Researchers have always been interested in code-switching (CS) and the interchange between languages. Richards and Schmidt (2010) define code as “a term which is used instead of language, speech variety, or dialect. It is sometimes considered to be a more neutral term than the others” (p. 87). The concept of code was first introduced by Bernstein (cited in Jingxia, 2010). Moreover, Nunan and Carter (2001) characterize CS as “a phenomenon of switching from one language to another in the same discourse” (p. 275). This research interest is intensified when it is realized that not only English is spoken in the EFL classroom, but also first language (L1) can be used. In addition, researchers have become more curious about merits and demerits of using L1 when they were involved in doing CS while teaching in the EFL classroom. In FL classrooms worldwide, using CS is a frequent practice.

English has been practiced as a foreign language in Iran and the necessity of learning it is growing every day. But there have always been heated controversies regarding using L1 in the EFL classroom setting. Recently, teachers have been expected to teach good quality English through English only classrooms because they have low chance of getting English input from the environment (Chi, 2000; Cook, 2001). However, Nordin, Ali, Zubir, and Sadjirin (2013) assert that there are some cases in which teachers have no choice except resorting to CS that “always comes with a price” because many teachers believe that CS should not be supported.

The present study intends to look for making a contribution to the way CS has been viewed, in general, and the various functions using CS in foreign language education can have, in particular. A major objective of the study is to make recommendations for teachers of foreign languages according to the results of the analyses regarding the students' personal attitudes across various proficiency levels toward teachers’ CS. As teachers’ CS and the students’ attitudes across various proficiency levels toward this phenomenon have been scantly paid attention to systematically in foreign language classrooms, the study seems to be of significance.

LITERATURE REVIEW

Code-Switching is a dynamic phenomenon of language use as a result of which the definitions of CS have alternated during time. Ting-Toomey and Chung (2005) refer to CS as switching to another language or dialect to control intergroup distance. Also, Cook (2000) concludes that CS is the process of “going from one language to the other in mid-speech when both speakers know the same languages” (p. 83). Levine (2011) called it a code choice which is made by the speaker.

There have been heated debates between proponents of the L2 norm and supporters of teacher CS practice. Chowdhury (2013) asserts that from the emergence of Grammar Translation Method to Communicative Language Teaching various teaching methods have witnessed two opposing sides regarding the use of L1 in the classroom. Proponents of TL only, prescribe an inflexible diet of the FL for students of a foreign language. They believe that use of the native language in the FL classroom can be counter-productive, dissuade learners from acquiring the target language (Cook, 2001; McMillan & Rivers, 2011). Those who believe L1 should not be used in the FL classrooms agree with Krashen's (1982) “comprehensible input” which asserts that language which is slightly beyond the student’s current level of competence should be used i.e. we should go for meaning.

According to Richards and Schmidt (2010), comprehensible input hypothesis is “spoken language that can be understood by the listener even though some structures and vocabulary may not be known. According to Krashen’s theory of language acquisition (1982), Comprehensible Input is a necessary condition for second language acquisition” (p. 108). Central to Krashen’s comprehensible input hypothesis is the belief that maximum exposure to the target language structures is critical in ensuring successful language acquisition.  This belief implies the omission of the L1 in the language classroom as it was believed that using L1 could hinder the development of proficiency in the second language (Then & Ting, 2011). Also, it is believed that exposure to the appropriate sample of language is conducive to learning and development (Skiba, 1997). In the same vein, Cheng (2013) suggested in a quantitative research that maximum exposure of the TL was essential.

Moreover, Ellis (2008) asserts that teachers' use of CS is still a sophisticated issue and raises a lot of challenges. Further, its use depends on the instructional context and also concerned with the interactionist viewpoint, in which learners must be exposed to maximum TL. As Cook (2001) states, strong and weak versions of this monolingual rule, being “taken for granted as the foundation of language teaching” (Cook, 2001, p. 404), influenced later language teaching approaches and methodologies like the audio-lingual method and task-based learning. This trend is a reflection of the doctrine that Macaro (2001) calls a maximal position, in which the L1 is necessarily an evil rather than a pedagogical resource. In addition, Sert (2005) argues that when a teacher makes a habit of clarifying instruction in L1 after introducing it in L2, there is a chance that learners, assured of receiving the L1, ignore the L2 instruction.

       Contrary to the belief that L1 use is destructive in the foreign language classroom, many researchers are now arguing that L1 can be beneficial as a cognitive tool that aids in TL learning (Dailey-O’Cain & Liebscher, 2009). These claims have encouraged them to argue in favor of some sort of principled switching between first and second language use in the foreign language classroom. However, according to Skiba (1997), CS is not a language interference as it supplements and facilitates speech. For instance, where it is used due to lack of language knowledge, CS ensures flow of speech rather than presenting an interference in language. In this regard CS is more of a supporting toolbox in exchanging information and in social interaction. Likewise, Sert (2005) concludes that using CS bridges the gap between known and unknown and is an important tool in teaching language if used efficiently.

A number of researches have indicated that CS does not necessarily show the lack of competence but it can be considered as nobility for communication (e.g., Hansen 2003; Shin 2005). Although it might interrupt the flow of the conversation, it still provides opportunity for students to learn (Skiba, 1997). Some researchers (e.g., Cook, 2001; Macaro, 2001) support principled and judicious use of the L1 as a sign of responsive and effective teaching for which educators should not be embarrassed. Cook (2001) has questioned the norm of monolingual education in FL classrooms and has recommended that teachers had better aim to “create bilinguals” instead. Also it is argued that if the ultimate goal of language instruction is to create bilinguals, then the aim of incorporating systematic CS behavior into the classroom is both worthy and appropriate.

In a study by Ahmad and Jusoff (2009) on students of low English proficiency, it was shown that students support CS mainly for understanding teachers’ instructions, learning vocabularies and improving their English generally. Besides these various functions, CS provided students with affective supports and learning success. There existed a significant relationship between teachers’ CS and low proficiency learners’ positive affective states as Ahmad and Jusoff (2009) argued, “the more teachers code-switch, the stronger is the learners’ affective state” (p. 51). This is based on the responses of the participants who indicated that CS by teachers enabled them to feel more comfortable and less anxious during lessons as they were able to comprehend the L2 input. In addition, most participants have attributed their language learning success to the use of CS by teachers.

Concerning affective state, Yao (2011) believes that teachers can enliven their class using various methods, one of which is shifting codes to tell a joke. In the same study when it comes to students’ perception of teachers’ CS, she asserts that students mainly support teachers’ using L1 because of several reasons: they desire that teachers make use of L1 to explain the cultural topics, grammatical and lexical items, and points concerning the lesson content. Also students want their teachers to use L1 to clarify what they want students to do and to engage their attention. By using CS on the part of the teacher, students are more encouraged get the better feedback. Telling a joke in L1 enlivens the class and the learning atmosphere became more interactive. Further, a study by Jingxia (2010) on Chinese universities suggests that teachers and students are nearly of the same attitudes toward using CS to Chinese in the classroom. The teachers (80%) and students (66%) testify the positive attitude which both of them share.

In a study by Nordin et al.(2013) on students at tertiary level,the functions of the CS favored by the students at 40% and above were as follows: giving instruction, giving feedback, checking comprehension, explaining new words, explaining grammar, helping students feel more confident and comfortable, explaining differences between first and second languages, discussing assignment, test, and quizzes. In this study the students maintained that using CS for explaining new words and helping students feel more confident was more important. In addition, Alenezi (2010) suggests that even though most of the students agree that using one language is useful for their learning, they support CS for some main reasons: it makes it easy to understand, it dismisses any confusion. Along with this study, Then and Ting (2009) also conclude that on the condition that the proficiency level of the students in the target language is not sufficient, CS is a necessary tool for teachers to facilitate students’ comprehension of the content area.

PURPOSE OF THE STUDY

The present study attempts to make a contribution to the views on CS, in general, and different functions of using CS in foreign language education, in particular. More specifically, a primary goal is to make recommendations for teachers of foreign languages, based on the results of the analyses of personal attitudes of students across various proficiency levels toward teachers’ CS. Much less attention has been systematically paid to the study of teachers’ CS and their students’ attitudes across various proficiency levels toward this phenomenon in foreign language classrooms. A considerable number of researchers (e.g., Alenezi, 2010; Evans, 2009; Rezvani & Rasekh, 2011; Yao, 2011) have investigated the students’ perception of CS but, to the knowledge of the researchers, none of them has been conducted across various proficiency levels of English. This study will investigate and show students’ attitudes to CS to find out how learners view switching between Farsi and English in EFL classrooms and when and why teachers and learners switch codes. Thus, this study is an attempt to investigate whether learners’ code-switching perceptions differ across various proficiency levels of English.

METHOD

Participants                                                                                                                                            A total number of 105 students were approached to participate in this study. These participants, selected based on the availability criterion, were all from two branches of the same English language institute, University of Tehran Language Center (UTLC) located in Tehran and Karaj, Iran. The participants were adult male and female students whose ages ranged from 20 to 50. Their genders have been taken into account as a variable.

Instrumentation

For the purpose of this study, a closed 35-item Persian questionnaire was administered to student participants with a standardized five-point Likert scale for all of the items (see the appendix). A CS Perception Questionnaire (see the appendix) was developed by the researchers. The items in the questionnaire, specifically the functions of CS were adopted from the functions of CS based on current views of CS in target language classrooms found in the literature. The items for each section in the CS questionnaire were then grouped into narrower categories (as Part 1: Teacher’ persona (9 items), Part 2: Subject access (11 items), Part 3: Classroom management (8 items), and Part 4: CS for interpersonal relations (7 items)) to identify and explore the relevant correlations and discrepancies in the data.

Data Collection Procedure

Upon obtaining approval to conduct this research from the supervisors of UTLC, the research was conducted at the two selected branches. In order to increase the internal validity of the study, all the participants, i.e., teachers and students, were advised to feel free to take part. To accomplish this objective, the researchers went to each class personally and required the teachers to spare five minutes of their class time to this study.  After receiving the consent, the purpose for conducting this research was explained to the students. A consent form including short demographic information such as name and e-mail address was distributed among those students who showed their willingness to participate in the study. Their e-mail addresses were used to send them the questionnaire for collecting the necessary data. The completed questionnaires were to be e-mailed back to the researchers later.

Data Analysis

Since the data did not meet the assumptions of the parametric tests, especially the assumption about normally distributed data, non-parametric data analysis tests were used. Moreover, the data obtained were ordinal. Therefore, non-parametric data analysis measures had to be employed. To answer the research question, the Kruskal-Wallis test (a non-parametric alternative to parametric one-way analysis of variance) was used which is a rank-based nonparametric test that can be used to determine whether any statistically significant differences are present between two or more groups of an independent variable on a continuous or ordinal dependent variable. The scores are converted into ranks and mean ranks are compared for every group (Pallant, 2013). In the case of this study, EL, IN and UI are the three groups of the independent variable.

RESULTS

The results of the Kruskal-Wallis H test showed that there was a statistically significant difference in students’ perception of this phenomenon across different proficiency levels (elementary, intermediate and upper-intermediate). Table 1 shows that the Elementary group recorded higher median score (Md=98) than the other two proficiency levels which recorded median values of 76 for Intermediate and 80 for Upper-intermediate levels.

Table 1:  Comparison of views towards CS across proficiency levels

Furthermore, as shown in Table 2, for three subparts of CS (Subject access, Class management and Interpersonal relationship), the EL students had the highest median compared to the other two groups of IN and UI.

Table 2: The Frequencies of Medians of CS Parameters across the Three Proficiency Levels

As to teacher persona, which refers to the personality of the teacher, the three groups’ medians were 22, 14, and 16 for EL, IN and UI students respectively. Also, for subject access, which means whether the use of L1 facilitates learning and teaching particular language skills, the medians for the three groups were 24, 10, and 16, respectively. In the case of class management, i.e., the degree to which the students preferred CS for class management, the medians were 24, 13, and 13 for the three groups of students respectively. Finally, regarding interpersonal relationship, which is the relationship developed between the teacher and the students, the medians are 23, 9, and 18.  

As indicated in Table 3, there were significant differences among the three groups of students with respect to three of the parameters, i.e., 'subject access', 'class management', and 'interpersonal relationship' as the p values turned out to be far less than 0.05. However, there was no significant difference among these three groups of students regarding 'teacher persona' (p = .259).

In order to know about the level of the students' attitudes towards CS, the responses made by the three groups of EL, IN, and UI students, to each item were analyzed. To do so, the percentage of each option, i.e., 'strongly disagree', 'disagree', 'not sure', 'agree', and 'strongly agree', to each item of the questionnaire was calculated. This would help to come to an understanding of possible differences among the students' perception of CS across various proficiency levels. Table 4 indicates all the percentages. Further explications are given below the table.

Table 4: Students’ attitudes toward CS across various proficiency levels

EL=Elementary, IN=Intermediate, and UI= Upper-intermediate

As indicated in Table 4, the upper-intermediate (UI) group had the highest strong disagreement percentage with Item 15 of the questionnaire, i.e., sing Farsi to start new topic by teacher, with a percentage of 88.9, while, the elementary (EL) and the intermediate (In) groups showed a strong disagreement to this Item with percentages of 52.8 and 60 respectively. Compared to the other level groups, the IN group strongly disagreed with Item 29, which is 'The use of Farsi by the teacher encourages students’, participation in classroom activities' with a percentage of 74.3. Regarding the same Item, the EL and the UI levels expressed 33.3% and 63% strong disagreement respectively. The EL group strongly disagreed with Item 4, which is 'Teacher's CS demolishes L2', and the relevant percentage was 69.4 while the IN and UI groups strongly disagreed with this Item with percentages of 42.9% and 40.7% respectively. On the other hand, the UI group did not at all strongly disagree with Item 7 which dealt with their preference of the fact that 'the teacher should minimize the use of the students' first language during lesson'. With respect to the same Item, the El group showed 11.1% strong disagreement and the IN one only 2.9% strong disagreement. The same held true about Item 9, 'using CS when students are not capable of expressing themselves in English'. The IN group showed no strong disagreement, while the EL group declared 5.6% and the UI one 7.4 % strong disagreements. Finally, as shown in Table 5, the average of the strong disagreement percentages of the EL group concerning all the Items in the questionnaire turned out to be 25.85%. Those for the IN and UI groups were 38.94% and 38.61% respectively.

       With respect to the percentages of those students who disagreed with the Items of the questionnaire, it should be stated that the IN group expressed the highest disagreement to Item 31, which said 'Using Farsi to build/strengthen interpersonal relationships between the teacher and students' (51.4%). While the EL group showed 33.3% disagreement and the UI only 22.2% disagreement. It was also found that 44.4% of the students in the UI level disagreed with Item 12, i.e., 'Teacher uses Farsi to explain new words'. As to the same Item, the EL and the IN groups expressed 25% and 20% disagreement respectively. Moreover, another highest disagreement percentage leveled against any of the Items by the level groups was 38.9% for Item 32 of the questionnaire, 'Using Farsi to receive feedback from teacher' by the El group, while the other two level groups, i.e., IN and UI, showed 34.3% and only 11.1% disagreement to this Item. The lowest percentage of disagreement made by the EL group was 2.8% which was expressed in the case of Item 9, 'using CS when students are not capable of expressing themselves in English', the lowest percentage for the IN group was 2.9% for Item 7, 'prefer the teacher to minimize the use of my first language during lesson', and the lowest one for the UI group was 0% declared on Item 15, 'Using Farsi to start new topic by teacher'. Table 5 indicates that average disagreement percentage of the EL group concerning all the Items in the questionnaire turned out to be 21.34%. That for the IN group was 25.89% and the one for the UI group was 22.71%.

       To the other end of the continuum of the Likert scale options used in the questionnaire, the UI group showed the highest agreement among all percentages. This happened in the case of Item 20 of the questionnaire, 'I use CS when explaining difficult words and sentences to my peers', with a 40.8% agreement. As to the same Item, the EL and the IN level groups expressed 27.8% and 31.4% agreements respectively. The IN group showed no agreement to Item 25, 'Teachers who use CS can better evoke responses and reactions from students', while the EL group showed a low agreement of 13.9% to this Item and the UI group only 11.1%. Moreover, the UI group expressed no agreement to Items 11, 'Teacher uses Farsi to explain meaning of sentences and translation', 23, 'Teachers who use CS can better direct students', and 32, 'Using Farsi to receive feedback from teacher'. These three Items also were very poorly welcomed by the other two groups. The percentages for the EL and IN groups for Item 11 were 11.1% and 5.7% respectively. To Item 23, the EL group's agreement was only 2.8% and the IN group's agreement was 8.6%. Finally, 8.3% of the responses made by the EL group supported the agreement with Item 32 and just 2.9% of those made by the IN group was in favor of this Item. The results indicated in Table 5 reveal that the average percentage of the El group's agreement with all the Items of the questionnaire was 13.26%, that for the IN group was 11.43%, and for the UI group it was 11.30%.

Regarding strong agreement with the Items, Item 7, 'prefer the teacher to minimize the use of my first language during lesson', turned out to have the highest percentages for all the three groups. The strong agreement percentage for the El group was 50%, for the IN group 71.4%, and for the UI group 63%. The EL group showed a 0% strong agreement to Item 3, 'Teachers who use CS are deficient in English'. The IN group expressed a low percentage of 8.6% strong agreement to this Item and the UI group indicated just 3.7% strong agreement to the Item. The IN group did not at all strongly agree with Items 5, 'CS should be included as an integral part of the FL lesson', 11, 'Teacher uses Farsi to explain meaning of sentences and translation', 12, 'Teacher uses Farsi to explain new words', 16, 'Teachers who use CS can better clarify the lesson substance taught', 18, 'CS will facilitate the language learning process', 23, Teachers who use CS can better direct students', 24, 'Teachers who use CS can better engage students’ consideration and attention', 26, 'It is more effective and time-saving for the teacher in explaining what s/he is teaching', 27, 'It allows me to focus better on the content of learning and to reduce distraction', 29, 'The use of Farsi by the teacher encourages students’ participation in classroom activities', 30, 'The use of Farsi by the teacher makes me feel more confident and motivated in learning English', 32, 'Using Farsi to receive feedback from teacher', and 33, 'Teacher uses Farsi to express empathy or solidarity toward student'. The UI group also expressed 0% strong agreement to Items 6, 'The use of my first language by the teacher helps me to enjoy the lesson', 15, 'Using Farsi to start new topic by teacher', 23, 'Teachers who use CS can better direct students', 29, 'The use of Farsi by the teacher encourages students’ participation in classroom activities', and 32, 'Using Farsi to receive feedback from teacher'. It is interesting that the percentages of other level groups to almost all these Items were also very low. Table 5 shows that the average percentage of the El group's strong agreement with all the Items of the questionnaire was 21.12%, that for the IN group was 8.82%, and for the UI group it turned out to be 12.27%.

Table 5: The average of the percentages of the students' attitudes towards CS across various proficiency levels

DISCUSSION

The research question in this study sought to explore the perceptions and beliefs of selected students about their use of CS in Iranian EFL classrooms in two branches of a language center. The findings were analyzed for three proficiency levels of EL, IN and UI students regarding their perceptions and beliefs toward code-switching.

Based on the students’ responses to the questionnaire, it was found in the current study that EL students had positive opinions toward the use of CS in general, and teacher CS in particular which is in agreement with Ahmad and Jusoff's (2009) findings which showed learners support CS in ELT classrooms. The finding for the EL students suggested that students agreed with their teachers' use of Farsi while teaching English and about half of them considered teachers who used CS were quite proficient in English. Another finding was that most of the students believed that Teacher's CS did not affect the languages. Surprisingly, it was found that half of the students agreed and the other half disagreed with CS to be included as an integral part of the ESL lesson.  So, in this case both languages are preferred. This can be interpreted as EL students showed more positive attitudes towards the teachers who employed Farsi in their English language teaching.

In keeping with Yao (2011), this study found that CS can be a useful tool for pupils’ language learning when it comes to understanding grammatical rules better. Furthermore, it is also indicated that it can be useful to explain instructions in pupils’ L1, when instructions in the TL are too complex but this function of CS should be practiced with too much care as two third of the students disagreed with teachers’ explaining instructions in pupils’ L1. This result further supports the idea of Wright (2010) who believes the reason for this is that as the pupils are receiving instructions in their L1, they do not need to attend to the TL and thereby, they learn less English.

Elementary students reported to have positive view toward CS as a useful technique when they need to gain better comprehension mainly while providing students’ comprehension improvement as well as giving classroom procedures. This was in line with several researchers’ findings (e.g., Ahmed & Jusoff, 2009; Ellis & Shintani, 2013; Hamidi & Najafi Sarem, 2012; Macaro, 2005). It can therefore be assumed that the teachers should not take it as a burden on their shoulders not to use L1 in EFL classes when dealing with Elementary students.

Also, the results support Ahmed and Jusoff (2009) that CS helps facilitate the flow of classroom instruction since the teachers do not have to spend so much time trying to explain to the learners or searching for the simplest words to clarify any point. This in turn prevents from confusion that might arise. The most obvious finding to emerge from the analyses is that proficiency (perceived competence) in a foreign language had an effect on students’ perceptions of and attitudes toward using CS both by teachers and learners. Furthermore, it also had an impact on students practicing CS and the prevalence with which they used in it. Students who perceived that their competence in the foreign language was high had less favorable perceptions of teachers and students’ CS and also had less favorable attitudes toward CS by both intermediate and Upper-intermediate groups than Elementary students who considered their competence low in the language. In addition, the high perceived language competence group was also less likely to engage in CS and did so less frequently which match those observed in earlier studies (e.g., Ariffin & Husin, 2011; Bailey, 2011; Then & Ting, 2009; Yao, 2011).

Another important finding with respect to the proficiency was that virtually no differences were found between the Intermediate and Upper-intermediate students' perception of CS to the large extent which shows when learners reach a certain level of proficiency, i.e. Intermediate, they do not tend to support CS either by teacher or classmates. CS was therefore considered as an inevitable and conscious choice when the students did not understand the message when working with Elementary students that the teacher wanted to convey during class which supports the findings of Agneta and Ana (2010), but this function for Intermediate and Upper-intermediate levels is not required and they prefer anything to be explained in English.

In addition, as with previous studies, this study revealed that low proficiency students’ benefit mainly from CS utilized as tool, both when used by the teacher, as well as by the students themselves. On the other hand, High proficiency students seem to prefer an English only classroom for the most part. So, it is clear that as the proficiency level of students goes higher, they need less CS. It is not easy to adapt the lessons in a way that meet every learner’s individual need as high proficiency learners benefit more from lessons held in only the TL, while low proficiency learners benefit from lessons when the L1 is used as a tool in the language classroom. This may lead to teachers’ obligation to use learners’ L1 because most of the time there are low proficiency learners in risk of failing the course. However, teachers should try to inspire students to rephrase or paraphrase the language used in the TL, as a strategy to motivate them to use English, if they practice CS owing to the fact that they are not understandable. Also, according to these data, it can be inferred that maximum exposure to the TL is crucial. This finding supports the Interactionist Viewpoint, in which leaners' exposure to maximum of TL is emphasized.

CONCLUSION AND IMPLICATIONS

In this study, the aim was to investigate the possible differences between students across various proficiency levels’ perception of CS in EFL classroom. The findings suggest that, in general, it may be illogical to expect the exclusive use of the TL in the FL classroom because teachers are expected to use every possible tool within their disposal to fulfill their duty of educating their students and to ensure the smoothness of classroom interaction. In this regard, CS can be considered as a precious asset for bilingual teachers in foreign language classrooms especially when it comes to reducing the stress of learners and enhancing teacher-student relationship. But finally, this is the teacher who decides whether L1 can be used in the class or not. Moreover, although most of the teachers agree that CS has some applications in the classroom, it is highly recommended that they receive enough training about its usages and the advantages that they can gain in their teaching foreign language.

At least the results from the current study imply that teachers’ CS in the classroom is not harmful to the learning process. This implies that it is possibly less logical to criticize teacher code-switching. These findings may help us understand that despite the number of objections against using CS, if judiciously used, it can be a precious tool within teachers and students’ access. As reported earlier CS can be employed as a tool for enhancing students' learning. Therefore teachers can benefit from the findings of the current study.

The study dealt with three levels of proficiency, i.e., El, IN, and UI. This can be as its delimitation. Further studies can be carried out by taking care of other proficiency levels. Moreover, the majority of the learners were adult students. Other research studies can be conducted with other age groups.