 4.2 shows their performances in the post-test regarding the accuracy.

Figure 4.2 Mean differences of performance of the experimental and control groups in the post-test regarding the accuracy

Notes: 1=Experimental Group 2= Control Group

4.4. Summary

This section started with a brief discussion about the background information of the students of both control and experimental groups as the participants of the study. Then, the background information of the teacher of both groups (who was one person) and the author of this thesis was represented. Next, the only research question of this paper was stated and the methods for testing it were reviewed. At the end, it was concluded that based on the achieved results, the experimental group outperformed the control group in writing accuracy. Thus, the alternative hypothesis was confirmed.

CHAPTER V

DISCUSSION AND CONCLUSION

5.1 Introduction

In the previous chapter, the collected data were statistically analyzed and the findings of the study with respect to the research question were displayed in different tables and figures. In this chapter, first, the summary of the main findings of the study is presented. Then the only research question is responded, discussed, and compared with the results of similar studies conducted in this field. After that, the pedagogical implications and limitations of the study will be mentioned and in the final step, some suggestions will be given for further researches.

5.2 Summary of the Study

As mentioned before, the main focus of this study was on the investigation of the effects of practicing pair work in EFL classes on improving writing performance (in this case grammatical accuracy aspect) of EFL learners at intermediate level in the language institute of Pardis in Marand Town. In order to gather evidences to do this, two classes (one experimental group and one control group) were formed whose students were all at the intermediate level of proficiency and their homogeneity was assured by the records of Language Institute of Pardis. Then, a pretest was conducted for both groups with a prompt of narrative genre for writing. Next, the writings were collected and scored manually by the teacher using Storch and Wigglesworth (2009)’s measure as a criterion to measure the accuracy of the writings. To meet rater reliability, the scoring procedure was monitored by the supervisor of the institute. Then, the experimental group was paired based on writing scores with the student with lower mark named writer and the student with the higher mark called helper. The experimental group got a pamphlet including the steps of pair work and guidance by the teacher about narrative genre that was the focus of the class. The control group also received the same guidance by the teacher about narrative genre. Then, both groups had eight sessions of writing practices and at the tenth session, a post-test was conducted and the papers were scored like pretest. Then, using the measure of independent t-test and some other descriptive measures (mean, significance, and standard deviation), scores in experimental and control groups in pretest and post-test were compared. Based on the achieved results, the mean difference between experimental group and control group was not significant in the pre-test since the significance value was above the cut-off value (i.e. p 0.05). Therefore, the two groups were homogeneous in their abilities in writing accurate sentences.

Examining post-test results, the mean score of experimental group was achieved to be 0.3679; while, the mean score of control group was 0.2747. Regarding the standard deviation value which was 0.0614 for the experimental group and the standard deviation of 0.0757 for control group, it was observed that the mean difference between experimental and control groups was significant in the post-test with experimental group, implying that this group has been outperforming the participants in control group since the significance value (i.e., 0.001) was less than the cut-off value (i.e. p 0.05).

5.3 Discussion

The main goal of this paper was to probe the influence of peer work on textual accuracy. The reason for selecting pair work rather than other collaborative techniques was the ample body of literature, and theoretical and pedagogical support for practicing pair work. As an illustration, Peregoy and Boyle (2001) pointed out that pairing students up in writing is an ideal way to promote learning effectiveness. Storch and Wiggleworth (2012) stated that a common teaching strategy in the language classroom is to assign students to work on a task in pairs or small groups. Furthermore, Brown (1994) supports the use of small group writing and pair writing stating that students often carry out peer correction in these activities and this is more productive than teachers always correcting students.

Based on the findings summarized in the previous section, the research question of the study was answered and discussed as follows:

The research question was” Does instructing pair work strategies affect writing performance of Iranian EFL learners?”. Examining the results of pretest and posttest for both experimental and control groups and their detailed