The Effects of Feedback Types on Learners ’ Recognition of Lexical Collocations

This study investigated the effects of three types of feedback on Iranian EFL learners’ recognition of lexical collocations. 78 pre-intermediate students from among 90 were selected to participate in this study. A KET (Key English Test) was used in order to homogenize them. A teacher-made pre-test including all 150 target lexical collocations was used to make sure that the students did not know the target lexical collocations beforehand. Then the participants were divided into three experimental groups. All groups received the same instruction during 10 sessions of treatment. However, they received different types of feedback. Each session, the teacher gave a handout including 15 English sentences to each student. There was one lexical collocation in each sentence which was written in parentheses. Students were asked to write 15 collocations within twenty minutes and then submit the paper to the teacher. Then the teacher underlined the errors and gave them back to the students. Students in group A received direct feedback from the teacher. Students in group B received indirect feedback and the last group received peer feedback on their collocational errors. At the end of the treatment, 30 multiple-choice items were used to test students’ recognition of lexical collocations. The result of One-Way ANOVA procedure revealed that the group that received indirect feedback had the best performance, followed closely by the group that received peer feedback. Students in direct feedback group had the lowest performance. The findings suggest that different types of feedback have different effects on EFL learners’ recognition of lexical collocations.


Introduction
Learning a language involves learning both grammar and vocabulary.However, there is evidence that even without using grammar accurately, communication can continue, but without using vocabulary correctly, understanding will be disrupted.Carter (2001) states that "knowing a word involves knowing its spoken and written contexts of use; its patterns with words of related meaning as well as with its collocational patterns; its syntactic pragmatic and discourse patterns" (p.43).
Knowing collocations is one of the most important parts of knowing a word.Lesniewska (2006) points out that collocations are word combinations that are not exactly fixed.In spite of difficultly, learners and teachers are interested in collocations.This is because as Shin and Nation (2007) state, it helps learners to develop "language fluency and native -like selection of language use" (p.2).Fahim and Vaezi (2011) note that lack of collocational knowledge makes learners sound odd and not competent in using language.When learners do not know two or more words that collocate with each other, they use some long and complicated sentences instead to express their idea.To help irradicate this problem, teachers are always in doubt whether they should correct students' errors, and if so, what kind feedback will be the most effective?
As Semke (1984) points out, traditionally linguists used to assume that if errors were not corrected, learners would repeat them in future and using language correctly would be difficult for them.However, Truscott (1999), in response to the question of whether teachers should correct errors, holds that teachers should "make decisions about what to do and what not to do in their classes" (p.121) according to certain conditions of each class.He adds that these decisions are never made under conditions of certainty.

Statement of the problem and purpose of the study
Although there are many studies (Balci & Cakir, 2012;Ganji, 2012;Karoly, 2005) that show the positive effect of collocations in vocabulary learning and teaching, many teachers still continue teaching new words using traditional ways such as definitions, synonyms and antonyms.Providing feedback is another factor that helps learners to be aware of their progress during their learning.It can be used in a variety of forms in different classes.With respect to the merits of providing feedback in language learning, the aim of this research is to find answer to the following question: Are there any significant differences among the effects of feedback types on EFL learners' recognition of lexical collocations?Huang (2001) classifies collocations into four groups.He notes that on the one end of the collocational continuum, there are free combinations that are dramatically productive.On the other end, there are idioms that have the least degree of productivity.Two other types of restricted collocations are placed between these two ends.So, the four types of collocations along the continuum include: 1) Free Combinations, 2) Restricted Collocations, 3) Figurative Idioms, and 4) Pure Idioms (p.3).

Classifications of collocations
On the other hand, Wei (1999) believes that there are three collocational categories: 1) Lexical collocations: "lexical collocations are recurrent word combinations that involve mainly content words, such as perform an operation but not perform a trip" (p.5).2) Grammatical collocations: "grammatical collocations are recurrent word combinations that involve mainly preposition or grammatical structure" (p.5), such as wait for but not wait to.
Idiomatic expressions: idiomatic expressions are the most frozen word combinations so that it is almost impossible to substitute one element with another like give somebody the ax, as a matter of fact.Balci and Cakir (2012) classify collocations into two types.They define lexical collocations as "combinations of noun, adjective, adverbs and verbs such as verb + noun, adjective +noun, noun + noun and verb + adverb" (p.23).They add that "grammatical collocations are composed of content and grammatical words" (p.23).
Similarly, Moehkardi (2002) is of the opinion that transfer of L1 elements in students' combinations are the more common source of collocational errors.In addition, Jukneviciene (2008) also claims that to compensate their lack of academic vocabulary knowledge, Lithuanian learners rely on L1 translation when creating collocations.
Hong , Abdul Rahim, Hua and Salehuddin (2011) investigated the types and sources of verb-noun collocatinal errors in EMAS (The English of Malaysian School Students), a sub corpus of Malaysian learners groups.The data was essays written by 872 students, each essay containing about 270 words.The inter-language theory was the base of this study and error analysis was used for analyzing data.To process the data, the linguistic software of word smith tools was used.First, collocations in the groups were identified.Then, the Oxford Collocation Dictionary (2009) and the online British National Groups were applied to determine the acceptability of collocations.After identifying the errors, they were classified into seven types of error.Errors were related to verb, noun, preposition, determiner, two types were related to the way elements were combined, and one was about singular or plural use of nouns.It was found that preposition was the most problematic linguistic category in writing of Malaysian learners, followed by verb errors; noun errors were the third erroneous items.There were three major sources of errors: inter-lingual transfer, intralingual transfer and paraphrase.Results revealed that among these three major sources of collocation errors, intralingual transfer had the most significant role.

Insufficient exposure
Insufficient exposure is another common source of difficulty in the recognition of EFL learners' collocations.There are many studies that confirm this finding.(Jing, 2008;Shokuhi & Mirsalari, 2010;Shin, 2007).In support of this point of view, Shokuhi and Mirsalari (2010) state that one of the main sources of learners' difficulty and unfamiliarity with English collocations is insufficient exposure.They also believe if learners encounter a certain type of collocation, they will comprehend it better.
In a study by Shin (2007), it was revealed that Korean students use artificial teeth instead of false teeth, lying story instead of tall story, etc.Like Shokuhi and Mirsalari, Shin (2007) notes that such wrong uses of collocations happen because learners do not encounter these word sequences repeatedly.So they rely more on translation from their first language.
Ying (2009) refers to six issues that cause collocational errors for Chinese English and non-English majors.However, it seems that they are sources of collocational errors for all EFL/ESL learners and they can be generalized.First, he reports that there is no detailed explanation of what collocation is in Chinese university textbooks.Second, lexical transfer or overgeneralization is another factor.Third, grammatical irregularity may be the cause of many misused collocations.Fourth, the collocatioal errors may be the result of a semantic choice error.Referring to Lombard (1997, p. 85), Ying regards semantic choice as using a near-synonymy English word instead of using appropriate word for that special context.Fifth, these errors may result from the existence of different categories of words in English.Content words are open sets and can be changed and added in different contexts.Due to this property the acquisition of lexical collocations become more difficult than grammatical words that are limited in number and are closed sets.Finally, miscellaneous uses are another reason for collocational errors.Although they may be similar to collocations, their meanings are different from the intended target collocations.

Some possible solutions
Wei (1999) suggests some exercises, techniques, and activities that can be used by teachers in the instruction of collocations.
1) Peer correction, he considers peer correction as a student-centered procedure and an effective technique in learning collocations.
2) Sentence making (individually or in a group) Nesselhauf (2003) makes the following suggestions: First, teaching only lexical elements that go together is not enough; teaching all combinations such as prepositions, articles, etc., are necessary.For example, teaching learners pass judgment on has more positive effect than teaching pass judgment.Second, emphasis should be put on L1-L2 differences.Learners tend to produce the L1 equivalents while they have learned the correct collocations.So, L1 and L2 collocations should be contrasted.Third, in teaching verb-noun collocations, emphasis should be put on the verb, because it is the main source of difficulty.Moehkardi (2002) believes teachers should address collocations when they see them as a by-product of other skills.They should make students aware of these word combinations.Teachers should not introduce words as individual words but as lexical units.

Feedback
There are different opinions about giving feedback.Erdogan (2005) points out that the process of second or foreign language learning is similar to first language acquisition process.Children commit plenty of mistakes while learning their mother tongue and they receive feedback from adults during the natural acquisition process.By getting feedback, they learn how to produce acceptable sentences in their first language.He believes learning a foreign language is no exception.On the other hand, Semke (1984), who investigated the effects of correction on 141 German students, asserts that correction does not increase learners' language competence.Similarly, Glover and Brown (2006) note that feedback provided to students is not often effective.

Direct Corrective Feedback
In direct corrective feedback, teacher provides the correct form of the language for students.(Bitchener, 2008;Ellis, 2009).Similarly, Ellis et. al., (2008) state that "direct feedback entails supplying learners with the correct target form" (p.365).
In a study conducted on the effect of written corrective feedback on 75 low intermediate international ESL students in New Zealand about the use of English article system, Bitchener (2008) came to the conclusion that students who received two kinds of direct corrective feedback (written as well as oral metalinguistic explanation and direct feedback without metalinguistic feedback) performed much better than the control group that did not receive any kind of corrective feedback.

Indirect Corrective Feedback
In indirect feedback, the teacher indicates that there is an error but s/he does not actually correct it.Indirect feedback has different modes such as underlining the errors, placing a cross in the margin by the line in which error has occurred.(Bitchener, 2008;Ellis, 2009;Ferris & Robberts, 2001;Robb, Ross & Shortreed, 1986).

Peer feedback
There are many studies ( Noonan & Duncan, 2005;Rollinson, 2005;Lima, 2011) that have been done on the effect of peer feedback on different language skills, and they have revealed that it has positive effect on language learning.Roscoe and Chi (2004) believe that tutors learn as a result of instructional explanation and monitoring their understanding while they are teaching.Results of a study by White (2009) shows that students have positive views toward peer feedback.Rollinson (2005) mentions some disadvantages for peer feedback.Peer feedback, whether oral or written, is timeconsuming.Other problematic aspects of peer feedback are students' characteristics and teachers' roles.Some students cannot accept their peers as substitutes for their teacher; it is also difficult for teachers not to interfere and leave providing feedback only to students.Teachers also doubt the implementation of peer feedback in particular situations.Nagano and Kitao (2008) considered two main issues that the present study tries to investigate, collocation and feedback.Nagano and Kitao's (2008) study investigated the role of negative feedback on the acquisition of collocations.Participants were 41 Japanese learners of English who were living in Japan.However, a limited number of them were university graduate students living in the U.S. The effect of proficiency was controlled in this study by dividing participants into three groups (beginners=13 participants, intermediate=22, and advanced=6).Classification was done according to answers to a questionnaire and students' self-evaluation of their proficiency in English.In addition to the classification of participants, in order to control the effect of learners' L1 on the acquisition of collocations, collocations were also classified into three groups.This study was done online by utilizing "a module written in flash" (p. 6).
Collocations were shown one by one on the screen as well as a picture that presented the interpretation of each collocation in the module.The acceptability of collocations was rated by learners using a 5-point Linkert scale (from good to not good).The grammatical judgments and reaction times of the participants were recorded.44 English collocations were randomly selected from different sources.There were three types of collocations.Type-0 was collocations that were grammatical in both languages (English and Japanese) such as a bicycle chain.Type-1 was collocations that were grammatical in English but not in Japanese like a strong student.Type-2 was collocations which were acceptable neither in English nor in Japanese.
The results showed that high-proficient learners did not perform better than the two other groups.It also turned out that collocation is a source of difficulty even for advanced learners.In addition, it was revealed that proficiency level had no significant effect on collocation learning, and that students transferred their L1 collocational knowledge to L2.It was also concluded that English collocations which had grammatical equivalents in L1 were rated higher than those which had ungrammatical equivalents in L1.
In another study Fahim and Vaezi (2011) studied the effect of visually enhanced input on the learning of verb -noun lexical collocations.In this study 128 male students, from different language schools in Esfahan, Iran were chosen.The Comprehensive English Language Test (CELT) was administered to homogenize and to choose the required number of subjects.Among them 96 subjects participated voluntarily and they were divided into three groups.Pre-test was used to ensure that all participants in three groups have the same knowledge of verb -noun lexical collocations.
The participants were randomly divided into two experimental groups and one control group.For eliciting the target items in question, pre-test was administered to all three groups.There were ten instructional treatment sessions.
Students were given reading passages that had the same semantic context.The first experimental group one (EG1) noticed the Verb -Noun lexical collocations saliently in the reading passage because those collocations were bolded or capitalized.In other words, they were thought via implicit means of visual/textual input -based treatment.
Conventional instruction was used for the second experimental group (EG2).In this method, the instructor presented the rules and examples, and then students were expected to engage in an immediate production task.In the comparison group (CG), students read the texts and asked for help if they had difficulty.After the treatment period, a post-test including 30 -multiple-choice items was administered.Results revealed that although groups one and two received different types of intervention, there were no significant differences between these two groups.In other words, implicit method of formal teaching like visual/textual input enhancement has the same positive effect as conventional instruction which involves a lot of practice and explanation.
In a different study, Noonan and Duncan (2005) studied teachers' use of self-assessment and peer-assessment as classroom assessment strategies.They believe formative assessment helps students to improve their performance by involving them in the learning process.They hold that formative assessment involves teacher feedback to students and one of its components is peer and self-assessment.They also emphasize that peer assessment is an integral part of formative assessment.In this study data were collected from 118 high school teachers' assessment practice in a school in Western Canada.
The survey involved both 34 forced choice items and some open-ended questions about teachers' assessment practices.One of the questions was related to the extent to which teachers use self and peer-assessment in classrooms.The answers were analyzed in three stages.First, the responses were classified according to the subject taught, three general categories of mathematics and sciences, social studies and English, and others were used.Then the teachers' responses were classified into two categories A) those who reported they used the mentioned strategies in their classroom and B) those who reported they did not use them.Group A were reviewed again for receiving more specific information.Some reported 'little' use of strategies and some reported 'some' use of them.
At the third stage, a constant comparison form of analysis was used to explore the nature and frequency of teachers' use of peer and self-assessment.Results showed that 24% of the teachers reported that they did not use those strategies.They reported different reasons such high school students' capability of being truthful and objective in self and peerassessment.49% reported they used these strategies 'a little', 5% used them 'seldom' and 27% reported using strategies 'somewhat'.Teachers who reported using strategies 'a little' mentioned that these strategies facilitate students' reflection on their achievement.They were effective in group work assessment and in students' projects and presentations assessment.Results also revealed English teachers and social studies teachers used these strategies more frequently than math and science teachers.

Participants
The participants of the present study were initially 90 pre-intermediate level students in two language institutes in Abhar and Hidaj, Iran.Participants were both male and female students, ranging in age from16-30.They were high school students who had been studying English for about 8 semesters in the institutes.Some participants were university students with different majors who were studying English in the institutes.So, in order to homogenize them, all participants were given a KET (key English Test) test.Students with very high and low proficiency level were eliminated from the study.As a result, 78 participants remained, who constituted the participants of this study.It should be mentioned that there was no control group in this study and the number of male and female students was not equal in each group.

KET TEST
The vocabulary subtest of the Key English Test (KET) was given to all 90 participants in order to homogenize them.The test contained 20 items in multiple-choice format.

A teacher-made pre-test
To see whether or not the students had any prior knowledge of the 150 lexical collocations selected for treatment, a teacher-made pre-test was used.It included all 150 target lexical collocations and was in fill-in-the blanks format.

A handout of sentences including lexical collocations
During ten treatment sessions, ten handouts were given to each student (one handout in each session).Each handout included 15 English sentences including 15 lexical collocations (one collocation in each sentence).The Persian equivalents of lexical collocations were given in parentheses.

Post-tests
A post-tests was used in order to test learners' recognition of lexical collocations after receiving different types of feedback.It was constructed by the researcher and included 30 multiple-choice items which were randomly selected from among the 150 target lexical collocations.

Procedures
The first step was the selection of 90 pre-intermediate-level participants from two language institutes in Abhar and Hidaj, Iran.Second, they were homogenized using the vocabulary subtest of a KET test.Based on the results of this test, 12 participants were excluded from this study.Then a teacher-made, fill-in-the blanks test including all the 150 target lexical collocations was administered to make sure that the students did not know the selected lexical collocations beforehand.Results showed that students knew none of the selected collocations.
Participants were divided into three experimental groups.All groups received the same instruction but with one of the different types of direct, indirect, or peer feedback.The treatment period lasted 10 sessions.Each session, the teacher gave a handout including 15 English sentences.There was one lexical collocation in each sentence that was written in Persian and was put in parentheses.Students were asked to write the English equivalents of the 15 Persian lexical collocations within twenty minutes in each session.The teacher checked them and underlined lexical collocational errors.As mentioned before, each group received different types of feedback, as explained below.It should be mentioned that students in all three groups were allowed to use dictionary in order to find the correct words that collocated with each other, after teacher underlined their errors.
In group A, Direct feedback was provided to the students.After collecting the papers, the teacher underlined the lexical collocational errors and gave the papers back to the students and wrote the correct form of all collocations on the board for students.Students in group B received indirect feedback on their lexical collocational errors.After underlining their lexical collocational errors, the teacher gave students' papers back to them.Students were asked to correct their errors at home and bring it next session.Next session, the teacher checked the papers again and wrote correct form of collocations on board if they still existed.
After underlining lexical collocational errors, the teacher gave papers back to students in group C to find and write the correct lexical collocations.Students in this group were divided into small groups of three and worked with their peers to find the correct collocations within 30-40 minutes.At the same time, they received feedback from their peers in the same group.Then the teacher went over each group and if students had not corrected lexical collocational errors yet, the teacher provided the correct form of collocations for them.Thus, this group received peer feedback.

Data Analysis
A One-Way ANOVA procedure was used in this study to analyze the effect of types of feedback on EFL learners' recognition of lexical collocations.

Investigation of the Research Question
The research question aimed to investigate if there were any significant differences among the effects of feedback types on EFL learners' recognition of lexical collocations.To this end, a One-Way ANOVA procedure was utilized for analyzing the data.Table 4.1 presents the descriptive statistics.Based on the results of Table 4.1, it can be observed that the indirect group has the highest mean ( =28.32), followed closely by the group that received peer feedback ( =27.25); coming third is the group that received direct feedback ( =20.92).The implication is that the indirect and peer feedback are more effective than direct feedback on learners' recognition of lexical collocations.
In order to see whether there are statistically significant differences among the groups, the One-Way ANOVA procedure was run.The results of the ANOVA procedure are presented in Table 4.2.The observed F value and the significance level shown in Table 4.2 are indicative of significant differences among the groups.Thus, it can be claimed that the differences among the three groups are statistically significant.At the same time, the index of the strength of association indicates that 33% of the total variance in the dependent variable (recognition of collocation) is accounted for by the independent variable; namely, feedback types.This means that the remaining 67% of the variance is left unaccounted for.In order to locate the differences between the means, a post-hoc Sheffe test was utilized.The results of the post-hoc comparison are presented in Table 4.3.The mean difference is significant at the 0.05 level The above results show that the difference between the effects of direct and indirect feedback on learners' recognition of lexical collocations is significant.The indirect group performed better than the direct group on the post test.As a result, it can be concluded that the difference between the effects of direct and indirect feedback on the learners' recognition of lexical collocations is statistically meaningful.
In addition, the results indicate that there is a significant difference between direct and peer feedback groups' recognition of lexical collocations.The results show that the participants who received peer feedback outperformed the participants of the direct feedback group.As a result, it may be concluded that the difference between the effects of direct and peer feedback on the learners' recognition of lexical collocations is statistically significant.
A look at Table 4.3 makes it clear that although there is a difference between the effects of indirect and peer feedback on learners' recognition of lexical collocations, the difference is not statistically significant and learners in both groups did nearly the same on the post test.The following graphical representation (Figure 1) shows the results more clearly.

Discussion
The findings of the study revealed that although the difference between the performance of the indirect and the peer feedback groups was not significant, the indirect group performed better than the peer feedback group in recognition of lexical collocations.The study also showed that receiving direct feedback did not help learners to improve their recognition of lexical collocations dramatically, and this group had the lowest mean in the study.It can be concluded that various types of feedback have different effects on EFL learners' recognition of lexical collocations.
Some of the findings of this study are similar to those of the previous studies (Moehkardi, 2002;Nesselhauf, 2003;Pishghadam et. al., 2011;Shokuhi & Mirsalari, 2010;Wei, 1999;Ying, 2009) in that they, like the present study, put emphasise on improving learners' knowledge of collocations for communicating easily and fluently.But the main difference of the present study with the above mentioned studies is in that, they emphasized improving learners' knowledge of collocations using various methods other than using feedback.For example, in the study conducted by pishghadam et. al., (2011), form-focused and meaning-focused instructions were used to improve learners' collocational knowledge.
In another study, Fahim and Vaezi (2011) investigated the effect of visually enhanced input on learners' acquisition of collocations.Another difference of Fahim and Vaezi's study with the present study is in that they utilized classical techniques like definitions, synonyms, antonyms and mother tongue translation to teach new vocabulary to students.
Their study was in accordance with the present study in that it focused only on lexical collocations and not grammatical collocations.Unlike the present study in which the participants were adults and were told what collocations meant in the first session of the treatment, Balci and Cakir (2012) taught collocations to young students without referring to the word collocation in their study.
Another significant difference of the present study with other studies is using different types of feedback to improve learners' collocational knowledge.Unlike the study which was conducted by Frantzen (1995) and showed that there is no significant difference between different types of feedback, the results of the present study revealed that there are significant differences among different types of feedback.It is worthy to note that Frantzen' s study was on writing while the present study was on lexical collocations, and this may be the reason for this difference.Another possible reason may be due to learners' different cultural background.In Frantzen's study, participants were Spanish, but in the present study, participants were Iranian.
Another study by Hyland (2003) showed that indirect corrective feedback causes misinterpretation and misunderstanding.However, it should be noted that in the present study indirect feedback did not lead to these problems maybe because learners knew that they were expected to write only collocations.On the other hand, the Persian equivalents of what students were expected to write were given to them in parentheses.
Unlike Hyland's (2003) findings, one of the findings of the present study is that learners achieved better results in learning collocations when they received indirect feedback.One reason for this finding may be due to the educational system of our country.In our educational system, students often receive direct feedback from their teacher; indirect feedback is something new in our educational system.This type of feedback may encourage students to search new materials and words on the net and different sources and engage them more in the learning process.
Another result of the present study is that learners who received peer feedback performed better than their counterparts who received direct feedback.Results also revealed that the peer feedback group did not perform as well as the indirect group, but the difference between the two groups was not significant.It is worthy to note that much like indirect feedback, giving peer feedback is not very common in our educational system.This may be one reason why the learners achieved good results when they received peer feedback.
In accordance with the present study, Noonan and Duncan (2005) collected data from high school teachers to see whether they use peer-assessment in classroom as assessment strategy.Those who reported they did not hold that high school students are not able to assess their peers' works.The finding of the study were similar to the present study in that participants of the present study who received peer feedback were also high-school students, but the difference is that in the present study students were not assessed by their peers, but they were helped to find the correct collocations by their peers.This may account for the improvement in the peer feedback group in the present study.
In another study, Nagano and Kitao (2008) came to the conclusion that learners' collocational knowledge did not improve even in advanced level may be because they did not receive negative feedback on their collocational errors.In comparison with the present study, we can conclude that giving different types of feedback either positive or negative can improve learners' collocational knowledge.
Another finding of the present study is that giving direct feedback was not very effective on improving learners' collocational knowledge.This result is in contrast with Bitchener's (2008) finding, which revealed that students who received direct feedback performed better than the control group.One possible reason for such a difference may be partially due to the fact that in the present study direct feedback was compared with two other types of feedback (indirect and peer), and it was revealed that it was less effective than the two other types of feedback.But in Bitchener's study direct feedback was not compared with any other kind of feedback and students in the direct feedback group were compared with a control group.
The above mentioned conflicting areas are indicative of the need for further research.Compared with most of the above mentioned studies which were carried out in ESL setting, the present study was conducted in an EFL context and perhaps this is what makes this study different from other studies.

Conclusion
The present study aimed at investigating the question of whether there are significant differences among the effects of feedback types; direct, indirect, and peer feedback on Iranian learners' recognition of lexical collocations.The results showed that the group which received indirect feedback had the best performance, followed closely by the group which received peer feedback.The lowest performance was related to the group which received direct feedback which was noticeably lower than that of the other two groups.By way of conclusion, it seems that using different types of feedback is an effective method for teaching and learning collocations, but the amount of their effectiveness needs to be further explored.

Table 4 .
2The results of the ANOVA on learners' recognition of lexical collocations