The Use of Speaking Strategies by Pre-Service EFL Teachers

Speaking is a language skill that dominates the notion of communicative language competence. Language teachers, especially early starters’ pre-service teachers, should undergo very intensive programmes of pronunciation practice as they will in many cases present the only models for their learners to imitate (Vilke 1993). To develop such fluency in speaking and propositional accuracy, students and prospective teachers should not only use but also be aware of a range of speaking strategies. This study examines pre-service EFL teachers’ perceived use of speaking strategies, as defined in the Oral Communication Strategy Inventory (Nakatani 2006). Since previous studies have identified various factors associated with learners' strategy use, we focused on determining whether the participants’ perceived strategy use is related to their EFL and speaking proficiency and their preference for engaging in speaking activities in their EFL classes. The results confirmed rather high strategy use, but the relation between the tested variables was only partially confirmed.


Introduction
Subject knowledge, which includes the knowledge of second language acquisition, pedagogical knowledge, curricular knowledge, cultural knowledge, language awareness and language proficiency (Richards et al. 2012), is the foundation of a teacher's professional experience. With respect to language proficiency, it goes without saying that EFL teachers need to possess knowledge about the grammar and vocabulary of the English language, as well as the reading, writing, listening and speaking skills necessary to use the language. "Having an excellent command of the target language is indeed one of the most important characteristics of outstanding foreign language teachers" (Shin 2008, 59), and for the nonnative teacher "language proficiency will always represent the bedrock of their professional confidence" (Murdoch 1994, 254). According to Richards (2011), there is a threshold proficiency level the teacher needs to have reached in the target language in order to be able to teach effectively in this language. In this paper we will focus on speaking, which as a language skill contributes to the overall development of communicative language competence, and is also represented as the basis of that competence (Martínez-Flor, Usó-Juan, and Alcón Soler 2006).
Speaking enables people to send their intended message to others using speech sounds which they produce themselves, activating their speech organs. However, speaking is more than just the production of sounds which are put together into meaningful units, since "learning speaking, whether in a first or other language, involves developing subtle and detailed knowledge about why, how and when to communicate, and complex skills for producing and managing interaction, such as asking a question or obtaining a turn" (Burns and Seidlhofer 2010, 197). The development of speaking skills can be more effective if language learners employ speaking strategies.
According to Martínez-Flor, Usó-Juan and Alcón Soler (2006, 151), "speakers need to become competent in using strategies in order to overcome limitations due to a lack of competence in any of the other components" (e.g. discourse, linguistic, pragmatic, intercultural and strategic competence) "integrating the proposed communicative competence framework". Many factors, such as the type of speaking assignment, students' cultural background or their level of proficiency can influence the choice of speaking strategies used (Chamot 2005). The teacher's role is to make students aware of the strategies they are already using and those they could use, so they are less restrained while deciding upon an appropriate and effective strategy for their speaking task. In that way, the teacher directs students towards new possibilities and eventually makes them use new strategies which they have not previously used or even taken into account. In that respect, Chamot (2005, 123) proposes explicit strategy instruction which "includes the development of students' awareness of their strategies, teacher modeling of strategic thinking, identifying the strategies by name, providing opportunities for practice and self-evaluation". Teachers should make students aware of the many existing strategies and explain each strategy in terms of its role and function, i.e. instruct students on "how, when, and why to use the strategy" (Anderson 2005, 758). In this way, students will be able to discover for themselves which strategies they find most beneficial (Anderson 2005). On the other hand, research conducted by Eslinger (2000as cited in Anderson 2005) draws attention to implicit strategy instruction, which could also be beneficial to students since "there may be a natural tendency to grow in strategy use without explicit instruction".
Another issue concerning the instruction of speaking strategies is whether using L1 in an EFL classroom to teach speaking strategies should be a common practice or whether it should be avoided. Chamot (2005) proposes the use of L1 if learners are not proficient enough to understand the teacher's explanation of a specific strategy in English. Usually, younger and/ or beginner learners are the ones who do not understand English well enough and may thus benefit from the teacher's decision to explain these strategies in L1. Nevertheless, using L2 to explain speaking strategies should be introduced gradually in the process of language teaching.
It is important for pre-service EFL teachers to be familiar with a range of speaking strategies so that they can teach them in EFL classrooms. In this paper, we will take Nakatani's (2006) eight categories of strategies learners use for coping with speaking problems as the theoretical foundation. These eight categories are the following: 1) social affective strategies, where learners try to control their own anxiety and enjoy the process of oral communication, they are willing to encourage themselves to use English, to risk making mistakes, and attempt to give a good impression and avoid silence during interactions; 2) fluency-oriented strategies, used to speak as clearly as possible so that their interlocutors can understand them, paying attention to the cultural context in which their conversation takes place to avoid misunderstandings; 3) negotiation for meaning while speaking strategies, used when checking whether their interlocutors have understood them or not, repetition and providing examples in order to enhance the listener's understanding of their intended message; 4) accuracy-oriented strategies, employed to self-assess the grammatical structures used to determine whether these structures are correct or not, and the attempt to sound like a native speaker; 5) message reduction and alteration strategies, where learners try "to avoid a communication breakdown by reducing an original message, simplifying their utterances, or using similar expressions that they can use confidently" (Nakatani 2006, 155); 6) nonverbal strategies while speaking, used when the message is not communicated through speech only, but through gestures, eye-contact, and so on; 7) message abandonment strategies, used to bring the interaction to an end by giving up on communicating the intended message to the interlocutor (e.g. by asking someone for help, although in this case learners do not give up on conveying their message completely); 8) attempt to think in English strategies, when learners try to think in English while speaking in English.

Literature Review
Speaking is a complex skill that consists of many sub-skills including phonology and pronunciation. In order to be able to speak English well, speaking strategies for phonology and pronunciation should be employed. Moyer (2014as cited in Oxford 2017 conducted research which confirmed that learners who were exceptionally good at phonology reported using learning strategies such as self-monitoring, imitation of native speakers, attention to difficult phonological terms and explicit concern for pronunciation accuracy. Dadour and Robbins (1996) showed that students who were taught speaking strategies used these strategies more often, and therefore their speaking skills were better than those of students without strategy instruction. Students also reported that they wanted to continue with strategy instruction, as this would help them develop a satisfying level of oral proficiency. Other researchers have also concluded that by employing various learning strategies, learners' language performance improves (Anderson 2005).
In line with those findings, Kawai (2008) showed how the use of speaking strategies by two very good Japanese EFL speakers contributed to their speaking proficiency. In order to build their confidence, they practiced orally in advance of any English language encounters; gathered information on potential discussion topics through books, the Internet, and interviews; sought help from native speakers if available; anticipated the comments of others; planned and prepared flexible conversational expressions to employ; reviewed discussion procedures; anticipated communication breakdown and the strategies to use if it happened; and made and followed plans to speak English every day. Moreover, the research showed that learning strategies of a non-compensatory sort (e.g. metacognitive strategies such as planning and monitoring, cognitive strategies which enhance grammar and vocabulary, affective strategies and social strategies) are helpful for improving speaking. Kawai (2008) concluded that those learners who develop good oral skills appear to be frequent strategy users, regardless of culture and learning context. Zhang and Goh's (2006) research with 278 Singaporean secondary school learners of English confirmed that the number and level of strategies used are related to learners' proficiency in the foreign language. Méndez López (2011) also compared the use of speaking strategies by university students of English with their proficiency levels (beginners, intermediates and advanced). The results showed that the use of speaking strategies was not the same at the three proficiency levels. More specifically, all students reported using similar speaking strategies, but the frequency of strategy use was related to the students' proficiency level (Méndez López 2011). The three strategies that were employed the most were: asking for repetition, the use of paraphrasing or a synonym for unknown words and asking for clarification of a message (Méndez López 2011). The author concluded with the suggestion that strategy training should be implemented in language courses, interspersed with communicative activities (Méndez López 2011).
A more recent study by Pawlak (2018) investigated what speaking strategies higher-proficiency English language learners use prior to, during and after a speaking task. The research showed that "the employment of SSs is bound to be conditioned by the type of activity, the demands it places on interlocutors, and the communicative goals it sets" (Pawlak 2018, 286). The results with regard to the specific speaking strategies the participants employed before, during and after performing two different speaking tasks showed that the participants mostly relied on metacognitive and social strategies (Pawlak 2018). Some of the metacognitive strategies they used were preparing for their speech by choosing suitable vocabulary and deciding upon the arguments which would support their opinions. They also reflected upon their grammatical accuracy during the communication tasks. An example of a social strategy that was employed was students cooperating in order to complete the two speaking tasks by asking each other different questions (Pawlak 2018). ENGLISH LANGUAGE AND LITERATURE TEACHING In view of these studies, the present research 1 focused on identifying speaking strategy use by pre-service EFL teachers and its possible correlations with several factors which have been identified as important in previous studies.

The Study
The present study investigates pre-service EFL teachers' perceptions of their use of speaking strategies with respect to their self-assessed language and speaking proficiency, as well as preference for speaking in EFL classes. The aim of the study is to establish whether pre-service EFL teachers use the speaking strategies which are available for them as language learners and future EFL teachers. This is especially important because their knowledge and use of speaking strategies will have a significant and immediate impact on their learners.
The research hypotheses are as follows: 1. The results will show significant use of speaking strategies by pre-service EFL teachers.
2. Pre-service EFL teachers who assess their proficiency in English with a higher grade will report using speaking strategies more often.
3. Pre-service EFL teachers who assess their speaking skills in English with a higher grade, and who claim that speaking is their favourite activity in the English classes, will report using speaking strategies more often.

Participants
The research comprised 50 participants, students at the Faculty of Teacher Education University of Zagreb studying within the Integrated graduate and undergraduate study of primary teacher education with English language.

Research Instrument
An online questionnaire, designed for the purpose of this study, was available for participants to complete throughout March, April and May 2018. Prior to responding to the two-part online questionnaire, the participants completed a consent form.
The initial part of the questionnaire contained questions regarding the participants' age, gender, current year of study and questions related to their English language learning history and self-assessment of their proficiency in English and speaking.
The second part of the questionnaire was an adapted version of the Oral Communication Strategy Inventory (OCSI) developed by Nakatani (2006). The original inventory contains two sets of items, the first one addressing strategies for coping with speaking problems and the second containing strategies for coping with listening problems. For the purpose of the present research, the set of items on strategies for coping with listening problems 1 The research described in this paper is part of a study conducted in the process of writing the first author's graduation thesis.
was excluded from the questionnaire. The OCSI, "a reliable and valid strategy inventory for communication tasks" (Nakatani 2006, 152) consists of 32 items indicating strategies for coping with speaking problems. Each item is evaluated on a five-point Likert-type scale whereby 1 indicates Never or almost never true of me; 2. Generally not true of me; 3. Somewhat true of me; 4. Generally true of me; 5. Always or almost always true of me. The strategies were divided into eight categories, based on the factor analysis results. Hence, Factor 1 are social affective strategies (items 23,25,26,27,28 and 29,Factor 2 (items 9,10,11,12,13 and 14) are fluency-oriented strategies, Factor 3 (items 19, 20, 21 and 22) are negotiation for meaning while speaking strategies, Factor 4 (items 7, 8, 17, 18 and 30) are accuracy-oriented strategies, Factor 5 (items 3, 4 and 5) are message reduction and alteration strategies, Factor 6 are nonverbal strategies (items 15 and 16), Factor 7 (items 6, 24, 31 and 32) are message abandonment strategies, and Factor 8 (items 1 and 2) are attempt to think in English strategies.
All of the items from the set of strategies for coping with speaking problems were included in the questionnaire, with some of them being slightly adapted to aid comprehension by the students.
In addition, although the instructions for the original OCSI state that it is to be completed after a specific speaking activity, for the purpose of the present research the questions were posed and analysed so as to refer to the general speaking behaviour of the participants.
The results obtained from the OCSI were analysed using descriptive statistics, i.e. mean and SD values indicated for the overall strategy use as well as for individual strategy use.

General Information on the Participants
The sample of participants comprised 50 students, of whom 48 were female and two were male. The distribution of students according to the year of study was as follows: 19 students in the second year, five students in the third year, four students in the fourth year and 22 students in the final, fifth year of study. The age span of the participants ranged from 19 to 27, with most of the participants being 23 (N = 15) and 20 (N = 13) years old. All of the participants reported having learned English prior to their enrolment at the Faculty of Teacher Education.
The majority (N = 44) reported that their final grade in English in primary school was 5 (i.e. excellent), while only six were graded with a 4 (i.e. very good). Their final grades slightly deteriorated in secondary school with N = 35 of them being graded with 5, N = 14 with 4, and N = 1 with 3 (i.e. good).
Students' exposure to the English language in their free time (e.g. listening to music, watching films, reading, surfing the Internet, communicating in English, etc.) was estimated as follows. More than half of them (N = 27) reported being exposed to English more than 10 hours per week, 11 participants being exposed to English up to 10 hours per week, while 12 participants stated that their exposure does not exceed 5 hours per week. Finally, the majority of the participants (N = 42) reported communicating with a native speaker of English at least 21 ENGLISH LANGUAGE AND LITERATURE TEACHING at some point in their lives, while the rest (N = 8) said they have never been in contact with a native speaker.

Self-Assessed Language and Speaking Proficiency and Awareness of Strategies
Speaking (N = 33; 66%) and listening (N = 23; 46%) were selected by the majority of participants as their favourite activities in English class. Reading silently (N = 19; 38%), writing (N = 17; 34%) and reading out loud (N = 11; 22%) were chosen by fewer participants. When asked to choose one or more skills in English at which they consider themselves to be very good, the most frequently chosen answers were listening (N = 30; 60%) and speaking (N = 28; 56%). Reading silently and writing were chosen the same number of times (N = 26; 52%), while reading out loud was chosen 23 times (46%). Further along, the participants self-assessed their EFL proficiency on a scale from 5 to 1 (5=excellent; 4=very good; 3= good; 2= satisfactory; 1=fail) Most of them assessed their EFL proficiency with 4 (N = 33; 66%), followed by 5 (N = 13; 26%) and 3 (N = 4; 8%). The mean grade for participants' self-assessed general knowledge of English was M = 4.18 (SD = 0.56).
The self-assessment of their speaking skills showed that the most common grade was 4 (N = 27; 54%), although there were also some 3s (N = 8; 16%) and 5s (N = 15; 30%). The mean grade for the participants' self-assessed speaking skills in English was M = 4.14 (SD = 0.67).
We also compared the self-assessment of the participants' speaking skills in the foreign language (English) with the self-assessment of the speaking skills in their mother tongue (Croatian). The results were as follows: an equal number of students (N = 24; 48%) graded their speaking skills in the mother tongue with 5 and 4. Only 4% (N = 2) graded it with 3. The mean grade for the self-assessed speaking skills in Croatian was M = 4.44 (SD = 0.58). Even though the participants assessed their speaking skills in Croatian with slightly higher grades than their speaking skills in English, the difference is not relevant. Some of the possible reasons might be that they feel confident while speaking in both their mother tongue and English, or that their standards for assessing these two skills were not the same (maybe they had higher standards for Croatian than English). Also, since the participants are pre-service EFL teachers and their proficiency is targeted at the C1 level of the CEFR (2001) (see Cindrić, Cergol, and Davies 2010), it is not surprising that their self-assessed English language proficiency is high.
The participants also had to explain what they thought speaking and learning strategies were. With regard to the former, the answers showed that some of the participants did not know what speaking strategies are, or had difficulties explaining them. These are some of their answers: "The way we say things" and "Mechanism to help you speak more easily and fluently". Secondly, they were asked to explain/say the meaning of learning strategies, and some of their answers were: "To find the best way to learn something.", "Watching movies, communicating, repeating the words you've studied, writing them down, connecting them with Croatian words...", and "visual, auditory, kinesthetic and multimodal strategies". The obtained answers show that some participants confused learning strategies with learning styles, some did not differentiate between strategic and non-strategic learning, some identified learning strategies in general with social strategies as one type of learning strategies, while some provided good examples of learning strategies.

Self-Reported Use of Speaking Strategies (OCSI)
The results obtained from the OCSI indicate that the average use of strategies by the participants was M = 3.77 (SD = .34). Strategies defined by inventory items 1, 24 and 32 were excluded from the total calculation, as they relate to speaking behaviours which generally do not have a positive impact on one's speaking skills. Those speaking behaviours are, for example, first thinking of what to say in one's native language and then constructing the English sentence, leaving a message unfinished because of some language difficulty, and giving up when one cannot make oneself understood.
In the text below, the results of the OCSI will be presented according to the eight factors, i.e. types of strategies. Variables in Factor 1 address social affective strategies, and they include six strategy items for which the following mean results were obtained: 1 I try to use fillers when The results obtained from the OCSI indicate that the average use of strategies by the participants was M = 3.77 (SD = .34). Strategies defined by inventory items 1, 24 and 32 were excluded from the total calculation, as they relate to speaking behaviours which generally do not have a positive impact on one's speaking skills. Those speaking behaviours are, for example, first thinking of what to say in one's native language and then constructing the English sentence, leaving a message unfinished because of some language difficulty, and giving up when one cannot make oneself understood.
In the text below, the results of the OCSI will be presented according to the eight factors, i.e. types of strategies. Variables in Factor 1 address social affective strategies, and they include six strategy items for which the following mean results were obtained: 1 I try to use fillers when  It may be observed that the highest mean for the perceived use of speaking strategies is for the strategy I try to enjoy the conversation, while the lowest mean was recorded for the strategy I try to use fillers when I cannot think of what to say. Since these students are instructed in and frequently made aware of the importance of their fluency in communication and the importance of using fillers in achieving this goal, such low results are rather surprising. On the other hand, it may be that due to their perceived high EFL/speaking competence they did not consider fillers   It may be observed that the highest mean for the perceived use of speaking strategies is for the strategy I try to enjoy the conversation, while the lowest mean was recorded for the strategy I try to use fillers when I cannot think of what to say. Since these students are instructed in and frequently made aware of the importance of their fluency in communication and the importance of using fillers in achieving this goal, such low results are rather surprising. On the other hand, it may be that due to their perceived high EFL/speaking competence they did not consider fillers to be a useful strategy, or they may simply not be aware of their use.   The results show that the highest mean value for the perceived strategy use was recorded for the strategy I try to speak clearly and loudly to make myself heard. The lowest mean was recorded for the strategy I take my time to express what I want to say, and it may be a reflection of the participants' awareness of their high language proficiency in EFL which enables them to communicate confidently and without hesitation. According to the Common European Framework of Reference for Languages (2001), the functional success of the learner/user is described through fluency and propositional accuracy. Fluency is the ability to articulate, to keep going, and to cope when at a dead end, whereas propositional precision is the ability to formulate thoughts and propositions so as to make one's meaning clear (Cindrić, Cergol and Davies 2010). Considering the participants' training as future teachers of English, it is expected that their fluency is at the C1 level of CEFR, which indicates the following 'can do' statement: Can express him/herself fluently and spontaneously, almost effortlessly. Only a conceptually difficult subject can hinder a natural, smooth flow of language (CEFR 2001). The results of this research are in line with the set level.
Factor 3 is negotiation for meaning while speaking strategies, and the following values were obtained for these four strategy items: 1 While speaking, I pay attention to the listener's reaction The results show that the highest mean value for the perceived strategy use was recorded for the strategy I try to speak clearly and loudly to make myself heard. The lowest mean was recorded for the strategy I take my time to express what I want to say, and it may be a reflection of the participants' awareness of their high language proficiency in EFL which enables them to communicate confidently and without hesitation. According to the Common European Framework of Reference for Languages (2001), the functional success of the learner/user is described through fluency and propositional accuracy. Fluency is the ability to articulate, to keep going, and to cope when at a dead end, whereas propositional precision is the ability to formulate thoughts and propositions so as to make one's meaning clear (Cindrić, Cergol and Davies 2010 Figure 3 shows the participants' assessment of their negotiation for meaning while speaking strategy use. (M = 4.10, SD = .91), and 4 I make comprehension checks to ensure the listener understands what I want to say (M = 3.68, SD = 1.10). Figure 3 shows the participants' assessment of their negotiation for meaning while speaking strategy use.   (Byrd et al. 1993), and since the participants are pre-service EFL teachers, such significant use of this speaking strategy is understandable.
For the five strategy items grouped under Factor 4, i.e., accuracy-oriented strategies, the following mean values were calculated: 1 I pay attention to grammar and word order during conversations (M = 4.30, SD = .86), 2 I try to emphasize the subject and verb of the sentence (M = 3.04, SD = .95), 3 I correct myself when I notice that I have made a mistake (M = 4.62, SD = .57), 4 I notice myself using an expression which fits a rule that I have learned (M = 4.08, SD = .80), and 5 I try to talk like a native speaker (M = 4.12, SD = .98). Participants' selfassessed accuracy-oriented strategy use is presented in Figure 4.   (Byrd et al. 1993), and since the participants are pre-service EFL teachers, such significant use of this speaking strategy is understandable.
For the five strategy items grouped under Factor 4, i.e., accuracy-oriented strategies, the following mean values were calculated: 1 I pay attention to grammar and word order during conversations (M = 4.30, SD = .86), 2 I try to emphasize the subject and verb of the sentence (M = 3.04, SD = .95), 3 I correct myself when I notice that I have made a mistake (M = 4.62, 25 ENGLISH LANGUAGE AND LITERATURE TEACHING SD = .57), 4 I notice myself using an expression which fits a rule that I have learned (M = 4.08, SD = .80), and 5 I try to talk like a native speaker (M = 4.12, SD = .98). Participants' selfassessed accuracy-oriented strategy use is presented in Figure 4. The strategy I correct myself when I notice that I have made a mistake is reported to be used the most, while I try to emphasize the subject and verb of the sentence is used the least. It is not surprising that self-correction was the strategy used most in this category, as well as the second most used strategy in the entire inventory, since self-correction has been identified as an essential form of error correction and a procedure that may contribute to learner autonomy in contemporary learner-centred educational settings (Edwards 2000, Sultana 2009). I try to emphasize the subject and verb of the sentence.
I correct myself when I notice that I have made a mistake.
I notice myself using an expression which fits a rule that I have learned.
I try to talk like a native speaker.
Note: 1. Never or almost never true of me; 2. Generally not true of me; 3. Somewhat true of me; 4. Generally true of me; 5. Always or almost always true of me.
The strategy I correct myself when I notice that I have made a mistake is reported to be used the most, while I try to emphasize the subject and verb of the sentence is used the least. It is not surprising that self-correction was the strategy used most in this category, as well as the second most used strategy in the entire inventory, since self-correction has been identified as an essential form of error correction and a procedure that may contribute to learner autonomy in contemporary learner-centred educational settings (Edwards 2000, Sultana 2009).
The perceived use ( Figure  The highest and lowest perceived mean uses of message reduction and alternation strategy were recorded for the strategies I use words which are familiar to me, and I reduce the message (what I want to say) and use simple expressions, respectively. However, this category had an overall lower mean result in comparison to most other categories, which may be related to the participants' high EFL proficiency, since Metcalfe and Noom-Ura (2013) found that the mean result for this strategy category was lower for the group of high proficiency learners in their research.
Strategies concerned with the aspect of communication which does not include speech are grouped under Factor 6, and are called nonverbal strategies while speaking. These were assessed by the participants in the following way: 1 I try to make eye-contact when I am talking (M = 4.46, SD = .86), and 2 I use gestures and facial expressions if I do not know how to say something (M = 4.28, SD = .83). Figure 6 shows how the participants assessed their use of nonverbal strategies while speaking. The results were, as expected, rather high for both of the assessed strategies, although the participants perceive that they use the strategy I use gestures and facial expressions if I do not know how to say something somewhat more than I try to make eye-contact when I'm talking. The importance of nonverbal strategies was recognized by Canale and Swain (1980, 30), who claimed that strategic competence relies on "verbal and nonverbal communication strategies to compensate for breakdowns in communication due to performance variables or due to The results were, as expected, rather high for both of the assessed strategies, although the participants perceive that they use the strategy I use gestures and facial expressions if I do not The highest and lowest perceived mean uses of message reduction and alternation strategy were recorded for the strategies I use words which are familiar to me, and I reduce the message (what I want to say) and use simple expressions, respectively. However, this category had an overall lower mean result in comparison to most other categories, which may be related to the participants' high EFL proficiency, since Metcalfe and Noom-Ura (2013) found that the mean result for this strategy category was lower for the group of high proficiency learners in their research.
Strategies concerned with the aspect of communication which does not include speech are grouped under Factor 6, and are called nonverbal strategies while speaking. These were assessed by the participants in the following way: 1 I try to make eye-contact when I am talking (M = 4.46, SD = .86), and 2 I use gestures and facial expressions if I do not know how to say something (M = 4.28, SD = .83). Figure 6 shows how the participants assessed their use of nonverbal strategies while speaking. Note: 1. Never or almost never true of me; 2. Generally not true of me; 3. Somewhat true of me; 4. Generally true of me; 5. Always or almost always true of me. know how to say something somewhat more than I try to make eye-contact when I'm talking. The importance of nonverbal strategies was recognized by Canale and Swain (1980, 30), who claimed that strategic competence relies on "verbal and nonverbal communication strategies to compensate for breakdowns in communication due to performance variables or due to insufficient competence".
Message abandonment strategies, categorized under Factor 7, are employed when speakers give up on delivering their original message because they do not feel capable of doing so. This group of strategies (Figure 7) was assessed by the participants as follows: 1 I abandon the execution of the original message and just say some words when I don't know what to say (M = 2.32, SD = 1.12), 2 I leave a message unfinished because of some language difficulty (M = 1.98, SD = .74), 3 I ask other people to help when I can't communicate well (M = 3.74, SD = 1.14), and 4 I give up when I can't make myself understood (M = 2.04, SD = .99). The overall analysis shows that the results for message abandonment strategy use were lower than those obtained for the previously presented strategy groups. The highest perceived mean use was recorded for the strategy I ask other people to help when I can't communicate well, while the lowest mean was recorded for the strategy I leave a message unfinished because of some language difficulty, which is also the lowest recorded value in the entire inventory. This particular strategy category has been found to be significantly more used by low proficiency learners (Metcalfe and Noom-Ura 2013). Therefore, this result actually speaks in favour of the participants and reflects positive practices in teaching, as this particular group of students are continually encouraged throughout their studies to employ various speaking strategies to avoid breakdowns in communication.
Factor 8 includes attempt to think in English strategies, and the following results were obtained for these two strategies: 1 I first think of what I want to say in my native language and then construct the English sentence (M = 2.34, SD = 1.00), and 2 I first think of a sentence I already know in English and then try to change it to fit the situation (M = 2.62, SD = 1.18) (Figure 8). The overall analysis shows that the results for message abandonment strategy use were lower than those obtained for the previously presented strategy groups. The highest perceived mean use was recorded for the strategy I ask other people to help when I can't communicate well, while the lowest mean was recorded for the strategy I leave a message unfinished because of some language difficulty, which is also the lowest recorded value in the entire inventory. This particular strategy category has been found to be significantly more used by low proficiency learners (Metcalfe and Noom-Ura 2013). Therefore, this result actually speaks in favour of the participants and reflects positive practices in teaching, as this particular group of students are continually encouraged throughout their studies to employ various speaking strategies to avoid breakdowns in communication.
Factor 8 includes attempt to think in English strategies, and the following results were obtained for these two strategies: 1 I first think of what I want to say in my native language and then construct the English sentence (M = 2.34, SD = 1.00), and 2 I first think of a sentence I already know in English and then try to change it to fit the situation (M = 2.62, SD = 1.18) (Figure 8). The results for these two strategies were also lower in comparison to the other strategies, with the lower mean results recorded for I first think of what I want to say in my native language and then construct the English sentence. This indicates that most students do not use this strategy, which is positive since thinking in the language of communication as much as possible is considered important, and has been proven to be more useful than thinking in one's mother tongue (Nakatani 2006, 155-56).
Based on the presented results, it may be concluded that, apart from the message abandonment and attempt to think in English strategies, most of the obtained mean values were generally rather high as well as the overall result. Therefore, the first hypothesis, H1. The results will show significant use of speaking strategies by pre-service EFL teachers, was confirmed.

The Relationship Between the Participants' Use of Speaking Strategies and Their Self-Assessed EFL Knowledge and Speaking Skills
A series of Spearman rank-order correlation analyses were conducted to test the second hypothesis: H2. Pre-service EFL teachers who assess their knowledge of English with a higher grade (4 -very good and 5 -excellent) will report using speaking strategies more often. The results confirmed a statistically significant weak positive correlation between the participants' self-assessed knowledge of English and their perceived use of the following four speaking strategy categories: social affective (rs (50)=.34, p< .05), fluency-oriented (rs (50)=.32, p< .05), negotiation for meaning while speaking (rs (50)=.39, p< .01), and accuracy-oriented category (rs (50)=.38, p< .01). In other words, the participants who assessed their EFL proficiency to be (relatively) high reported using the above-mentioned speaking strategies more often. Therefore, The results for these two strategies were also lower in comparison to the other strategies, with the lower mean results recorded for I first think of what I want to say in my native language and then construct the English sentence. This indicates that most students do not use this strategy, which is positive since thinking in the language of communication as much as possible is considered important, and has been proven to be more useful than thinking in one's mother tongue (Nakatani 2006, 155-56).
Based on the presented results, it may be concluded that, apart from the message abandonment and attempt to think in English strategies, most of the obtained mean values were generally rather high as well as the overall result. Therefore, the first hypothesis, H1. The results will show significant use of speaking strategies by pre-service EFL teachers, was confirmed.

The Relationship Between the Participants' Use of Speaking Strategies and Their Self-Assessed EFL Knowledge and Speaking Skills
A series of Spearman rank-order correlation analyses were conducted to test the second hypothesis: H2. Pre-service EFL teachers who assess their knowledge of English with a higher grade (4 -very good and 5 -excellent) will report using speaking strategies more often. The results confirmed a statistically significant weak positive correlation between the participants' self-assessed knowledge of English and their perceived use of the following four speaking strategy categories: social affective (r s (50)=.34, p< .05), fluency-oriented (r s (50)=.32, p< .05), negotiation for meaning while speaking (r s (50)=.39, p< .01), and accuracy-oriented category (r s (50)=.38, p< .01). In other words, the participants who assessed their EFL proficiency to be (relatively) high reported using the above-mentioned speaking strategies more often. Therefore, it may be proposed that the second hypothesis was confirmed partially, as the correlations were weak, and they were confirmed for only four out of eight speaking strategy categories.
In order to test the third hypothesis (H3. Pre-service EFL teachers who assess their speaking skills in English with a higher grade, and who claim that speaking is their favourite activity in the English classes will report using speaking strategies more often), a correlation analysis and two one-way ANOVA tests were applied. A series of Spearman rank-order correlations confirmed a statistically significant weak positive correlation between the participants' self-assessed speaking skills in English and their perceived use of only one speaking strategy category: social affective (r s (50)=.38, p< .01). A statistically significant weak correlation was confirmed for one more strategy category, message reduction and alteration (r s (50)= -.29, p< .05), but this one was negative, i.e. the participants who assessed their speaking skills with a higher grade reported using this speaking strategy category less often.
The first one-way ANOVA test confirmed a statistically significant difference in favour of the participants who mentioned speaking as their favourite activity in EFL classes for the social affective: F(1,49)=22.958, p=.000, and accuracy-oriented strategy categories: F(1,49)=6.817, p=.012. In other words, their use of these speaking strategies was perceived to be greater than that of the participants who did not refer to speaking as their favourite activity. A statistically significant difference was also found for message reduction and alteration strategies: F(1,49)=5.757, p=.020), but it was in favour of the participants who did not identify speaking as their favourite activity in EFL classes. The results show that having speaking as a favourite activity was determined as a relevant factor only for two out of eight strategy categories.
The second one-way ANOVA test confirmed a statistically significant difference in favour of the participants who see themselves as being good at speaking for the social affective: F(1,49)=20.308, p=.000; fluency-oriented: F(1,49)=4.032, p=.050; and accuracy-oriented strategy categories: F(1,49)=4.712, p=.035. For two categories the difference was statistically significant, but it was in favour of those who did not perceive themselves as being good at speaking (message reduction and alteration: F(1,49)=5.645, p=.022, and attempt to think in English: F(1,49)=4.131, p=.048). The fact that participants who do not perceive themselves as being good at speaking reported using message reduction strategy category more often is in accordance with previous findings indicating that this strategy category was used more often by low proficiency learners (Metcalfe and Noom-Ura 2013).
The results of the second ANOVA test show that being good at speaking was proven a relevant factor only for three out of the eight strategy categories. Therefore, it may be proposed that the third hypothesis was only partially confirmed, as the relationship between the perceived use of speaking strategies and the three tested variables was confirmed only for some of the strategy categories.

Conclusion
The study presented students' (pre-service primary school English teachers) perceptions of their use of speaking strategies with respect to their self-assessed language and speaking proficiency, as well as preference for speaking in EFL classes. The results showed significant use of speaking strategies by pre-service EFL teachers (H1) although, when asked to define both learning and speaking strategies, some participants were not able to provide satisfactory explanations. Hence, it may be proposed that the majority of the target group of learners predominantly showed satisfactory use of strategies acquired in the course of their studies. The somewhat weaker knowledge or understanding of strategies recorded in this sample may be related to the fact that some of the participants were only in their second year of study, and they have not yet had any EFL teaching methodology classes.
Considering that the students in this sample will in the future teach the English language to young learners, their expected level of accuracy and fluency in the four skills is a high one. This in particular refers to grammar and pronunciation (Cindrić, Cergol, and Davies 2010), as it is important that their speech is comprehensible and grammatically correct. This may be the reason for high mean results obtained for the majority of fluency-and accuracy-oriented strategies, and for the fact that in addition to the social affective and negotiation for meaning while speaking strategies, the fluency-and accuracy-oriented strategies were positively correlated with the participants' self-assessed EFL proficiency. It is also worth mentioning that some of the strategies for which the lowest mean results were obtained are actually considered less efficient, and students are warned not to use them, e.g. the message abandonment strategy category and the strategy I first think of what I want to say in my native language and then construct the English sentence.
Since the second hypothesis (Pre-service EFL teachers who assess their knowledge of English with a higher grade will report using speaking strategies more often.) and third hypothesis (Preservice EFL teachers who assess their speaking skills in English with a higher grade, and who claim that speaking is their favourite activity in the English classes will report using speaking strategies more often.), were only partially confirmed, we propose considering additional factors that may be related to EFL learners' speaking strategy use, or even expanding the inventory with additional open-ended questions related to speaking strategies.