The Inclusion and exclusion criteria eligible for this study were constructed based on the research questions guiding the present study. Numerous research has been published about self- efficacy (SE) since 1960, including systematic reviews, literature reviews, and meta-analysis. In this regard, the researcher intended to come up with the latest excellent articles published in the last six years, because there are a few research papers on self-efficacy in the field of education. By determining the specific keyword through abstract reading, the papers were narrowed. First, the articles were classified by years, and 61 articles on the self-efficacy of teachers were received. Then, through selected studies that exclusively pertain to the self- efficacy beliefs of in-service EFL teachers during the last six years, the articles collected have been refined. The abstracts were reexamined again until the number of articles for in-service EFL teachers was reduced to 31.

This part is a discussion of articles related to the self-efficacy beliefs of in-service EFL teachers. As it is shown in Appendix 1, the study of articles covered authors, titles, years, context, participants, data collection tools, and findings for each article. 31 self-efficacy (SE) articles were identified including in-service EFL TSE beliefs. These articles cover the last six years. As seen in Figure 1, the study included different countries around the world including Turkey, Iran, Oman, Vietnam, Indonesia, Japan, Israel, and Pakistan participating in 2.959 in-service EFL teachers (See Figure 2). This shows the relevance of the self-efficacy of teachers in the field of education. Rawahi et al. (2019) concluded that the relationship between the academic outcomes of students and self-efficacy is significantly good. High SE affects learners' motivation, students 'achievement, and teachers' teaching practices. It also affects teachers’ job satisfaction and burnout levels, as well as their psychological wellbeing (Alibakhshi et al., 2020). Sabet et al. (2018) also indicate that teachers with high self-efficacy motivate their students more successfully and enhance their cognitive growth. However, those with a weak perception of efficacy prefer a “custodial orientation that relies heavily on negative sanctions to get students to study” (Bandura, 1994 as cited in Mojavezi & Tamiz, 2012, p. 489). Language teacher education can have a positive influence on language teacher self-efficacy beliefs, in both pre- and in-service contexts. Regarding this matter, Wyatt (2016) indicates how Omani English

teachers, reflecting and trying, developed practical knowledge and long-term LTSE convictions via focusing on self-oriented action research activities. Wyatt and Dikilitaş (2016) state that engaging in teachers' research as a continuous profession allowed Turkish English teachers to gain higher efficiency concerning their actual tasks.

Figure 1
Distribution of studies according to the countries between years 2015 and 2021

Figure 2
Distribution of the participants based on Countries

Two-thirds of all the studies (21/31) conducted between the years 2015-2021 are quantitative methods. Sabet et al. (2018) explored the relationship between Iranian EFL Teachers' Self- efficacy, their Personality, and Students' Motivation and they found that, unlike less efficacious teachers, highly efficacious teachers are good at motivating their students and improving their cognitive growth. Rawahi et al. (2019) found similar findings. There are other correlational studies including self-efficacy and job satisfaction, burnout, empowerment, etc. in our sample ( Azizifar et al., 2020; Babaei & Abednia, 2016; Fuchs et al., 2021; İpek et al., 2018; Marashi & Azizi-Nassab, 2018; Ortaçtepe & Akyel, 2015; Özkara, 2019; Ravandpour, 2019; Sabet et al., 2018; Safari et al., 2020;). As seen in Figure 2 above, the distribution of participants’ numbers also supports that most of the studies were conducted via correlational studies or used quantitative methods. When we analyze these studies we can conclude that they have a reductionist point of view. Their basic aim is to generalize so there is a linear type of interaction. As suggested by complexity theory the effect of dependencies among components cannot be fully represented by traditional approaches based on statistics. Those components are variables and they are dependent upon each other, so they cannot be fully represented by traditional methods like correlation because the system is changing. The growth of self-efficacy cannot be predicted. The relationships between the cause and effects do not easily determine the dynamic changes of teachers’ self-efficacy beliefs.

Figure 3
Distribution of the methods used in the articles

In the 31 articles reviewed, it is also seen in Figure 3 above that only eight articles adopted a qualitative study, and two conducted a mixed-method study, which consists of EFL teachers’ self-efficacy beliefs. However, a mixed-methods study conducted by Lailiyah and Cahyono (2017) did not truly address the challenges of directly eliciting the self-efficacy of language teachers (LTSE), since qualitative research techniques were mostly applied to extract background or contextual information for other purposes and support the data collected from the questionnaires. This study examined the self-efficacy of technology integration (SETI) of Indonesian EFL teachers and their usage of technology for teaching EFL. The study looked for a linear relationship when we looked at the findings. Data analysis shows that the EFL teachers' SETI is linked with their usage of technology in teaching EFL. Therefore, we could not accept this study as a mixed-method study. It is quantitative indeed. I would like to criticize it in this way.

However, Indahyanti (2016) used interviews as a data collection instrument in his research to explore Indonesian EFL teachers’ successful teaching performances and their beliefs in self-efficacy. The researcher maintained that he explored this influence in a classroom setting and discovered that experienced teachers have a higher perception of self-efficacy beliefs than novice teachers. We can conclude that generalizable results were found like in most other qualitative studies (Batool & Shah, 2018; Indahyanti, 2016; Listiani et al., 2019) in our study sample. Zonoubi et al. (2017) discovered self-efficacy to be a dynamic structure and to be nurtured via teacher reflection and collaboration opportunities like those provided by professional learning communities (PLCs). As Stoll and Louis (2007) describe, PLCs allude to collaborative activities that enhance "teachers critically interrogating their practice in ongoing, reflective, and collaborative ways" (p. 2) to support student learning. It can be claimed that the researchers were aware that self-efficacy is a complex structure, but their research method was not suitable to gauge this dynamism.

In contrast, some other researchers have succeeded in centralizing observations on various aspects of their research. In the framework of interviews, reflective assignments, etc. Phan and Locke (2015), Phan (2016), Wyatt (2016), and Wyatt and Dikilitaş (2016) attempted to incorporate and contextualize the knowledge they learned from them. The number one principle in designing research in a complex system perspective is including context as a part of the system under investigation, in which the researcher should consider the contextual factors and variables as a part of the research to get a holistic understanding of the issue in demand. Including the context means including any variable that has an impact on the issue under investigation. Although this research did not follow the principles of complexity perspective, there was a further step to be ecologically valid.

This section examines the number of articles that are often used for 31 articles chosen from 2015 to 2021, depending on theories and data collection tools. Based on the articles evaluated, most of them employed the theory of Bandura (1982) and Tschannen-Moran and Woolfolk (2001) in their articles or the context of the theory as well as to discuss their findings. Every article summarizes the need for self-efficacy (SE) of a teacher in detail, and its significance is also highlighted in depth. Bandura (1982) states that the SE of a teacher influences organization, the teacher's academic appointment in classrooms, and the way a teacher evaluates a student based on his/her abilities.

The notion of a new theory of self-efficacy based on Bandura's study was developed by Tschannen-Moran et al. (1998). Four things are said to impact teachers’ self-efficacy beliefs. Self-efficacy is defined by Tschannen-Moran et al. (1998) as a natural cycle of excellent efficacy, leading to scaffold and continuity to improved performance and returning to excellent efficacy. It is discovered that basically, all the papers debated in the last six years formed the basis of these theories.

The data collection tools used in 31 papers between 2015 and 2021 revealed that some tools were used extensively. For studies relating to in-service EFL teachers’ self-efficacy, The Teacher Sense of Efficacy Scale (TSES) (Tschannen-Moran & Woolfolk Hoy, 2001) was used for articles In2, In5, In10, In18, In 20, In,22, In23, In28, and In31. Meanwhile, an adapted or modified version of the Teacher Sense of Efficacy Scale (TSES) (Tschannen-Moran & Woolfolk Hoy, 2001) was used in articles In3, In8, In11, and In25. On the other hand, Self- Efficacy Survey (Praver, 2014) was used in article In17 to gauge the Turkish EFL teachers’ self-efficacy and burnout levels. The use of these tools in studies helps researchers analyze the amount and relevance of teachers' self-efficacy when used in classroom teaching and learning.

We can conclude that these mentioned researches above started with a claim and some assumptions, and the researchers tested them. So, in terms of research design, the contemporary approaches are mostly large-scale and cross-sectional studies. It includes a large group of participants because the aim is to make generalizations. However, as Wyatt (2016) mentioned, self-efficacy is a complex, dynamic and evolving construct. In this regard, the whole system and the pattern of its growth are studied, thus a transition from the objective of a causal explanation to a more descriptive, exploratory retrodiction is needed. We have to make comments on what we observed. Self-efficacy is needed to minimize the stress and concerns of teachers in carrying out their assigned tasks. To achieve this, we need to know the causal mechanisms of their self-efficacy levels.

From 2015 to 2021, this section will discuss frequent research variables in 31 papers. There are several self-efficacy variables studied in the 31 research articles over six years from 2015 to 2021. The self-efficacy is investigated to see whether there is a relationship with these variables or not. Variables selected by researchers for these groups of teachers include age (Özkara, 2019), teacher motivational strategies, teachers’ general beliefs, student’s motivation and achievement, classroom management, teaching practices, ( Al Rawahi et al., 2019; Alibakhshi et al., 2020; and Sabet et al., 2018), cultural context (Barabadi et al., 2018; Phan, 2016; Phan & Locke, 2016), various instructions, and professional development programs (Lailiyah & Cahyono, 2017; Ravandpour, 2019; Wyatt & Dikilitaş, 2016), job satisfaction ( Safari et al., 2020), burn-out (Fathi, & Saeedian, 2020; Roohani & Iravani, 2020), psychological wellbeing (Fathi et al., 2020), teachers’ empowerment (Azizifar, et al., 2020), self-regulation (Noughabi & Amirian, 2020) and online teaching Lee & Ogawa, 2021). The self-efficiency of teachers depends heavily on these variables since they help increase teachers' confidence in their capacity and competence to carry out tasks without giving up (Bandura, 1997). Studies focused on these variables should thus be pursued with a view to the theory of complexity to receive positive feedback from teachers and to ensure a high level of self-efficacy.