PAIR-DISCUSSION TASKS’ EFFECTS ON STUDENTS’ ORAL PERFORMANCE

Communicative skills have gained importance in a wide range of general education courses and students are aware that they need effective speaking skills that will help them succeed in academic and social contexts and will provide better job opportunities. To be able to have a successful academic career or to achieve their dream job they will have to take several oral exams and interviews mostly done in pair-discussion formats. Thus, the purpose of the present study is to investigate the effects of pair-discussion tasks on students’ overall speaking exam scores and their interaction scores. The data of the study was gathered from 40 prep-school students studying at Namik Kemal University/ Turkey. 20 of the participants were assigned to experimental group and 20 of them were assigned to control group. The experimental group received 10 pair-discussion task sessions together with their usual class instruction whereas the control group didn’t receive any special treatment. Total speaking exam scores and interaction scores of participants of both groups were gathered to be analyzed using independent samples t-test. The results of the study indicated that pair discussion tasks have a positive effect on students’ speaking exam scores and their interaction scores. Participants in experimental group benefited from the treatment and improved their speaking abilities and peer interaction.


INTRODUCTION
The world is changing rapidly and now we are living in an era with technological inventions and developments. These improvements in technology have made communication much easier than it used to be and gave us the chance to communicate well beyond our speech communities (Ellis, 2003). To be able to keep up with the changes of this fast developing world, learning a second language has become indispensable. This need has been recognized by education authorities all over the world and students learn foreign languages starting from primary school. According to Nunan (1999) language is the feature that makes human species unique and an educated individual should have a good command of at least one other language apart from his or her first language. Within language learning, speaking skill has been recently considered as a priority in language teaching by many methodologists. In other words, knowing a language may be compared to the ability to speak that language well (Nunan, 1999). We gain this skill in our native tongue at a very early age that we forgot how we struggled trying to say just a few words. Thornbury (2009) stated that even though we are not aware, we use tens thousands of words just a day, and this number goes up when it comes to auctioneers or politicians (p. 1). That's why; speaking is a critical skill in language learning. Learners need to develop this skill so that they can be a part of this global world. As Troike (2008) mentions, tourists ask directions and seek information about hotels and entertainment, immigrants shop for goods, seek services, and describe health problems; foreign students negotiate for housing, utilities and currency exchange as well as to express themselves in academic or professional contexts (p. 161). Consequently, the need for communicative skills in English has been increasing both for business and in private life.

Communicative Language Teaching
Within the last quarter century, communicative language teaching (CLT) took interest of many researchers and methodologists as the new and innovative way to teach English as a second or foreign language. Teaching materials, course descriptions, and curriculum guidelines have been prepared based on its principles. One of the principles of CLT is that it puts the focus on the learner. CLT as all learner-centered methods is concerned with learner's needs, wants and situation (Kumaravadivelu,2006). Kumaravadivelu (2006) defines learner-centered methods as such: These methods seek to provide opportunities for learners to practice preselected, presequenced linguistic structures and communicative functions through meaning-focused activities, assuming that a preoccupation with form and function will ultimately lead to target language mastery and that the learners can make use of both formal and functional repertoires to fulfill their communicative needs outside the class. (p. 91) For CLT, language is a system for the expression of meaning, and primary function of language is interaction and communication. According to Littlewood (1994) learners need to be taught how to use the language spontaneously and flexibly in order to express their meaning. For this reason, they should be provided situations where they must use language as an instrument to carry out meaningful tasks and objectives that reflect the need of the learner. The focus should be on functional effectiveness rather than structural accuracy. Thus CLT benefits from real-life situations that necessitate communication (Nunan, 2004).

Task Based Language Teaching
CLT has also been argued a lot and these arguments have led to the rise of task based teaching. TBLT emerged as an effective language teaching methodology to develop a purpose-driven communicative language learning built around the use of real-world tasks 507 (Kasap, 2005). The focus in task based teaching is communicative and meaningful tasks and the process of using language communicatively rather than just uttering correct sentences. For this reason, TBLT can be considered as one model of Communicative Language Teaching (CLT) in terms of regarding real and meaningful communication as the primary feature of language learning (Kumaravadivelu, 2006;Ellis, 2003). Today, many language teachers and researchers consider TBLT as a fruitful and beneficial approach to L2 instruction that has the potentiality to provide them with new insights into teaching language in the most effective ways. Moreover, Task Based Instruction is considered to be an effective approach that promotes a learning environment in which learners can choose and use the target language forms which likely to achieve to accomplish their objectives. (Hedge, 2000) Task-based instruction (TBI) is considered as an alternative method to traditional language teaching methods because it favors a methodology in which functional communicative language use become prominent (Ellis, 2003). In 2005 Nunan summarized the difference between the traditional classroom and the TBLT classroom (Table 1).  (Nunan, 2005, p.49) TBLT took the responsibility from the teacher and share it out between the teacher and learner giving learners the chance to be more active participants in the learning process. Nunan (2004) also stated that Task-Based language teaching has strengthened some beneficial and effective principles and practices. First of all, when choosing the content, students' needs are taken into consideration. The emphasis on learning to communicate through interaction in the target language has strengthened. Bringing authentic texts into the learning situations gained importance. Moreover, learners have the opportunity to focus not only on language but also on the learning process itself. Learners' own personal experiences have enhanced as important contributing elements to classroom learning. Finally, classroom language learning is linked with language use outside the classroom 2.2.1. What is Task Task Based Language Teaching is built around tasks. From the day TBLT was proposed, many researchers and methodologists have made different definitions of task based on previous studies and their own experiences. The table 2 summarizes most cited ones. Table 2.

Researchers
Key concepts Long (1985) What people do in their everyday life, at work, at play, and in between?

508
Breen (1987) Carroll (1993) A range of work for exercise and activities in language instruction. Any activity in which a person engages to achieve specific objectives. Skehan (1996) Meaning, task completion, the real-world and outcome are focused. Willis (1996) Bygate et all. (2001) A classroom undertaking for a communicative purpose to achieve an outcome. An activity which requires learners to use language with emphasis on meaning, to attain an objective. Ellis (2003) A work plan that requires learners to process language pragmatically to achieve an outcome. Nunan (2005) A piece of classroom work to convey meaning rather than manipulate form. (Izadpana, 2010;Branden, 2006) When students perform a task, they are not expected to use a newly learned item to practice it. On the contrary, students are supposed to make use of their all language knowledge to complete a task as in real life. That's why, speaking tasks provide the chance to rehearse the language that the students need to use outside the class and prepare learners for real-life language use (Harmer, 2003;Thornbury,2009;Pinter, 2007). Communicative tasks also encourage the automatization of language knowledge. Another advantage of tasks is that they can be used as a feedback tool both for the teacher and the students. When teachers observe their students during the performance of tasks, they can identify the problems and also observe their progress. Moreover, good speaking tasks engage and motivate students by encouraging them to communicate with each other in real time (Harmer, 2003). According to Thornbury (2009) criteria should be designed for speaking tasks so that students can benefit from tasks most. Table 3. Criteria for speaking tasks Productivity: a speaking activity needs to be productive in order to provide the best conditions for autonomous language use. Purposefulness: a speaking activity should have a clear outcome, a common purpose.

Interactivity:
activities should require learners to take into account the effect they are having on their audience. Challenge: students should be forced to draw on their available communicative resources to achieve the outcome. Safety: students need to feel confident that they can complete the activity without to much risk.

Authenticity:
speaking tasks should have some relation to real-life language use. (91) Communicative activities and tasks help learners to activate and integrate forms for meaningful communication. Some methodologists took interest in task types that can be applied in a language classroom. Pattison (1987) proposed seven tasks and activity types: (Pattison, 1987; as cited in Nunan, 2004, p. 57) 1.
Questions and answers 2.
Pictures and picture stories 6.
Puzzles and problems 7.

Discussion Tasks
Discussion task is one of the many useful tasks that TBLT proposed. Discussion tasks are advantageous since learners aren't supposed to use a particular structure, but they are able to use any linguistic resource they can attain to complete the task (Nunan,2004). In addition, in order to succeed in universally proven and accepted tests such as IELTS or TOEFL, students must be able to take part in a discussion successfully (Stroud, 2015). For this study, discussion task type was chosen since the oral exam is administered in the form of discussion tasks, and also discussion tasks have several advantages. Discussion activities;  give students the chance to express their own ideas, feelings and preferences, and opinions. Personalized practice makes language more memorable and it is motivating as learners see the ways in which they can make use of language resources in interpersonal situations. (Harmer, 2003)  can provide opportunists for developing certain aspects of fluency.  involve students in talking about a range of topics which engage their interests, opinions, histories, and experiences.  encourage students to use the language needed to sustain conversation over a period of time by drawing other speakers.  provide opportunities for students to practice the strategies required in interpersonal communication. (Hedge, 2000, p. 277) For a successful discussion activity, one of the important things is that we should give students time to assemble their thoughts before starting the discussion activity and teachers should make students engaged with the topic (Harmer, 2003). One of the best ways of encouraging discussion is to provide activities that force students to reach a decision or a consensus. (Harmer, 2003, p.273)

2.3.Pair Work
For the present study students performed discussion tasks in pair because of the benefits of pair work. Pair work;  increases the amount of speaking time for each student including the opportunity to provide and obtain feedback from other students.  allows students to work and interact independently without necessary guidance of the teacher, thus promoting learner independence.  allows teachers time to work with one or two pairs while the other students continue working.  promotes cooperation and helps the classroom to become a more relaxed and friendly place.  is relatively quick and easy to organize. (Harmer, 2003, p.116)  students do not feel as "exposed" as if speaking in front of the full class, so shy students feel comfortable.  most students prefer to share their ideas with closer classmates than with the whole class or the teacher. This may help to increase students' motivation. (Harmer, 2003, p.116;Taylor & Wigglesworth, 2009, p.326). Student-student pair work can take two forms: open or closed. Open pair work is when two students perform a dialogue while the rest of the class observes. In closed pair work, adjacent students perform the dialogue while all pairs work at the same time. (Thornbury, 2009, p.73) Participants of this study performed the task as closed pair work so that they get the chance to work at the same time.

Interaction
To convey messages and to negotiate meaning students need to interact with each other which makes interaction an indispensable part of language teaching. Learners need skills to interact with the interlocutor and the other learners by initiating and responding appropriately without long pauses and hesitations to successfully complete the task. Learners should be able to maintain the coherence of the discussion and should be able to ask for clarification (Thornbury,2009, p. 129). Learners benefit from this negotiated interaction in many ways. Negotiated interaction helps learners to improve their language ability, to progress faster, to adjust their speech, to notice the gap between their present language and what was expected in a particular situation (Kumaravadivelu, 2006). Long (1985) proposed interaction hypothesis which states that language development is promoted by face-to-face interaction and communication. Long suggests that the best input for language acquisition is that which arises when students have the opportunity to negotiate meaning in exchanges where an initial communication problem has occurred. The interaction hypothesis may be summarized thus: 1) Interactional modification makes input comprehensible. 2) Comprehensible input promotes acquisition. Therefore, 3) interactional modification promotes acquisition. According to Long, interaction facilitates acquisition because of the conversational and linguistic modifications occurring in such discourse (Long, 1985; as cited in Tulung, 2008, p.108). Oral discussion tasks require students to read and understand the task, listen to each other carefully, ask for clarifications when necessary without having any breakdowns in communication thus they measure listening, reading and interaction (Ockey, Koyama, Setoguchi& Sun, 2015) PREVIOUS STUDIES Communicative Language Teaching and Task Based Language Teaching have been a fruitful research area for many years for EFL and ESL researchers. In recent years, researchers conducted studies related to the effectiveness of communicative tasks and activities. Sung (2010) investigated the effect of newly designed three different communicative language teaching-based projects. The results of the study showed that students considered the tasks as fun and interesting. Moreover, the students reported that the projects provided more opportunities to practice the language and also fostered student autonomy and helped them express their individual personality. The projects gave the chance to use their creativity and helped improving their skills in the target language. The participants also suggested that there should be more emphasis on individual learner difference. Incecay and Incecay (2009) conducted another study to examine the perceptions of Turkish EFL university students to understand the appropriateness and effectiveness of communicative and non-communicative activities. The study results revealed that participants showed positive attitudes towards both communicative and non-communicative activities which means that students believed the effectiveness of both traditional and communicative activities. Kasap (2005) explored the effectiveness of task-based instruction in the improvement of learners' speaking abilities as well as student and teacher perceptions of TBI at Anadolu University School of Foreign Languages.
Control and experimental groups were administered questionnaires, interviews and oral tests. The T-tests revealed no significant differences in any of the comparisons. The study demonstrated, however that students' general perceptions of task based instruction were positive, and the interview with the study teacher also yielded a positive result. The questionnaire results demonstrated that students had neutral or partially positive reactions to the treatment tasks but found these helpful in developing their oral skills.

511
Communicative tasks and their effectiveness has been another study interest. Some researchers investigated the effectiveness of various task types and found different results. Huei-Chun (2007) carried out a study to examine the effectiveness of task type on the performance of EFL speaking tests. Overall, the participants agreed that the oral test should include more tasks and didn't perform better in one of these tasks. Dumont (2017) investigated the effects of speaking tasks on production of fluency features of Frenchspeaking learners of English. The results showed that speaking task characteristics influence fluency of the speaker. For example, the monologic picture description task is characterized by a higher frequency of unfilled pauses and a lower speech rate as compared to the set topic and the free discussion task. Some researchers designed studies to examine the effect of a particular task type. The effectiveness of role-play activities on motivation and learners' speaking ability has been investigated. Results of these studies showed that role-play activities are motivating, have a positive effect on students' speaking ability and students became more active in class. Moreover, researchers suggested that communicative tasks such as role-plays should be included in class procedures to motivate students (Liu & Ding, 2009. Aliakbari&Jamalvandi, 2010. Liu, 2010. Taskın, (2009) designed a study to explore how role-play can improve students' English speaking proficiencies and overcome students' shyness while speaking English. Experimental design was used in the study. Both qualitative and quantitative data was gathered. Experimental group were taught role-play in speaking lessons by the instructor, whereas control group were taught the activities in speaking book. The result of the study yielded that experimental group showed positive attitudes towards role-play. After the applications of role-play, students in the experimental group not only improved their speaking proficiency but also overcame their shyness while speaking English. At the end of the study, it was observed that there were no significant differences in control group in terms of their speaking proficiency. Gan, Davison and Hamp-Lyons (2008) focused on a different task type namely discussion task and conducted a study to investigate the effect of topical talk of students in a group negotiation task under an assessment condition. The results of the study yielded that group oral discussion tasks provided learners with opportunities to demonstrate their linguistic competence together with their interactional abilities such as initiating, expanding or closing a topic. Most activities and tasks in CLT and TBLT are performed in the form of pair or group work. Many researchers have been interested in pair work and made valuable contribution to the field. Hamano-Bunce (2011) investigated whether chat room pair work was a comparable alternative to face-to-face spoken interaction. The results of the study yielded that the participants spent much more time to complete chat room tasks than oral tasks whereas the participants performed better and produced more language orally and more turns were observed. Furthermore, more negotiation about the language the learners produced, more selfcorrection the language production was observed in the oral pair work negotiations. Achmad and Yusuf (2014) examined students' pair-work interactions to develop their speaking skills and pair-work interaction task had a major impact on students' language learning achievement and proficiency in the classroom. Mcdonough and Sunitham (2009) carried out another study to examine the collaborative dialogues between participants when they were working together to carry out self-access computer activities. The participants considered collaborative activities as helpful and motivating. There are also some studies which investigated student-student interaction under assessment conditions. Brooks (2009) and examined the effectiveness of two interactional test types 512 namely student-examiner interaction and student-student interaction on students' oral performance. The study results revealed that overall students gained higher scores from pair interaction test and there was not a significant difference between test scores on the paired and individual formats with regard to the order or testing. Another result of the study showed that the paired format testing elicited a wider range of features of interaction than did the individual format and seeking confirmation was found to be the most frequent interaction feature. During student-examiner interaction, what was common is that the tester initiated most of the questions and students replied. On the other hand, pairs showed a greater range of interaction such as elaboration, finishing each other's' sentences, paraphrasing their utterances etc. which resulted in negotiating of meaning and consequently a better performance. To sum up, the participants scored higher and performed better in pairs. May (2009) also examined reactions of four trained raters to paired candidate. The results of the study showed that paired speaking test had the potential to elicit features of interactional competence, including a range of conversation management skills better than traditional testing interviews.

Study Context
The study participants were all prep-school students. They receive 28 hours of English instruction per week. For the testing procedures, students attend 6 midterms once in five weeks and a final exam at the end of the year. Midterm exams consist of listening, use of English, reading/ vocabulary and writing parts with a separate oral exam. Each part constitutes 20 points of total score. For the speaking exam students are provided approximately 15 topics two days before the speaking exam to get prepared. In the first part of the exam students pick a number, think about the topic that the number refers to and they are given 3 minutes to talk about their topic. In the second part of the exam students are given a situation card that isn't provided beforehand. They get prepared for the situation in one minute and have 4-5 minutes to discuss it with their partners. One example situation is that "You are attending a conference in Antalya, what should you bring with you?". When students complete their discussion, they leave the classroom and two different raters score their performance individually by using the school's oral assessment rubric. When they decided on the grade the raters compare their grades and reach a final grade. Although pair-discussion tasks are used as a testing instrument and class materials provide a variety of oral communication tasks, pair-discussion activities are limited and also neglected by the teachers. The study aim is to integrate this task type into classroom procedures hoping that it will improve their speaking skills and higher oral exam scores.

Statement of the Problem
Even though communicative language teaching has its impact on language teaching for years, Turkish students still experience problems in terms of speaking in class and outside the class. As a result, they got low grades from the speaking exam. These low grades demotivate them and caused them to feel anxious, inadequate and they become unwilling to take risks. Their unwillingness to communicate in class hinders their learning. Another problem is that when learners complete a task, they prepare their answers in advance and memorize it. They don't listen to their partners and give answers accordingly but they just wait for their turn and utter their memorized answers. Thus, real communication doesn't occur. This arises from the lack of activities that encourage students to listen to each other. Therefore, hypothesis of the study is pair discussion type of task which was prepared similar to testing procedures will increase pair-interaction and speaking exam results. The research questions that were posed to guide the study are: 513  Do pair-discussion tasks have an effect on students' oral exam scores?  Do pair-discussion tasks have an effect on students' pair interaction? 4.3. Participants 40 preparatory school students fromNamık Kemal University in Tekirdağ took part in the study. The participant's age range was 18 to 20. Their English backgrounds were similar since they were all exposed to same education that was determined by the Minister of Education, in Turkey. All students were Turkish EFL students. Their proficiency level was pre-intermediate assessed by the school exam. Students are placed in their classes according to their midterm results. For the present study 40 students who received 10-15 from their speaking exam was chosen. 20 students were assigned to control group who did not receive any discussion tasks in class whereas 20 students were assigned to experimental group who received 10sessions of pair discussion tasks in a 5-week period. Experimental group's instructor was the researcher herself while the control group'sinstructor was a colleague who agreed to cooperate. Participants were informed about the procedure and the aim of the study.

Materials
Oral exam: The speaking scores students received from two midterms were used as the pre and post-test. The exam scores that students took just before the experiment was used as pre and the exam scores they took right after the treatment was used as post test to see if there was a significant increase in their scores. Oral exam is assessed with a speaking rubric which consists of grammar and vocabulary, fluency and ability to express ideas, pronunciation, and interaction parts. To observe also the change in students' pair-interaction, participants' interaction scores were compared separately. Pair-discussion tasks: For each week's session, the class instructor of experimental group prepared two discussion tasks related to the current unit topic. Students were given a situation and given time to get prepared. Participants performed their tasks first in the form of closed pairs and then in the form of open pairs. Open pair performances were video recorded for the reliability of the study.

Data Analysis
This study was conducted to investigate the effectiveness of pair-discussion tasks on students' speaking abilities and their interaction with their pairs. The scores of the two groups in the first oral exam and the second oral exam were analyzed to see whether there was any support of 10 discussion tasks to the scores of post-test of groups and second oral exam. The data of the study were gathered through oral exam scores. The oral exam part of the midterm was used as pre and post-test. Both experimental and control groups oral exam scores and interaction scores were taken and compared before the treatment to see how homogenous the group was using both descriptive and inferential statistics. After five weeks of discussion task treatment, students received another oral exam. Their scores were gathered and compared using independent-sampled t test to see whether there is a statistically significant difference.

Figure 1. Normal Distribution of pre-test results in total
As it can be seen from the Figure 1, first oral exam scores of both groups had a normal distribution. Both control and experimental group were administered the same oral exam to see whether there is a difference in terms of their speaking ability and level before the treatment. Experimental and control group were similar based on their pre-test results. Moreover, Independent Samples t-test was run to see if there was a statistically significant difference between groups before the treatment. The significance level was ,349. Mean scores, df, t and significance showed that experimental and control group were similar. There was no statistically significant difference between groups. This means that both groups shared similar characteristics in terms of students' speaking ability and level.

Post-test results of experimental and control group
In order to determine whether the two groups' speaking abilities were improved after 10 sessions of pair-discussion tasks, post-test was administered after five weeks. The scores of both groups were gathered to be analyzed. The table 5 shows the results of both groups. According to the results of Independent Sample t-test there was a significant difference between the two groups (00,2). Experimental and control group scores were very close to each other before the treatment. Both groups mean scores increased, but the experimental group received higher scores which mean that pair-discussion task treatment had a positive effect on participants' speaking ability. Participants in the experimental group who received discussion tasks improved their score more than the control group.

Research question 2: Do pair-discussion tasks have an effect on students' pair interaction?
To determine students' interaction ability with their peers, interaction scores of speaking exam were used. According to pre-test results mean score of experimental group's interaction score was 11, 45 and mean score of control group's interaction was 10,95. Mean difference was found as,500. The t-test results show that there wasn't a significant difference in terms of participants' interaction scores before the treatment. Post-Control Group 20 12,25 Table 6 shows that experimental group scored 16,40 on the post-test in terms of interaction, while the control group scored 12,25. Mean difference was calculated as 2,350. Both groups increased their scores between two examinations. However, experimental groups score increased from 11,45 to 16,40 whereas control group's score increased from 10,95 to 12,25. It can be seen from the table that experimental group increased their scores more than the control group. Moreover, t-test result indicates that there is statistically significant difference between two groups when their post-test interaction scores were compared. This result shows us that pair-discussion tasks have a positive effect on students' pair interaction scores. All in all, the experimental group benefited from pair-discussion task treatment both in terms of their overall oral exam scores and their interaction scores.

DISCUSSION
This study investigated whether pair-discussion task type has a positive effect on learners' oral exam scores and their peer interaction. The first research question that guided the study was related to whether pair-discussion tasks have an effect on students' oral exam scores. The study participants consisted of a homogenous group. Participants' pre-test scores were similar and no difference was found. Results of the study yielded that when participants' post-test scores were compared a significant difference was found. Participants in both groups showed an increase in their post-test results. Even though, they didn't receive any special treatment, control group also increased their scores to a certain extent. This may result from the fact that students in the control group received their normal instruction for five weeks before the posttest. Their knowledge of English also increased and this reflected their scores in the exam. However, experimental group increased their scores more than control group. Experimental group benefited from pair-discussion tasks and received higher scores. Another factor that caused experimental group receive higher scores is that the oral exams are administered in the form of pair-discussion tasks. The reason why pair-discussion tasks are used as the testing procedure of the school is that students cannot give memorized answers and need to reach a consensus with their partner. When learners reach a decision or reach a consensus, the tasks became more efficient (Harmer, 2003). Therefore, raters can see the real performance of the students and evaluate accordingly. Additionally, students are expected to give their own ideas about a certain topic so this makes discussion tasks personalized activities which in return result in motivation. A successful task necessitates learners' own ideas and meanings not others together with real life situations that necessitate communication (Skehan 1998;Harmer, 2001;Savignon, 2002;Nunan, 2004). Moreover, discussion topics were chosen carefully to attract students' attention. Students should be given a clear purpose to complete the tasks (Savignon, 2002;Thornbury, 2009). Lots of learning materials also include discussion tasks, however most of them fail to took learners attention. These activities are added just to have some variety, but some of them are prepared beyond learners' level. When teachers expect their students to complete these tasks, students fail and become demotivated. That's why; the study tasks are prepared cautiously. Finally, discussion tasks that participants completed throughout five weeks may helped them to notice the gap between what they think they know and what they actually know (Swain, 1995). They might have worked on harder on their week points which resulted in higher performance. The second research question was posed to investigate the effect of pair-discussion task on students' interaction scores. To answer this question interaction scores that participants get from their midterms were used. The pre-test scores showed no difference but a significant difference was found when post test scores of experimental and control group were compared. Students' interaction with their partners and also with the examiner increased. This may result from the nature of discussion tasks. Most students prepare their answers and give memorized answers during a speaking task and think that they are speaking since they can trick their teachers. Discussion tasks do not allow one sided conversation. A student needs to listen to the partner so that s/he can support her ideas. In addition, to be able to reach a decision, partners need to listen carefully, make eye contact with their partners, prepare their answers accordingly and take turns effectively. Student-student interaction helps acquisition through fostering learner production, feedback, and noticing of form. Peer interaction complements teacher-fronted interaction since it provides a context for practice and meaningful use of the target language (Philp, Walter &Basturkmen, 2010).

CONCLUSION
Students in Turkey as all students around the world need to communicate in a second or foreign language for different reasons, in different contexts, and under different conditions.

517
Effective speaking skill in a second language is an indispensableasset for students to pursue career goals, and to catch up with world trends. For this reason, teaching effective speaking skills is a high priority for the teaching profession. However, students in Turkey still have problems when it comes to speaking a foreign language despite of all the efforts and emphasis on importance of speaking. Therefore, the present study was conducted to investigate the effect of a particular communicative task namely pair-discussion tasks on students' oral exam and interaction scores. More specifically, the study aims to examine the effect of a particular task type on students speaking skills by using the oral exam scores. Participants in the experimental group received 10 sessions of pair-discussion task in the form of two tasks per week. Participants in the control group didn't get any different treatment than their usual class procedures. The findings indicated that pair-discussion tasks have a positive effect on students' speaking ability and their interaction with their partners. Participants in the experimental group benefited from the task sessions and their oral exam scores and interaction scores increased. In addition, with these tasks, more time was allocated to speaking activities in class. This may cause learners to consider this skill asimportant as grammar, vocabulary or etc. since their previous teachers most likely focused on these skills rather than speaking. The outcome of the study revealed that pair-discussion tasks are effective in terms of improving students' speaking skills. Students have the opportunity to share their own ideas and also learn about their friends. When authentic and appropriate topics are chosen, this task type becomes more meaningful for students. They make a conversation that the can have with their friends in daily life by using their foreign language. In addition, to complete a discussion task, to reach a decision together they need to learn to listen to each other. This increases student-student interaction. Therefore, teachers should include this task type more in their classes especially if students are tested using this kind of a task. Moreover, test book writers should be aware of the effectiveness of this task type and add more examples. They also should revise existing once, since some of them are not applicable and suitable for students' language level. Furthermore, teachers should encourage their students to interact with their peers while completing a speaking task. Most teachers are consented when students utter their memorized answers but this isn't a real communication. However, there are some limitations of the study. The biggest problem was timing. The study only could take five weeks because it was the end of the semester. To obtain more reliable results longer studies are needed. Moreover, future studies should benefit from also qualitative data. Students should have been interviewed about the effectiveness of the tasks but because of spring festival and attendance problems qualitative data wasn't collected in this study.
In conclusion, only the effectiveness of pair-discussion tasks on students' overall speaking exam and interaction scores was investigated because there is no study conducted as such. Therefore, there is a need to do further research with regard to this concept. Since the impact of pair-discussion tasks on students' speaking skills was explored, it is suggested that the effectiveness of other tasks should be considered. Furthermore, other supplementary instruments like interviews and questionnaires about the effectiveness of pair-discussion task may be included to support the study data. Regarding the impact of pair-discussion tasks, future studies can be conducted with different participant groups and with the help of more participants. Over and above, it is worth mentioning that the pair-discussion tasks are not the