Indonesian EFL Students’ Perceptions of Effective Non-Native English Teachers

Effective non-native English teachers (NNETs) are essential to facilitate students to develop their English mastery. However, while students are directly affected by teachers’ instructions, they have been limitedly involved in teacher education and development research. The current study aims to explore the qualities of effective non-native English teachers (NNETs) from students’ perspectives in Tanjungpinang, Indonesia. This study was mixed-method research specifically an explanatory sequential design. In the quantitative phase, 380 students were selected using cluster sampling techniques. Meanwhile, in the qualitative phase, six students were purposely selected based on their English proficiency levels and gender. The instruments used were a questionnaire adapted from Park and Lee (2006) and an interview guide. To analyze the questionnaire data, descriptive and inferential statistical analyses were employed. MANOVA test was run to seek significant differences by students’ gender and English proficiency levels. Meanwhile, the qualitative data were coded to identify the emerging patterns. The results show that the qualities concerning teachers’ socio-affective skills gained the highest mean scores compared to those concerning teachers’ subject matter knowledge and pedagogical skills. This study also confirms that there was a significant difference among students in perceiving effective NNETs by English proficiency levels. This study is expected to help various stakeholders to improve preand in-service EFL teacher education and development in Indonesia. * Corresponding author, email: ummizurrahmi95@gmail.com Citation in APA style: Zurrahmi, U., & Triastuti, A. (2022). Indonesian EFL students’ perceptions of effective non-native English teachers. Studies in English Language and Education, 9(1), 299-317. Received July 13, 2021; Revised October 15, 2021; Accepted December 6, 2021; Published Online January 17, 2022 https://doi.org/10.24815/siele.v9i1.21720 U. Zurrahmi & A. Triastuti, Indonesian EFL students’ perceptions of effective non-native English teachers | 300


INTRODUCTION
Many English foreign language (EFL) students are taught by non-native English teachers (NNETs). Effective NNETs are essential to facilitate students to develop English mastery, particularly in this globalization era where English plays an important role as an international language. Nowadays, students are expected to meet world-class standards, compete globally, and succeed in coping with future challenges in the workplace and life (Malik, 2018). To do so, students need to be able to use English as a means of communication which is useful to expand their knowledge and catch up with the development of science and technology. Therefore, NNETs play an important role as a force that highly influences students' language learning (Usman et al., 2016). Hence, NNETs teachers have the responsibility to effectively prepare students to master English. With this in mind, investigating the characteristics of effective NNETs is an essential issue in the field of second/foreign language teachers' education and development.
Studies on effective NNETs have largely taken into accounts the voices of various stakeholders, such as teachers, schools, and administration staff (Bell, 2005;Brosh, 1996;Hidayat, 2017;Karim, 2021;Kwangsawad, 2017;Park & Lee, 2006;Shishavan & Sadeghi, 2009). However, little is known about effective NNETs as represented in students' voices (Barnes & Lock, 2010Çelik et al., 2013;Khaerati, 2016;Mudra, 2018;Said, 2017). It is in line with Tarone and Allwright (2005), as well as Tsui (2011), who claim that research on English teachers has made little reference to students, and if existed, students' voices have largely remained silent. It is problematic because, in fact, students act as the party who is directly affected by the instructions managed by the teachers. Therefore, students' perceptions are worth considering, particularly since the ultimate purpose of teachers is to enhance the quality of their student learning. Moreover, considering the students as the subject of learning mandated on the law No. 20 Year 2003 (Ministry of National Education andCulture, 2003), it is important to explore students' perceptions of effective NNETs.
This study is a conceptual replication (Mackey & Gass, 2005) of Park and Lee's (2006) study. This study is "conceptually" true to the original study and carefully considers the theoretical claims of the original research (Mackey & Gass, 2005, p. 22). Therefore, this study adapted the questionnaire and the framework of teacher dimension in order to collect and analyze the data. Replication studies have been argued to be conducted in social science research as it aims to establish and confirm the outcomes of research, including in second/foreign language teaching and learning research (Brown, 2011;Language Teaching Review Panel, 2008). A conceptual replication study of Park and Lee's (2006) is considered significant to be conducted in the Indonesian context, particularly in Tanjungpinang because the conceptions among people toward effective NNETs are different depending on the levels of schooling, curriculum, gender, major area, and background (Barnes & Lock, 2010Brosh, 1996;Kwangsawad, 2017;Mudra, 2018;Park & Lee, 2006;Shishavan & Sadeghi, 2009). Therefore, the purpose of conducting this conceptual replication study is to prove whether the findings in Park and Lee's (2006) study can be extended to a different setting and to a different group of participants (Mackey & Gass, 2005), which in this case is the Indonesian context. The differences between this study and that of Park and Lee (2006) are that this study focused on the perspectives of the participating students from Indonesia (i.e., Tanjungpinang) and employed a mixed-method approach to provide a complete understanding of the following research problems: 1. How do public senior high school students perceive effective NNETs? 2. Is there any significant difference between public senior high school students by their gender in perceiving effective NNETs? 3. Is there any significant difference among public senior high school students by their English proficiency level in perceiving effective NNETs? This study is significant to the body of knowledge on teacher education and development in EFL school contexts. This study is expected to inform NNETs about the students' needs in the EFL classroom, which perhaps could help to improve EFL teaching practices, especially in the Indonesian context.

The Conceptions of Effective NNETs
'Effective teachers' is one of the terms used by people when referring to required and desired teachers. People may say 'good' (Brown, 2001;Harmer, 2007), 'outstanding' (Faranda & Clarke, 2004), 'professional' (Amin, 2015), or 'best' (Liando, 2010) teachers. However, within the context of this study, the word used is 'effective'. Walker (2020) defines effective teachers as teachers who are successful in helping students to learn. In this study, the concept of effective NNETs is in agreement with Walker's (2020) definition, which is to put less emphasis on marks or grades to judge the students' learning outcomes.

Dimensions of a Teacher
Within the context of this study, the characteristics of effective NNETs focus on three major dimensions that have a direct impact on the students' learning outcomes from the students' perspectives; they are subject matter knowledge, pedagogical content knowledge, and socio-affective skills (Park & Lee, 2006). Vélez-Rendón (2002) defines subject matter knowledge as what teachers know about what they teach. Additionally, to describe what teachers need to know regarding their subject matter, Lafayette (1993), as cited in Vélez-Rendón (2002), points out the three components of language teachers' subject matter knowledge, which comprises language proficiency, civilization and culture, and language analysis. Within the subject matter knowledge dimension, the qualities of effective NNETs were explored in accordance to teachers' language proficiency, knowledge about civilization and culture (Vélez-Rendón, 2002), knowledge about language system, materials, and resources, classroom equipment, and up-to-datedness/insightful (Harmer, 2007). Added to these, pedagogical knowledge refers to what teachers know about teaching their subjects (Vélez-Rendón, 2002). Pedagogical knowledge alludes to second/foreign language acquisition theories, teaching methods, and testing (Vélez-Rendón, 2002). In this study, within the pedagogical skill dimension, the qualities of an effective English teacher were explored in accordance with the teacher's lessons preparation, delivery, methods and strategies, classroom management, and assessment.
Finally, the term socio-affective skills refer to the emotional side of human behavior (Brown, 2007) or social and emotional behaviors (Stronge, 2018). This present study intended to investigate the characteristics of effective NNETs within the socio-affective category in accordance with teachers' physical appearance, social capacity, cultural qualities, and psychological makeup (Penner, 1992, cited in Brosh, 1996.

The Standard of Professional Teacher Competences in Indonesia
In the Indonesian context, teachers are expected to have certain competences. The act of Republic Indonesian no. 14/2005 requires teachers to have four competencies, including pedagogical, professional, social, and personal competence (President of Republic of Indonesia, 2005). In the Ministry of National Education and Culture (MNEC) regulation (Ministry of National Education and Culture, 2007), the elaboration of the teacher competence standards is presented. The elaboration of the competence standard was developed integrally from the four main competencies. Professional competence has additional competences that are specific to each subject. In the English subject, teachers should know the English language as a system and as a means of communication and master English both in spoken and written forms. Since the aforementioned act and regulation are mainly aimed at teachers in general, the results of this study could reveal particular characteristics of effective NNETs in the Indonesian context from students' perspectives.

Research Design
The design of this study was mixed-method research, specifically explanatory sequential design (Creswell, 2012). By combining quantitative and qualitative methods, it provides a better understanding of the research problem and question. In this design, the researchers implemented two phases in a sequence. The quantitative phase followed with the qualitative phases to help explain the quantitative results.

Research Location and Samples
This study was taken place in Tanjungpinang, the capital city of Kepulauan Riau province, Indonesia. This study was conducted during the Covid-19 pandemic, in which all teaching and learning activities were conducted online. The research samples for the quantitative phase were 380 students. The samples were selected by using cluster sampling techniques. From seven public senior high schools in Tanjungpinang, four schools were involved to be the representatives from each district. With an average of 30 students in each class, the researchers randomly selected 3 classes from every school. All students from selected classes were involved in this study. In the qualitative phase, six students were purposely selected to be the samples in order to seek variations in the students' perspectives based on their different English proficiency levels and gender.

Instruments
The research instrument used in the quantitative phase was a questionnaire about the criteria of effective NNETs. The questionnaire in this study was adapted from Park and Lee (2006) and was translated into Indonesian in order to make the students understand and answer each statement easily. After the questionnaire was translated into Indonesian, the content was validated by a validator. The validator is a faculty member of a state university in Daerah Istimewa Yogyakarta (i.e., Special Territory of Yogyakarta). Corrections were then made dealing with clarity of the language in the translated questionnaire as advised by the validator. Moreover, one item regarding how to 'alleviate students' anxiety in English class' in item number 20 was revised from the original questionnaire. Since this item was categorized within the socio-affective skills, the researchers revised the item to 'alleviate students' anxiety by building more relaxed relationships in the classroom'. Moreover, the researchers decided to remove two items from the questionnaire because the two items had not statistically passed the criteria to be valid and reliable. The items were 'teach English tailored to students' English proficiency levels', and 'assess what students have learned rationally'. The original questionnaire consisted of 27 items but the questionnaire employed in this study consisted of 25 items. For further data collection and analysis, some items included eliciting personal information (e.g., name, school, grade, contact person, and English proficiency). As for the item focusing on the students' English proficiency, the researcher adopted the scales from Dörnyei and Taguchi (2009). The scales were used to ease the students to self-assessed their English proficiency from the criteria available for each proficiency level.
In the qualitative phase, an interview guideline was employed to gather in-depth information regarding students' perceptions of effective NNETs. The researchers designed the guideline by referring to the conceptual framework and the research questions of the study. The interview guideline consisted of six questions. Furthermore, follow-up questions were asked to the students, when necessary, to triangulate the data with the questionnaire results. To ensure the validity of the data, the instrument used was validated by a validator. Then, a pilot study was conducted to test the validity of each item by using the Pearson Product-Moment Correlation. Additionally, to establish the trustworthiness of the interview data, there were some strategies that the researchers employed. The strategies were detailed descriptions of the context, consistent research procedures, overlap methods, and the researchers' neutrality. Additionally, the researchers employed peer debriefing and member checking (Miles et al., 2014;Mills & Gay, 2019).

Data Collection
The data were collected by administering a survey and an interview. Due to the Covid-19 pandemic, the researchers used Google Forms to collect the quantitative data. An online questionnaire link was distributed to the samples of the study through their English teachers. The interview was conducted via phone call and was recorded to avoid problems due to an internet connection that might occur while making a video call. Each student was interviewed by following the semi-structured interview procedure. Thus, the students were able to freely explain their perspectives without straying out of the topic.

Data Analysis
To analyze the data obtained from the questionnaires, the researchers employed both descriptive and inferential statistical analysis. In descriptive statistics, the researchers calculated and compared the items' mean scores and the standard deviations. Then, the inferential statistical analysis was calculated to find out the possible significant group differences in perceiving effective NNETs. Although the normality test showed that the data were not statistically normal, by accounting for the robustness of a parametric test (Norman, 2010), and considering the sample size, the sampling technique used in this study was MANOVA tests (a parametric test).
The interview data were analyzed through the following phases; data condensation, data display, and drawing and verifying conclusion (Miles et al., 2014). In this study, the researchers transcribed the interviews and then translated them into English. After that, the data were coded deductively into the preconceived themes: subject matter knowledge, pedagogical skills, and socio-affective skills. The data were displayed by providing the interview excerpts. Finally, the researchers drew and verified conclusions. The interpretation of the findings was obtained by integrating the results and meanings from both quantitative data and qualitative data.

RESULTS
The results combined the quantitative and the qualitative findings. The quantitative result presented the statistical analysis of the questionnaire data, and the result of MANOVA tests to show whether there are significant differences in the effective teacher dimensions by students' English proficiency levels and gender. In the qualitative findings, interview excerpts were provided to clearly illustrate the students' opinions. The names of the students are coded to ensure the students' privacy and confidentiality. Table 1 displays the data of the students involved in the qualitative phase.

Subject Matter Knowledge
This subsection presents the findings from the quantitative and qualitative phases of the study categorized within the teachers' subject matter knowledge. Table 2 presents a summary of the students' responses to the questionnaire within the subject matter knowledge category. From the eight qualities, three items were categorized into strongly agree; items number 3, 4, and 6, while the other five items were categorized into agree. The highest mean score within this category was for item number 3 'read English well' (M=4.3; SD=0.76). The second, third, and fourth highest mean scores were item number 6 'pronounce English well', item number 4 'write English well', and item number 7 'speak English well', respectively. The lowest mean score was item number 2 'effective NNETs know foreign culture well' (M=3.63; SD=0.81). Overall, the majority of the students agreed that the eight qualities were important for NNETs to be effective concerning teachers' subject matter knowledge. Moreover, the quality regarding effective NNETs should 'be fully conversant with English grammar' was also emerged. One student, Student4, reported that she liked a teacher who explains in detail. The comment in E2 illustrates her opinion: (E2) I like an English teacher who explains the lesson in detail. For example, we are learning about past tense, then, the teacher explains in detail about 'to be'. Not just saying that it happened in the past. (Student4, female-lower intermediate level) The comment above illustrates Student4's preferences about learning grammar and how she wanted an English teacher to explain the materials thoroughly.
Furthermore, another quality that had also emerged was insightful and resourceful. The student argued about the importance of the teacher's knowledge, not limited to the language but also things or issues that can broaden the students' knowledge.
(E3) I think that a good English teacher is more like…who uses some references or stories. A good teacher is a teacher who isn't strict about the books, and also has a broad insight. (Student1, male-upper intermediate, and above level) From the statement in E3, he believed that being resourceful and insightful were the characteristics of effective NNETs. This indicates that he needed a teacher who could broaden his knowledge, not limited to teaching him what is in the book.
Based on the above findings of quantitative and qualitative data, the characteristics of effective NNETs within the subject matter knowledge category encompassed being proficient in English, fully conversant with English grammar, and being insightful and resourceful.

Pedagogical Skills
This subsection presents the findings from the quantitative and qualitative phases of the study categorized within the teachers' pedagogical skills. Table 3 shows the results of the students' responses within the pedagogical skill category. From the eight qualities, only one item (item number 13) was categorized into neutral (M=2.88), while the other seven items were categorized into agree (M>3.41). The highest mean score in this category was item number 14 'teach English incorporating students' various learning styles' (M=4.2; SD=0.82), followed by item 9 'prepare the lesson well'. The lowest score was item number 13 'teach English in fully English' (M=2.88; SD=0.87).
It can be seen from item number 13 that only a minority of the respondents agreed to the statement (16.05% agree, and 4.21% strongly agree), while 46.84% chose neutral. These results suggested that in general, the respondents were neutral towards the use of English as the only medium of instruction in the classroom. They neither agreed nor disagreed. Moreover, the mean scores of the other seven items indicated that these qualities were important for effective NNETs albeit the mean scores are lower than those revealed in the subject matter knowledge category. The quality regarding providing opportunities to use English through meaningful activities was also confirmed, in which one student said: (E7) I think the efforts were not effective because first, the teacher taught okay, she taught too strict to the theories, she did not teach us to speak, she asked us to make five sentences of present perfect tense/past tense. We should not learn the structures too much; we have to know the structures but not memorize them. (Student1, male-upper intermediate, and above level) The claim in E7 indicates that the teacher might not have provided the students with a number of communicative activities. Student1 judged the activities provided by the teacher did not meet his goal to improve his communication skills.
The quality regarding the teachers' medium of instruction that should be fully in English obtained the lowest score compared to the other twenty-five qualities. Most of the students answered neutral for this quality. In relation to this quality, E8 depicts a student's reason why she chose neutral: In her view, it was not necessary for the teacher to use English all the time in the classroom. She thought about her and her friends' limitation of vocabulary so that it would be better if the teacher mixed the language; English and Indonesia when explaining the lesson.
The quantitative and qualitative data revealed that within the pedagogical skill category, effective NNETs were perceived to have qualities such as considering students' various learning styles in teaching, preparing lessons well, providing activities that arouse students' interests to learn the language, and providing a number of meaningful/communicative activities.

Socio-affective Skills
This subsection presents the findings from the quantitative and qualitative phases categorized within the teachers' socio-affective skills. Table 4 shows the results of effective NNETs characteristics viewed by students within the socio-affective skill category. From the nine qualities, only one item (item number 17) was categorized into agree, while the other seven items were categorized into strongly agree. The highest mean score in this category was item number 23 'do not discriminate between students and treats them fairly' (M=4.66; SD=0.61). Meanwhile, the lowest score was item number 17 'is helpful to the students inside and outside the classroom' (M=4.01; SD=0.78). Overall, these results suggested that the qualities within the socio-affective skill category were the most important characteristics, compared to those within the subject matter knowledge category and pedagogical skill category. Over half of respondents strongly agreed with statements number 19, 21, 23, and 24. In summary, these results indicate that teachers' socio-affective skills (personalities and relationships with the students) were highly essential for the students. From the interviews, most of the students described effective NNETs by mostly concerned with the teachers' personalities and attitudes toward the students. One student reported that effective NNETs should treat the students fairly regardless of their ability in English. E9 illustrates her view about teachers' fairness: (E9) In the learning process, there are students with high ability and low ability, so the teacher has to guide the students until their (ability) improved. (Student6, female-beginner level) From the statement in E9, it is clear that she wanted an English teacher who could demonstrate fair treatment to all students, whether the students had high ability level or low ability level. NNETs were expected to guide the students until the intended outcomes were achieved. Some students considered the characteristic of friendly as the quality that could help their learning. When asked about his portrayal of effective NNETs, Student3 said: (E10) In my opinion, effective teachers are friendly…because not all students can learn English easily, then, maybe, if the teacher is friendly, the students must be happy and get closed with the teacher. It will be great. (Student3, male-intermediate level) All six respondents demonstrated the importance of teacher-student relationships to help their learning. One of them said: (E11) First of all, in my opinion, the personality of an English teacher is different from a math teacher.
We learn how to communicate, meaning that the teacher should be close to the students, must have a close relationship, more like a friend but still have (certain) boundaries. (Student1, maleupper intermediate, and above level) Student1 thought that the distinctive characteristic of NNETs was the teachers' attitude to the students. It was important because to learn a language, the teachers should use the language for communicating with the students. To do so, the teachers were expected to be like a friend figure to increase the students' willingness to communicate with the teacher.
In conclusion, both quantitative and qualitative data show that within the socioaffective skill category, the qualities of effective NNETs which were considered to be highly effective for the students were fairness, friendliness, and closeness to the students. Furthermore, effective NNETs should be able to arise the students' motivation, listen to the students, and increase the students' confidence.

Significant Differences by Gender
This subsection provides the results of students' differences by gender based on the quantitative and qualitative data.  Table 5 shows the result of the MANOVA test of effective NNETs dimensions by gender. Using Wilks's statistics, there was no significant difference between public senior high school students in Tanjungpinang in perceiving effective NNETs by gender (V= 0.98, F = 2.39b, p > 0.05).
In the qualitative phase, although in general, various perspectives were reported by the six students, in which there were some similarities between male and female students. All students mentioned some qualities related to socio-affective skills regardless of their gender such as 'close with students' and 'friendly'. Moreover, some of the male and female students shared the same opinions about their preference for English teachers' pedagogical skills. Qualities such as 'varying teaching methods' and 'face-to-face learning' were reported by the male and female students. Table 6 shows the similarities of the data from the six students based on their gender.

Significant Differences by English Proficiency Level
This section provides the results of students' differences by English proficiency level based on the quantitative and qualitative data. Table 7 presents the result of the MANOVA test of effective NNETs dimensions by English proficiency level. Using Wilks's statistics, there was a significant difference among public senior high school students in Tanjungpinang in perceiving an effective English teacher by English proficiency levels (V= 0.95, F= 3.17, p < 0.05).
The result of separate univariate ANOVAs on the outcome variables revealed statistically significant differences of subject matter knowledge category (F= 4.02, p < 0.05), and pedagogical skill category (F= 2.76, p < 0.05) among students in perceiving effective NNETs by English proficiency level. The result from post hoc Tukey analysis also shows that beginner level students reported a lower score in the subject matter knowledge category (M= 31.71, SD= 5.50) compared to those at intermediate level (M= 33.93,SD= 4.21) and those at upper-intermediate and above level (M= 35.83,SD= 3.30). There was no other significant difference found in the students' scores in the subject matter knowledge category by English proficiency level. Additionally, the result from post hoc LSD analysis shows that upperintermediate and above level students reported a higher score in the pedagogical skill category (M= 32.42,SD= 3.68) compared to those at the beginner level (M= 29.44,SD= 4.03) and those at the post beginner level (M= 29.89,SD= 3.71). Moreover, intermediate level students reported a higher score in the pedagogical skill category (M= 32.42,SD= 3.68) compared to those at the beginner level (M= 29.44,SD= 4.03). There was no other significant difference found in the students' scores in the pedagogical skill category by English proficiency level. However, there was no statistically significant difference in the socio-affective skill category (F = 1.48, p > 0.05) among public senior high school students in Tanjungpinang in perceiving effective NNETs by English proficiency level.
In the qualitative phase, the findings show that a significant difference among the students by their English proficiency levels was not fully confirmed. However, students from lower intermediate level and upper-intermediate and above level reported qualities within the subject matter knowledge as the characteristic of effective NNETs as more important compared to those at the lower level. In addition, within the pedagogical skill category, a variety of perspectives were expressed by the six students in the interview. Furthermore, within the socio-affective skill category, some students shared similar perceptions about the characteristics of effective NNETs, such as being close and friendly. Table 8 shows the comparison of the data from the six students based on their English proficiency level.  learning. Yet, Student6, beginner level, reported that there were too many rules in the textbook. She thought that the teacher should selectively choose materials to be delivered to the students. In this case, students in the lower level might think that the content in the textbook was too much and the teacher needed to simplify the lessons.

DISCUSSION
This study investigated public senior high school students' perceptions of effective NNETs in Tanjungpinang, Indonesia. The quantitative and qualitative data suggest the importance of teachers' socio-affective skills compared to those concerning teachers' subject matter knowledge and pedagogical skills. In contrast, Park and Lee (2006) revealed that the students in their study ranked the pedagogical skill category as the highest and the socio-affective skill as the lowest. This could be explained since, within the Indonesian context, the social and affective qualities of human beings are very important. In fact, the current curriculum implemented in Indonesia also had an emphasis on a character-based curriculum (Ministry of National Education and Culture, 2018). This curriculum has primarily emphasized characters values. This suggests that students should have good characters and acquire certain competences afterward, which means that affective qualities come first in Indonesia. This concept could also be applied to a teacher, in which most students will enjoy learning with a teacher who possesses certain good qualities. Therefore, teachers' socio-affective skills were important to help the students' learning processes. This is in line with Stronge (2007) who states that teachers' affective characteristics are often a primary concern to the students. This result has also confirmed the empirical evidence from previous studies (Barnes & Lock, 2010Shishavan & Sadeghi, 2009). Lock (2010, 2013) also believed that teacher rapport is essential, particularly to build respectful and understanding atmospheres in EFL classes. Similarly, Silviyanti et al. (2021) revealed that native English-speaking teachers (NESTs) shared similar expectations with students about a strong teacher-student emotional relationship as a handful aspect in EFL classes. It provides encouragement to students, helps students like the class, and makes students learn better.
Some qualities of effective NNETs discovered in this study include subject matter knowledge (teacher's reading and speaking skills, sound knowledge of grammar), pedagogical skills (incorporating students' various learning styles, preparing lessons well, stimulating students' interests, providing meaningful activities), and socio-affective skills (fair, friendly, motivational, and close with students). These findings provide particular characteristics within the English language teaching field and explain more about the teacher competences that have been stipulated in the act of Republic of Indonesia number 14/2005 (President of Republic of Indonesia, 2005) and MNEC regulation number 16/2007 (Ministry of National Education andCulture, 2007). In regards to the teacher-student relationship, Borg (2006) states the relationship between teacher and students is more prominent in language teaching since there is more communication between teacher and students. During class, students often have to talk about their experiences and their life. As a result, effective NNETs need to engage students in this kind of meaningful communication many times. Therefore, in the English language learning context, the quality regarding teacher-student relationships is very distinctive compared to other subjects.
Furthermore, the findings of the current study also revealed that there was no significant difference between public senior high school students in Tanjungpinang in perceiving effective NNETs by gender. This finding did not support Park and Lee (2006) who discovered a significant difference between the males and the females. However, the finding of this study was consistent with that of Mudra's (2018) who also conducted a study in Indonesia but on the tertiary level. He discovered that there was no significant difference between male and female EFL learners in determining the qualities of effective English lecturers.
In relation to the students' English proficiency levels, quantitatively, it was discovered that there was a significant difference among public senior high school students in Tanjungpinang in perceiving effective NNETs. Likewise, Park and Lee (2006) discovered a significant mean difference between the high achieving students and the low achieving students. Furthermore, based on the qualitative data, it was found that the students with 'lower' levels seemed to be less concerned about the teacher's subject matter knowledge as the characteristic of effective NNETs. They were more concerned with qualities related to socio-affective skills and pedagogical skills. The most likely quality that explained the differences among students with different English proficiency levels was demonstrated by one respondent who reported that there were too many rules in the textbook. In this case, students in the lower level might think that the content in the textbook was too much and the teachers need to simplify the lessons. On the other hand, Student4, the lower intermediate level student, and Student1, upper-intermediate and above-level student reported qualities such as having good speaking skills, being resourceful and insightful, having sound knowledge of grammar are important to help their language learning. Therefore, NNETs should selectively choose materials to be delivered to students based on their English level (Harmer, 2007).

CONCLUSION
This study offers insights from students' perspectives regarding the most effective qualities of NNETs that can assist students' learning process in the Indonesian EFL context. This study reveals the importance of teachers' socio-affective skills in students' learning process. Some qualities of effective NNETs are being fair, being friendly, arousing students' motivation, providing activities that engage students' interest to learn the language, and providing meaningful activities, speaking English well, and being fully conversant with English grammar. As replication of Park and Lee's (2006) work, the generalizability of their findings cannot be extended to the context of this present study. The difference between the results of this study and Park and Lee's (2006) study provides further evidence that different perspectives in perceiving the most valuable characteristics of NNETs exist across cohorts of students, countries, and time. Furthermore, this study also confirms that students perceived different needs regarding the most effective NNETs' qualities in order to help their learning. Thus, NNETs need to explore and meet students' needs so as to enable them to be more effective in managing their instructions.
Furthermore, this finding raises the need to consider students' needs analysis in order to evaluate and modify the teaching processes as mandated in the current curriculum implemented in Indonesia. It is also important to prepare teachers with knowledge and skills that are highly effective to facilitate students' language learning process. This study suggests future research on the characteristics of effective NNETs in various settings to establish a greater understanding of how to develop EFL teaching practices that incorporate relevant aspects of language teaching. This study has three limitations. First, since the data was collected from only public senior high schools in Tanjungpinang, meanwhile the perspectives of students from private schools were not included. Second, the findings of this study were obtained only from students' perspectives. Hence, readers' discretion is expected to judge which needs are feasibly accommodated into English language teaching. Third, only some qualities identified in the quantitative data were confirmed in the qualitative findings. It is expected that these limitations can be covered by future related studies on this topic.