You are on page 1of 11

Apr. 2010, Volume 7, No.4 (Serial No.


US-China Education Review, ISSN 1548-6613, USA

Cooperative learning as a correction and grammar revision technique: Communicative exchanges, self-correction rates and scores ∗
Sara Servetti
(Department of Applied Linguistics, University of Torino, Santo Stefano Roero 12040, Italy)

Abstract: This paper focuses on cooperative learning (CL) used as a correction and grammar revision technique and considers the data collected in six Italian parallel classes, three of which (sample classes) corrected mistakes and revised grammar through cooperative learning, while the other three (control classes) in a traditional way. All the classes involved in this study were in their first year of secondary school, and although most students had a level of English which was A2, they made a certain number of mistakes also in grammar topics like the present tense, which is the grammar topic under examination in this study. This paper analyses the sample students’ communicative exchanges while they performed the error correction activity through CL, compares the self-correction rates reported by the two groups of students after the two types of correction and revision activities (traditional and through CL) and the students’ scores in tests given to students one, four and eight weeks after the correction lessons. The aim of the study is establishing if the use of CL as a correction and grammar revision technique had a beneficial effect on the students who experienced it and in particular on the low achievers. Key words: cooperative learning; error correction; grammar revision

1. Theoretical background
1.1 Involving and helping students through CL Cooperative learning has been experimented and used in a variety of contexts and activities, both in order to enhance students’ knowledge and skills, and to foster their interpersonal relationships, two goals which are however intertwined within CL groups. A great number of research studies (Johnson & Johnson, 1981; Slavin, 1987) show the effectiveness of cooperative learning in many contexts and from many points of view. First of all, this way of working together influences positively both the learners’ achievement (Ream, 1990), they obtain higher achievement scores in comparison with individualistic groups (Sherman & Thomas, 1986) and they can learn material better than individual students (Yager, et al., 1985), and their interpersonal relationships, they build both interpersonal and higher-level cognitive skills (Michaelsen, 1992); they support their peers and engage in constructive conflict resolution (Johnson & Johnson, 1994); they develop a sense of social responsibility (Vermette, 1988) and of mutual respect (Pate, 1988). The positive effects of CL activities are particularly beneficial for the low achievers, as they can receive attention from the other group mates and help from more experienced peers (Johnson, et al., 1991). On the grounds of these benefits and the involvement in the task which characterizes CL activities,
This paper was presented at the International Conference “Intercultural Education: Paideia, Polity, Demoi” (Athens, Greece, June 22nd to 26th 2009) co-organized by the International Association for Intercultural Education (IAIE) and the Hellenic Migration Policy Institute (IMEPO), under the aegis of Unesco. Sara Servetti, Ph.D., Department of Applied Linguistics, University of Torino; research field: teaching foreign languages. 12

were at different proficiency levels (mainly A2) but made a certain number of mistakes also in simple structures. two moments of the language lesson which are not very involving and motivating for learners. they usually pay more attention to the mark than to their mistakes. textbook and number of English lessons per week. The total number of students involved in this study was 150. The author is grateful to both schools. Students in fact are usually rather passive during both activities: when they receive their tests back. were chosen for the study.1 Sample students Six parallel classes. four were in Istituto Superiore “Leonardo Da Vinci” in Alba (CN). and this attention to form can help language acquisition. so all the variables which could have had a biasing influence on the data collected were very similar in each couple of classes. which is welcomed by learners. It is a common practice in language classrooms. They were from the first year of the Italian secondary school: two parallel classes were in Istituto Superiore “G. followed English lessons for 3 or 4 hours a week.2 Written error correction and grammar revision in the language classroom In the language lesson. In this study. 13 1 . only the tests of students who were present during all the tests and during the error correction activities were taken into account: Two experiments were carried out in this study and data were collected from 128 students in the first experiment. being guided towards improvement by a mark or a comment. project. and from 135 in the second. Van Patten & Cadierno. to the English teachers (Gabriella Martinelli. All the students had studied English for at least three years. Govone” Liceo Classico in Alba (CN). The study 2. some researchers (Truscott. 1. while others (Ferris & Roberts.Cooperative learning as a correction and grammar revision technique: Communicative exchanges. who want and expect to be corrected (Leki. 1D and 1F in Liceo delle Scienze Sociali and 4C and 4D in Liceo Classico). 1996) state that correction is useless. Three different schools were involved in this study because in each school two parallel classes were chosen (1A and 1B in Liceo Linguistico. activities like written error correction and grammar revision aim at making students focus on the form of linguistic structures. two in the Liceo Linguistico and two in the Liceo delle Scienze Sociali 1. CL was used as an experimental technique to correct the students’ mistakes and to revise grammar rules. 1993). In order to compare the results of students who received different correction and grammar revision lessons. like the present tense. in order both to understand how the language works and to produce accurate language. in each school one parallel class (sample) was corrected through CL and the other (control) through traditional methods. However. in order to establish if this use of CL could help the students who experienced it and in particular the low proficiency ones. students rarely ask questions for clarification.D. and when grammar rules are revised in the classroom. Although no agreement has been found by research on the effectiveness of error correction. Through written error correction learners are made aware of their learning steps and their difficulties through the teacher’s feedback. 2001) show that it helps learners improve. Through grammar revision learners focus on rules. as shown by research (DeKeyser. 1995. made up of 14/15-year-old students of English as a Foreign Language. which had the same teacher. in order to involve students in these steps of the language lesson in an active way. which was part of a larger Ph. 2. self-correction rates and scores cooperative learning has been chosen for this study as an experimental technique for error correction and grammar revision. 1991). Consolata Sobrero and Lucia Toppino) and to the students who took part in this study for their kindness and cooperation. which is the object of this study.

During this activity all the CL groups were tape recorded. 1981). they received their pre-test back and they had to self-correct it individually. and their communicative exchanges analysed. individually). and in sample classes within mixed-ability CL groups (4-5 people). Analysis 3. in which students had also to motivate their corrections. self-correction rates and test scores over time. 3. in order to establish to what extent students were able to correct together the wrong sentences properly and motivated corrections referring to grammar. 1996). each correction task lasted 15-20 minutes. the first on simple present and the second on the difference simple present vs. four and eight weeks after the correction lessons. progressive. All the tests were checked and students’ errors underlined but not corrected. therefore grammar rules were revised. Although using different techniques. one hour lesson on the grammar topic of each experiment. and in control classes the list of wrong sentences that students had to correct individually was collected after the task. Students had time for questions in the end of the lesson. In all the classes the number of the successfully corrected sentences was divided by the total number of sentences in the list. the time given to such activities is not very long and then because it is focusing too much on errors can demotivate students (Ur. so the percentage of successful corrections was determined. Post tests on the same grammar topic were given to all the students one. One hour was dedicated to this activity because in language lessons. After all the classes revised a grammar topic with their teacher. This study had one main goal. 3.1 Successful corrections within CL groups versus individual correction As one of the aspects under examination in this study was to establish if students within CL groups were able to correct more mistakes than students working on their own. and the same methodology was used. that is establishing if the CL activities performed had any positive effect for the students who experienced it (and in particular for the low proficiency students) in any of the three aspects taken into account—amount of wrong sentences corrected (within CL groups vs. both the groups spent the same time focusing on error correction and grammar revision. Both experiments (the first on simple present and the second on simple present versus progressive) were carried out following the same methodology and the whole study lasted four months. The tape recordings lasted around 10 hours. After that. made up of both formal exercises and a text to write. In the control groups they had to correct it individually. the percentages of successful corrections made by sample and control students were compared: In every CL group tape scripts were analyzed. the type of CL used was the groups of four (Burns. as in each sample class 5-6 groups were formed. percentages were made for every group (in the sample classes) 14 . which lasted one hour. The following lesson all the students corrected their pre-test and they received an anonymous list of the most common mistakes (chosen by frequency and typology) found in their pre-test.Cooperative learning as a correction and grammar revision technique: Communicative exchanges.2 Methodology This study focuses on two experiments. and two error correction and grammar revision activities were performed. all the students corrected the wrong sentences in plenary giving reasons for each correction.1 Communicative exchanges within groups The students in sample classes were tape recorded while performing the CL activities. that is. they were given a grammar test on the grammar topic taught (pre-test). in order to test long term results and to establish if the CL activity had any influence on the students’ accuracy. self-correction rates and scores 2.1.

as the numbers in bold highlight. the percentages of successful corrections were much higher in sample than in control groups: CL groups were able to find the appropriate solution for nearly every sentence. while in control groups individual students showed more varied percentages (from 23% to 100%). so students within CL groups are able to correct sentences more successfully than individual students.09% (sd=20.36) 78. motivated their corrections and explained each other grammar rules.9% (sd=5. Therefore.6% (sd=5. and finally mean percentages were calculated.05) in every case.77% (sd=5.04 As reported in Table 1.77) 94. self-correction rates and scores and every student (in control classes) for both experiments. 3. Classico Group Sample Control Sample Control Sample Control Successful corrections in experiment 1 and 2 Experiment 1 96.05). The percentages underwent the analysis of variance (ANOVA) and. ask for explanation. as students talked about grammar and revised rules together within CL groups.2 Self-correction rates The pre-tests corrected by students were collected and analyzed. 15 .76% (sd=10. Also low proficiency students’ self-correction rates were compared and.75) 79. not only to decide among different possibilities but also to justify a correction on which everyone agreed. The statistical analysis of variance applied to these data shows that the difference between the two groups is statistically significant (p<0.68% (sd=14.94% (sd=17.Cooperative learning as a correction and grammar revision technique: Communicative exchanges.55) 80.25) 85. in four cases out of twelve this difference is statistically significant (p<0. and in three cases out of twelve. so in these cases the higher rates in favour of the sample classes show an actual better ability in self-correcting.07) p=0 p=0 p=0.1. the difference was statistically significant (p<0. and sometimes they explained each other how some rules worked. then the mean percentage for every group was calculated. as shown in Tables 4 and 5. The number of successfully corrected mistakes on the present tense was divided by the total number of mistakes on the present tense each student made in the pre-test.2) 94. Scienze Sociali L. so that all the members could understand the reasons for correction and.05). Data underwent the non-parametrical analysis of variance.2 Grammar explanation within CL groups CL groups were invited to correct the wrong sentences and motivate their corrections.07% (sd=5.93) 93. percentages were made for every learner in each class. with percentages of successful corrections ranging from 87% to 100%.88% (sd=18. The following Tables 2 and 3 report two different self-correction rates (one for formal exercises and the other for texts): If percentages in parallel classes are compared. Linguistico L.03% (sd=4. and the self-correction rates were calculated taking into account only the grammar mistakes on simple present (in experiment 1) and on simple present and progressive (in experiment 2). the objective of making students work together and help each other was reached.43% (sd=5.41) p=0 p=0 p=0 Experiment 2 95. Table 1 Item L.53% (sd=10.41) 89. in case. it can be remarked that every sample class self-corrected more than its respective control class. Tape scripts show that students referred to grammar in most cases (75% of the sentences on average). in most cases (eleven out of twelve cases) the students in sample classes self-corrected more than the ones in control classes. because both correction activities focused only on the present tense and no other grammar structure. as standard deviation shows.32) 93. 3.96) 76.

04 Texts 79.01 p=0. 3.12 p=0.33% (sd=6.03) 93.13 p=0 p=0.72 p=0.54% (sd=11. Linguistico L.1 p=0.39% (sd=8. Classico Table 5 Sample Control Sample Control Sample Control Sample Control Sample Control Sample Control Self-correction rates in experiment 1 Formal exercises 89.37) 80%(sd=40) 78.33%(sd=6.14 Low proficiency students’ self-correction rates in experiment 2 Formal exercises 85. self-correction rates and scores Table 2 Simple present L.74 Texts 85.76) 77.25%(sd=18.11) 65%(sd=30.96% (sd=7. the students’ scores will be examined separately for both types of exercises.37 Sample Control Sample Control Sample Control Present simple vs.6) 91. Classico 3.72 Low proficiency students’ self-correction rates in experiment 1 Formal exercises 76.3.58 p=0.19 p=0. Scienze Sociali L. Linguistico L.83) 82.98) p=0.83%(sd=5.21%(sd=14.74% (sd=11. Classico Table 4 Simple present L.79) p=0.67%(sd=17.26) 96.31) p=0.14) 83.67%(sd=27. Scienze Sociali L.58 p=0.59) 95.5%(sd=16.76) p=0.3 p=0.76) p=0.42) 83.78) 80% (sd=38.05% (sd=28.6%(sd=19.13) p=0. progressive L.04) 97.27 Texts 91.44% (sd=23) 97. Linguistico L.17%(sd=4.11) 81.77% (sd=11.6) 66.5%(sd=38.96%(sd=19.75% (sd=13. Test scores will be examined separately for the three schools.17) 61.74) 73%(sd=12. so the difference between parallel classes will be highlighted.44) 78.5%(sd=22.75%(sd=11.84) 95.59) 61%(sd=37.71) 75.19) 88.42) 86.42) 81.73% (sd=7.22 p=0 p=0 Self-correction rates in experiment 2 Formal exercises 92. Linguistico L.05 p=0.01 p=0. Scienze Sociali L.55) p=0. while in texts they had to express ideas through correct grammar forms.5) 67.39) 85.34) 87.92%(sd=29.4%(sd=12.96% (sd=27.5%(sd=16.Cooperative learning as a correction and grammar revision technique: Communicative exchanges. and consisted of exercises for 16 .6) 61.14%(sd=13.25% (sd=8.76) 72.31% (sd=17.73) 100%(sd=0) 94. progressive L.82%(sd=30.7% (sd=11.52) 80%(sd=7. as the abilities involved in performing them are different: In formal exercises students had to focus mainly on producing correct grammar structures.42) 39% (sd=23.04) 67.7%(sd=32) 75.2%(sd=15. Classico Sample Control Sample Control Sample Control Table 3 Present simple vs.5%(sd=5.75%(sd=14.13) 24% (sd=28. Scienze Sociali L.5) 18.63 p=0.4 Texts 82.19) p=0.29) 64%(sd=10.3 Test scores As all the four tests in both experiments included formal exercises and a text to write.09 p=0.64% (sd=5.58) 96.1 Formal exercises Formal exercises were very similar—the same number and type of items.67%(sd=11.

38 +8.14 p=0.98 +15.7 sd=14. the mean scores in test 1 (before the error correction and grammar revision activities) are quite similar in all the couples of classes.25 p=0.35 sd=10.91 sd=16.36 80.05 sd=17.87 p=0.35 sd=9. some statistically significant better results in favour of sample classes can sometimes be found only in the second or in the third test.26 +5.48 sd=7.71 79.07 p=0.33 sd=14.29 sd=7. Classico Control (20) Test scores in formal exercises in experiment 2 Test 2 74.05 74. self-correction rates and scores which only one correct answer was possible.92 p=0.84 p=0.87 +7.26 sd=11.09 sd=8.87 p=0.47 p=0. Secondly.15 sd=7. In all the other cases.14 sd=7.59 sd=23.85 sd=9.91 sd=9.94 sd=6.2 sd=12.71 sd=9.99 p=0.52 sd=12.48 sd=13.8 72.95 sd=10.14 p=0.23 +0.26 p=0.06 p=0.5 sd=15.21 p=0.25 sd=12.82 +0.91 92.15 sd=13.33 85.49 +5.16 87.92 80.3 sd=8.95 p=0.11 p=0.05 sd=8.09 70.76 p=0.1 Improvement Test 2-3 +1.21 87.1 p=0.92 3 80.57 +7.9 79.94 p=0.78 +10.61 sd=18. Thirdly.17 sd=8.83 sd=16.38 sd=18.52 69.68 73.38 sd=10.09 sd=11.3 sd=8.86 sd=13.23 94.4 80.13 p=0.64 The analysis of the tables shows some common trends.01 +11.72 L.12 +2.57 p=0.45 -0.25 71.83 88.5 Test 1-2 +1.07 Test 1-4 +8.39 p=0.06 sd=12.52 79.99 -0. either in scores or in improvement.06 -2.65 93.61 L. in some cases.03 p=0.99 +5. some sample classes had better results than their respective control classes.45 sd=6.48 sd=12.9 sd=12.85 p=0.78 p=0.71 sd=10.89 +5.35 sd=13.05 3 4 76.75 sd=15.12 +3.92 Test 3-4 +8.17 p=0.32 +5.99 L.8 77.17 p=0. however.58 87. Scienze 63.28 p=0.96 sd=10.22 sd=13.5 88.03 -0.8 p=0.Cooperative learning as a correction and grammar revision technique: Communicative exchanges.57 +12.13 sd=9.38 84.91 Test scores in formal exercises in experiment 1 Test 2 83. Scienze Sociali Control (23) Sample (17) L.71 +5.49 p=0. Linguistico 79. Classico 80.06 73.41 p=0.05 sd=6.91 p=0.07 p=0.88 89.09 sd=12.8 p=0. A quick glance at the tables below shows that although most test scores in formal exercises are rather similar in all the couples of classes.88 84.95 p=0. Table 6 1 82 sd=16.55 sd=14.76 p=0.1 sd=8.43 89.27 sd=10.6 sd=7.17 +6.43 Control (23) sd=25.55 Sample (22) sd=23.69 p=0.41 sd=8.71 sd=10.24 sd=10 sd=10.83 sd=11.02 +8.86 Improvement Test 2-3 -2.17 sd=9.41 sd=15.98 +6.96 sd=15.57 +4.46 p=0.14 Sample (21) sd=11. First of all.26 81.83 sd=15. so after eight weeks from the error correction and grammar revision activities the scores in all the classes are again similar.72 Sociali Control (25) sd=12.85 4 89.13 sd=12.6 sd=16.79 sd =18.05 sd=10.2 sd=11.48 sd=13. students in sample classes never outperformed students in control classes in the fourth test.61 Test 1-2 +5.59 p=0.25 +1.13 Test 3-4 +3.24 p=0.63 -0.97 88.03 p=0.79 +4.92 -0.95 sd=12.59 sd=11.2 sd=8.55 sd=9.87 p=0.7 sd=8.2 91.41 85 sd=11.7 Sample (23) L.32 sd=16. no difference can be found between sample 17 .05 sd=16.8 sd=11.9 sd=10.91 +1.11 81.28 sd=16.39 -0.18 88.43 +9.48 sd=15. All the students’ scores were calculated dividing the number of the correctly completed gaps by the total number of gaps.23 p=0.4 Sample (23) sd=13.85 sd=14.41 sd=17.6 +14 sd=17.09 sd=11.54 +6.4 sd=8.95 sd=6.43 sd=11.18 sd=11. Linguistico Control (22) Sample (23) L.98 Table 7 1 69.47 Test 1-4 +7.07 +10. the students’ improvement as well as standard deviation and the results of the statistical analysis of variance.59 sd=13.88 91.25 83.09 sd=12.52 p=0.95 sd=13.82 sd=10.86 80.94 -7.89 p=0.46 78.74 79.9 +5.11 -0.98 p=0.3 p=0. and this similarity could mean that the classes considered were at a similar level before the experiments.61 +1.04 -0.42 p=0.56 p=0. Tables 6 (first experiment) and 7 (second experiment) report both the mean scores calculated for each class in the four tests.76 Control (21) sd=10.

25 sd=20.61 sd=9.87 73.62 sd=10.08 65 sd=8.64 76.64 p=0.61 sd=15.6 81 -8.65 70.83 +14 sd=18.65 sd=18.39 p=0.67 sd= 0 sd=4.8 82.72 sd=10. students had to write a text.5 9 +20.2 74.67 +2.14 sd=12.43 sd=20.04 p=0.33 79.75 +23. then the number of the correct ones was divided by the total number of occurrences.4 -0.93 p=0.5 +14.05 p=0.8 -3.12 p=0.39 sd=14.4 sd=7.5 -11 +11.71 sd=17.8 +11 +6.44 sd=8.6 sd=14.15 sd=17.59 p=0.5 +35.41 p=0.07 Sample (5) L.59 Sample (6) L.6 sd=4.5 sd=10.75 83.5 +9.52 p=0. The following Tables 8 and 9 show a pattern which is quite similar to the one found in Tables 6 and 7: Again.65 p=0.75 83.4 71.84 p=0.94 sd=6.54 76 78.01 sd=5.74 sd=7.59 p=0.54 p=0.86 p=0.8 67 59.38 sd=7. neither in test 1 nor in test 4.17 77.65 sd=12.5 58.03 sd=12.54 sd=19.45 48 59.75 +11.02 p=0.53 p=0.96 sd=6.26 sd=7.94 p=0. As for the whole class.79 p=0.4 +7.25 +6 25.62 p=0.22 p=0.11 p=0.2 Texts In all the tests.2 +27.9 sd=16.05 p=0. Scienze Sociali Control (5) Sample (4) L.4 60 +16 +0.19 sd=11.4 -2.57 sd=8.5 sd=5.2 -7.25 +54.4 -3.13 sd=5.25 -3.5 sd=0 sd=3 sd=14.65 p=0.46 p=0.39 55.33 +12. Classico Control (3) Table 9 Test scores in formal exercises in experiment 2 for low proficiency students Test 1 53.2 sd=21 sd=6.22 sd=7.76 50 sd=9 29.49 62.17 sd=8.45 sd=18.63 70. no difference can be seen between the two low proficiency groups in neither experiment.09 sd=6.4 sd=9.31 p=0.64 p=0.38 p=0.32 sd=9.5 sd=16. Linguistico Control (5) Sample (3) L.85 sd=17.08 72.63 p=0.78 sd=18.75 sd=23. no difference between couples of parallel classes can be found.83 sd=14.1 sd=9.46 p=0.6 +15.2 57.5 +2 -6.97 sd=11. but some better results for the sample students either during the second test or the third.41 sd=14.67 sd=6.23 p=0.5 sd=12.81 p=0.27 p=0.26 sd=5.8 sd=6.25 +15. Linguistico Control (5) Sample (4) L.49 p=0.86 p=0.1 sd=11.67 0 +7.5 68.34 p=0.8 +10.8 sd=13. self-correction rates and scores and control classes.25 +11.86 p=0.95 sd=7.67 sd=10. Table 8 Test scores in formal exercises in experiment 1 for low proficiency students Test Improvement 1 2 3 4 Test 1-2 Test 2-3 Test 3-4 Test 1-4 64 77.8 +10.5 sd=5. the scores of the low proficiency students in both groups were compared and data underwent the non-parametrical analysis of variance.35 p=0. and this type of exercise aimed at establishing if students were able to use spontaneously the structures under consideration in a correct way and whether there was any difference between sample and control groups: The occurrences of simple present and progressive verbs were counted in each student’s text.19 sd=4.6 sd=15.75 -6 sd=12. in which the present tense was needed. Scienze Sociali Control (5) Sample (3) L.2 72.67 79.05 sd=11.07 p=0.5 sd=16.2 sd=24.08 sd=11.83 +13.35 sd=4.17 66. The other grammar components of the texts were not considered because the 18 . Classico Control (4) 3.67 +6 +12 +23.19 p=0.6 sd=11.9 sd=6.75 76 88 +5.Cooperative learning as a correction and grammar revision technique: Communicative exchanges.89 sd=5.15 sd=7.48 p=0.52 p=0.6 -9.5 sd=3 sd=8.12 p=0.6 sd=9.73 p=0.52 sd=6.24 2 69 sd=10.25 sd=15.33 86.11 76 85 +16.14 63.53 sd=12.5 +9.33 +5.4 +20.2 sd=21.91 p=0.93 46.07 p=0.78 p=1 p=0.5 52.98 sd=3.3.92 sd=11.45 66.87 sd=17.45 p=0. In other cases however.33 sd=14.31 sd=18.36 sd=11.9 sd=9.5 sd=19.99 sd=10.57 sd=10.28 sd=18.65 Improvement 3 4 Test 1-2 Test 2-3 Test 3-4 Test 1-4 69.52 58.43 sd=10.76 sd=6.25 72 87.

48 +5.77 sd=22.91 sd=10.15 +11.28 p=0.11 +9.08 p=0.89 -9.92 83. Classico Control (20) Table 11 Percentages of correct occurrences of present simple and progressive verbs in texts in experiment 2 Test 1 80.01 sd=11. The CL activities therefore.56 2 3 4 Test 1-2 Test 2-3 92.21 89.32 sd=22.49 p=0.04 sd=7.37 +8.32 sd=13.02 p=0.81 p=0.51 sd=9.35 -0.76 sd=8.8 +7.09 sd=8.67 p=0.91 sd=13.88 2 93.2 sd=25.79 sd=17. but this advantage is likely to be due to the better scores of the sample class during the first test.29 p=0.22 p=0.72 74.58 sd=8.85 sd=20.07 96.18 sd=13. Scienze Sociali Control (25) Sample (17) L.79 70.36 sd=29.13 sd=12.41 p=0.85 89.02 p=0.4 Improvement Test 3-4 +2 sd=9. Table 10 Percentages of correct occurrences of simple present verbs in texts in experiment 1 Test 1 82. the CL activities which sample students experienced seem to have had few positive effects on the accuracy of present simple and progressive forms in texts. 19 .21 sd=12.87 +2.07 p=0.47 p=0.43 92 94.86 sd=6.21 p=0.1 sd=13.15 sd=9.82 +5.72 +4. no difference can be found between classes.87 71.7 sd=12. Classico Control (20) The scores of the low proficiency students were compared also in texts and as shown in Tables 12 and 13.75 p=0.4 sd=14. In all the other cases.2 +4.82 sd=18.49 Sample (23) L.13 sd=10.47 p=0.37 Improvement 3 4 Test 1-2 Test 2-3 Test 3-4 Test 1-4 84.38 +9.75 83.25 p=0.05 96.06 +1.77 sd=12.51 sd=15.93 sd=10.08 93.49 62.38 p=0.47 sd=7.85 p=0.04 sd=14.85 83.22 89.95 sd=22.88 p=0.02 p=0.44 sd=8. Linguistico Control (21) Sample (22) L.56 sd=22.82 p=0.58 90.94 86.65 sd=18. Therefore.85 sd=7.41 +0.63 sd=18.6 -4.8 sd=22.48 94.14 p=0.86 +8.9 88.4 p=0.32 +10.79 sd=13.65 sd=10.29 sd=13.4 81.67 +1.8 p=0.76 sd=12.45 sd=12.43 +9. Scienze Sociali Control (23) Sample (17) L.13 p=0.31 83. neither in the long nor in the short term.63 sd=10. no remarkable difference was found.56 93.31 sd=18.49 sd=13.7 sd=11.38 sd=17. neither in the first nor in the second experiment.39 sd=6.38 Sample (22) L.85 p=0.02 sd=8.87 90.46 92 sd=8.Cooperative learning as a correction and grammar revision technique: Communicative exchanges.48 +2.05 sd=5.81 p=0.8 94.4 sd=6.76 +2.38 p=0.85 +1.69 sd=22.97 sd=8.24 +1 sd=16.13 sd=10.27 p=0.53 sd=6.26 91.95 +11.39 Test 1-4 +12.95 -2.04 +17.06 +5.53 p=0.36 sd=18.34 sd=9.68 p=0.58 sd=24.1 p=0. self-correction rates and scores correction lessons focused only on the revision of the present tense and no other structure.45 sd=12.87 p=0.2 sd=23.4 sd=15.46 sd=7.45 sd=11.99 sd=14 sd=19.89 sd=9.45 sd=17.5 p=0.23 +18.38 sd=9.64 +13.65 96.34 p=0.9 p=0 92.76 +10.58 sd=14.7 +1.62 84.48 sd=10.48 p=0.67 p=0.25 81.4 sd=13.73 -1.1 -0.35 +16.04 sd=21.17 +17.87 sd=10.4 sd=5.52 sd=20.57 sd=23.16 92.81 sd=17.14 sd= 24.79 93.43 sd=20.78 sd=15. the sample class in the Liceo delle Scienze Sociali performed significantly better than the control class in the second test.68 83.31 87.95 +6.84 sd= 9.39 sd=12.45 +5.05 p=0.36 p=0. The data in the following Tables 10 and 11 show that the percentage of correct occurrences are very similar in all the classes in both experiments: Only in two cases sample classes had better results than their respective control classes four weeks after the correction and grammar revision lesson.43 p=0.7 sd=12.9 sd=22.78 sd=17. In the second experiment.3 sd=24.25 80.14 p=0.33 sd=9.2 p=0. had no positive effects on low proficiency students’ accuracy in texts.76 sd=13.56 +3. as only in two cases better results for sample classes were registered.22 p=0.29 sd=16.83 -1.6 97 +4. Linguistico Control (22) Sample (23) L.02 82 89.05 +4.82 +7.13 sd=19.76 -5.95 p=0.98 sd=14.48 95.63 92.16 p=0.33 -0.56 92.58 p=0.5 sd=12.37 +8.67 79.45 89.23 +10.8 sd=23.17 p=0.

8 sd=4.17 +27 sd=23 +5.02 sd=6.64 sd=10.5 sd=17 sd=8.5 -3 +32.31 p=0. the discussion and comparison of different people’s opinions.42 +5.44 56.2 sd=12.5 sd=0.18 sd=10.41 sd=14.49 p=0. 4.75 +1 +34 sd=13. in particular the grammar revision within groups.64 p=0.79 sd=9.5 +28 sd=15.94 sd=9.4 sd=26. because they wrote that the sentences they had to examine were too easy for them.57 81 87.5 +7.5 75 sd=10. Discussion The data analyzed above show that the CL activities on correction and grammar revision had some positive effects on some of the aspects taken into consideration in this study.12 p=0.5 +6.4 sd=15 75.75 sd=15.69 p=0.81 p=1 p=0. appreciating different aspects.5 -17 -1.12 p=0.75 +11 -19.62 p=0.65 sd=8.2 sd=13.48 p=0.43 sd=20.5 +11.46 sd=7.5 sd=2.5 +4.91 79.89 97.5 sd=14.94 Sample (6) L.5 sd=21.05 p=0. All the students liked the activities.47 sd=4.75 +25 +24.Cooperative learning as a correction and grammar revision technique: Communicative exchanges.6 sd=24.62 sd=27.67 +16 +12 sd=6.5 -9.88 sd=11.46 p=0.5 sd=12.38 82.37 2 3 89.43 78 90 +18.5 sd=6 sd=16.75 +12 +33. Some students (12%) however did not find the activities very useful for their learning.58 sd=33. self-correction rates and scores Table 12 Percentages of correct occurrences of simple present verbs in texts in experiment 1 for low proficiency students Test 1 79. and the awareness of a higher self-correcting ability after the CL activities.36 65.5 82.34 sd=15. Scienze Sociali Control (25) Sample (3) L.25 sd=2.41 p=0.5 sd=19.5 sd=19.35 p=0.13 p=1 73.57 46.12 p=0.5 sd=7.84 96 84 sd=5.25 sd=19.06 p=0.5 +8.2 sd=12.5 sd=12.78 p=0.25 +19.16 p=0.14 sd=11.5 sd=6 sd=25.99 74.22 sd=18.13 p=0.44 p=0.72 Improvement Test 2-3 Test 3-4 Test 1-4 +5.28 4 94.5 +2 +27. the students who worked within CL groups corrected successfully many more mistakes than 20 .17 p=0.5 sd=6. Linguistico Control (5) Sample (4) L.33 p=0.66 p=0.8 p=0.5 p=0.4 Questionnaires Anonymous questionnaires were given to students who experienced the CL activities in order to collect their opinions.35 p=1 -8.65 p=0.02 sd=7.6 sd=33.6 +20 +14. Classico Control (4) 3.55 p=1 p=0.6 72. most students (88%) found them useful for their learning and wrote positive comments.8 sd=27.66 p=0.44 +22.5 sd=13.79 82. First of all.44 Improvement 3 4 Test 1-2 Test 2-3 Test 3-4 Test 1-4 70. Scienze Sociali Control (5) Sample (4) L.5 sd=3 sd=13.25 sd=9.85 p=0.52 p=0.86 sd=27.2 sd=14.98 sd=8.33 sd=17.5 78.75 54.17 65 90 +14.2 +5.5 54.42 sd=22.81 p=0. Classico Control (3) Table 13 Percentages of correct occurrences of simple present and present progressive verbs in texts in experiment 2 for low proficiency students Test 1 66.75 sd=28.5 p=1 2 90.5 sd=5.25 sd=5.53 -0.5 88 86.35 sd=6.58 p=0.75 sd=13.34 sd=14.75 +2.66 sd=13.35 p=1 81.84 sd=13.5 -14.81 -6.92 p=0.81 100 sd=0 95 sd=8.61 sd=8.5 sd=5 59.25 +25.75 93 sd=25.5 sd=13.6 sd=4.5 sd=5 +13 0 +18.4 sd=21. Linguistico Control (5) Sample L.11 Sample (5) L.48 56.33 sd=22.6 +6.75 sd=13.65 sd=17.5 sd=9.46 sd=13.75 90.46 sd=15.11 sd=8.14 76 sd=12 91 sd=9 p=0.6 sd=23.5 sd=22.6 +41.09 74 92.66 sd=7.25 sd=9.4 93.33 p=0.4 sd=7.44 p=0.4 +22.75 sd=19.81 59 91.39 Test 1-2 +23.81 sd=30.81 p=0.24 66.57 p=0.58 +18.49 p=0.8 sd=18.2 sd=27.42 p=0.62 85.68 p=0.66 p=0.7 p=0.09 sd=21.

in some cases in tests 2 and 3. this type of CL activity does not seem to have a particularly positive effect on sample students. motivated and compared the different corrections. 5. because some advantages are shown by some sample students in formal exercises. for both the sample classes and the low proficiency students in them. with an impact which is deeper for some linguistic structures (Saxon genitive. personal pronouns) were under examination. possessives. in particular for the self-correction rates: also in Servetti (2009). self-correction rates and scores individual students. in which four parallel classes (83 students) from the first year of the Italian secondary school were involved in the same project using the same methodology. Secondly. although in many cases sample and control groups had similar results. and in some cases explained each other some grammar rules. in the study by Servetti (2009). and this could be due to the composition of the CL groups: Mixed ability students had more chances to correct mistakes properly than students who worked on their own. no such a sharp difference in results can be found in this study. instead. However. possessives. no cases were found in which the 21 . and this means that the atmosphere within groups was truly cooperative. Some similar results can be highlighted. In test scores. in which the students’ attention is focused on form. although in the study by Servetti (2009). but in this study no such difference was found between groups. It can be however hypothesized that CL activities on correction and grammar revision may influence students’ accuracy in a positive way.05) in four cases out of twelve. neither for the sample classes nor for the low proficiency learners in them. the comparison of the self-correction rates between sample and control classes shows better results for the students who worked within CL groups. a result shared also by low proficiency students. either in test 2 or 3. In texts. Conclusion The data reported in this study show some better results for the students who experienced the CL activities and. sample classes self-corrected more than control classes and the advantage was statistically relevant in half of the cases. This study adds further data to the study by Servetti (2009). but not by all. although other English structures (Saxon genitive. Therefore. the difference between sample and control classes does not seem to be so clearly defined: In formal exercises better results were scored by sample students in both experiments. In texts the difference between groups is fuzzier: only in few cases sample students outperformed control students and no difference can be found in low proficiency students’ scores. however. in which students have to convey meaning and probably pay less attention to form than in formal exercises. The analysis of the tape scripts also show that students discussed about grammar rules in the greatest majority of the sentences. or of a higher motivation in performing the self-correction task. it can be hypothesized that the CL activities performed might have had quite a positive influence on students’ accuracy in formal exercises. most sample students produced more accurate texts than control students (and the difference was statistically significant) in text 2 (1 week after the CL activity). This advantage could be the result either of a higher level of attention during the CL activities or during plenary. in others in the medium (4 weeks) term. all the sample classes performed significantly better than control students during test 3 (4 weeks after the CL activity) in formal exercises. as well as by the low proficiency ones.Cooperative learning as a correction and grammar revision technique: Communicative exchanges. Every sample class in fact had higher self-correction rates than its respective control class and the difference was statistically relevant (p<0. Moreover. being able to correct successfully more mistakes than control students. in some cases in the short term (1 week). personal pronouns) than for others (present tense).

it can however be assumed that it helped students. M. 10. DeKeyser. (1992). (Eds. R. & Fagan. March 9-11). R. Valencia (Spain). School and classroom organization. Ortiz. P. Truscott. Learning (September). NJ: Erlbaum. Error feedback in L2 writing classes: How explicit does it need to be? Journal of Second Language Writing. K. R. Even though it is not sure that the results in favour of sample students are the direct consequence of the correction and grammar revision activities through CL. Correcting students’ written grammar mistakes through cooperative learning: A case-study. Proceedings of INTED 2009 (International Technology. (1994). Mathematics achievement in cooperative versus individualistic goal-structured high school classrooms. The case against grammar correction in L2 writing classes. Johnson. I. 3267-3277. Journal of Educational Pedagogy. Sherman. (Edited by Nicole and Lily) 22 . Martì Belenguer. (1991). Ream. M. (1988). increasing their level of attention and probably reinforcing their knowledge of some structures. R. & Johnson. In: Dupuis. Therefore. D. 11. Education and Development Conference). (Ed. S. & Johnson. Johnson. R. B. Journal of General Psychology. 203-218. Cooperative learning and student achievement. & Thomas. IATED. Foreign Language Annals. (1990). Johnson. 60-66. Ur. (Eds. W. Teacher education: Reflection and change. 454-459. R. Michaelsen. & Johnson. S. L. D. (1985). R. Hilsdale. Language Learning. Effects of cooperative and individualistic experiences on interethnic interaction. Cambridge: Cambridge University Press. Johnson. in order to add further data to this study. 118. 341-347. CL as an error correction and grammar revision technique could be used as an alternative to traditional correction.). (1991). Joining together: Group theory and group skills. R.. & Johnson. L. Further research on the matter should be however conducted. (1995). E. A course in language teaching. J. D. (1981). Johnson. 46. which have been found also for the low proficiency students in sample classes. 271-273. To Improve the Academy. & Smith.. The students’ involvement in the CL activities could justify the higher self-correction rates and some higher levels of accuracy in formal exercises. Groups of four: Solving the management problem. D. self-correction rates and scores control classes outperformed the sample classes.. (1989). (1986). R. In: Gomez Chova. interaction and attitudes. Pate. J. P. allowing them to talk about and revise grammar rules in an active way. 2009. (1988). D. M. Edina. 79(6). L. (1996). D. A. 23(3). Johnson. 46-51. Servetti. R.). and achievement in cooperative learning groups. Selected effects of cooperative learning. Impact of positive goal and resource interdependence on achievement. Studies in Second Language Acquisition. T. & Candel Torres. Team learning: A comprehensive approach for harnessing the power of small groups in higher education to improve the academy. MN: Interaction Book Company. Leki. Johnson. Cooperative grouping in the classroom. (1996). Vermette. which were appreciated by the greatest majority of students. 24. K. (1988). 77(1). USA: Pennsylvania State University. In: Slavin. Journal of Educational Psychology. Social Studies. I. (2001). 79(3).. MN: Interaction book company. 327-369. 287-289. Social Education. March 9-11. M. 52(4). 17(3). Ferris. M. Johnson. E. Cooperation and competition: Theory and research. Oral discussion. Yager. & Stanne. S. Endina. (2009. 107-122. References: Burns.). D. E. M. Research on reducing prejudice. Valencia (Spain). Journal of Educational Research. Active learning: Cooperative in the college classroom. G. (1981). Slavin. & Roberts. 161-184. group-to-individual transfer. (1991). D.Cooperative learning as a correction and grammar revision technique: Communicative exchanges.. 169-172. R. The preferences of ESL students for error-correction in college-level writing classes. Boston: Allyn and Bacon. A. 379-410. Learning second language grammar rules: An experiment with a miniature linguistic system.