The effect of teacher self-efficacy on learning engagement of secondary school students

Abstract

Less engagement in learning among students’ becomes an obstacle to their academic journey, which teacher self-efficacy can influence. This study examined the level and linkages between students’ engagement and teacher self-efficacy using data from 714 selected secondary school teachers in the Amhara region of Ethiopia. Descriptive and correlational survey designs were employed to determine the status and associations among study variables. CFA was used to ensure the validity of the instruments: student engagement and teacher self-efficacy scales. The data was analyzed using mean, standard deviations, one sample t-test, Pearson correlation, and multiple regressions. The study revealed that students’ behavioral and cognitive engagement were significantly below average, while teachers had significantly above-average self-efficacy in engaging students, managing classrooms, and using various instructional strategies. Nonetheless, a change in student engagement was significantly associated with teacher self-efficacy. The regression analysis disclosed that teacher self-efficacy accounted for 19.4% of the variance in student engagement. The study’s conclusions have implications for teacher efficacy, which is important in enhancing student engagement. Besides, diverse context-specific studies are required to ensure the stability of the study findings.

Keywords:

• Student engagement
• teacher self-efficacy
• behavioral engagement
• emotional engagement
• cognitive engagement

REVIEWING EDITOR:

• Lawrence Jun Zhang, Curriculum and Pedagogy, University of Auckland Faculty of Education and Social Work, Auckland, New Zealand

SUBJECTS:

• Educational Psychology
• Secondary Education
• Teachers and Teacher Education
• Classroom Practice

1. Introduction

The problem of student engagement in their learning activities is a global concern since empirical evidence links it with positive learning outcomes (Fredricks, 2014; Lei et al., 2018). Engaged learners are involved mentally, behaviorally, and emotionally in their education to be more likely to attend class, stay in school, and achieve better. However, student disengagement is a nationwide problem teachers face in their daily classes, and evidence shows that as students’ progress from elementary to high school, their engagement level decreases (Fredricks, 2014; Martin & Torres, 2016).

In this regard, to improve students’ academic outcomes, the government of Ethiopia has introduced and implemented various reforms. However, there are still challenges in the education system, such as many less competent and less engaged pupils, which concerns the public and researchers. The newly endorsed education and training policy stated in its introduction section that the quality of education in Ethiopia is deteriorating, as evidenced by students’ low academic outcomes. Moreover, the policy noted that most pupils’ academic results at each level of education are far below the policy’s desired level, i.e., below 50% (Ministry of Education [MoE], 2023). Likewise, the Ministry desk review report revealed that secondary school students’ engagement in academic work was low, as measured by motivation to learn, interest in academic activities, reading, and attendance (MoE, 2018). The report also showed high dropout rates challenge the country’s education system. Hence, improving student engagement might be needed to ameliorate the decline in academic achievement and minimize dropout rates because empirical evidence has revealed that engagement is tied to improved learning and academic success (Wang et al., 2011).

Student engagement in their learning activities is one of the predictors of positive learning outcomes (Lei et al., 2018). As empirical evidence suggests, a higher level of student engagement is associated with positive learning outcomes such as high academic achievement, satisfaction, and determination, a low dropout rate, and social engagement among students (Ali, & Hassan, 2018; Archambault et al., 2009; Walker & Greene, 2009). Put another way, the more actively students engage or invest time and energy in educationally advantageous activities, the more likely they are to learn, stick to their studies, and attain their academic goals (Kuh, 2009). Students get more out of their learning when they put more effort and time into studying, interacting with diverse peers and teachers, and engaging in relevant educational practices (Lawson & Lawson, 2013). On the other hand, a lack of student engagement in their learning leads students to quit school and is associated with disruptive behavior, alienation, and declining achievement (Appleton et al., 2008; Marks, 2000).

Engagement is a construct that is difficult to conceptualize and involves multifaceted concepts such as students’ behavioral, emotional, and cognitive involvement in educational activities that bring desirable outcomes (Ben-Eliyahu et al., 2018; Fredricks et al., 2004; 2016). Behavioral engagement includes active participation in class and extra-curricular activities, regular attendance at or in class, and a good attitude toward learning activities (Fredricks et al., 2016). Cognitive engagement involves students’ time and effort in academic tasks or dispositions toward schoolwork, applying metacognitive skills learning strategies, and valuing learning (Lawson & Lawson, 2013). Emotional engagement comprises students’ affective/emotional attachment and responses to the school and educational activities, such as psychological attachment or feelings towards their school and learning, as well as interactions with peers and teachers (Appleton et al., 2006; Lawson & Lawson, 2013).

In empirical definitions, phrases like student participation, academic involvement, student involvement, and involvement in school assignments have been employed to denote engagement (Ali & Hassan, 2018). In this regard, Astin (1999, p. 518) defines student involvement as “the amount of physical and psychological energy that the student devotes to the academic experience.” As to Astin, a student who is highly engaged spends a lot of time at school, studies a lot, is actively involved in school activities, and frequently relates with teachers and other students. From this conceptualization, it can be inferred that engagement involves student effort, time, emotional attachment to their education, and frequent interaction with teachers and other students (Kuh, 2009).

Engagement first emerged to understand school dropouts (Fredricks et al., 2004), and later, as a construct, it has attracted researchers’ attention since the mid-1990s due to its positive association with student intellectual, emotional, behavioral, social, and physical development in the learning process (Coates, 2005). In this regard, Fredricks et al. (2016) identified the main reasons for the rise in popularity of student engagement in research as engagement being significant for student academic success, the multidimensionality of the construct, its linkage with dropout prevention, and the malleability of the construct to changes in interventions.

Several factors might contribute to students’ engagement. As authorities in the field suggested, three contextual factors might be the main contributors to the engagement of students in their learning activities: family, peers, and school factors (Audas &Willms, 2001; Sahil, 2010). Family support and involvement in children’s education are mainly related to their engagement in their education and academic progress (Leithwood & Jantzi, 1999). Peers also play a role in student engagement because adolescents share similar characteristics with teenagers (Poulin & Chan, 2010). School-related factors such as culture, community, curriculum, and co-curriculum are critical factors that contribute to student engagement (Bardin & Lewis, 2011). As a school-related factor, teachers can use various techniques to affect student engagement in their learning. Empirical evidence suggested that teachers’ support and encouragement were among the significant factors for students to actively engage in learning tasks (Sahil, 2010; Trowler, 2010). Besides, Thien and Chan (2020) found that teacher self-efficacy influences students’ engagement in learning. Moreover, Ali and Hassan (2018) elaborated that teacher planning and accomplishing relevant activities are indispensable in promoting and improving student engagement.

As Ethiopia considered the education system to transform the country into a middle-income economy by the year 2030, the newly endorsed education and training policy showed that the quality of education is worsening. Most students’ academic results at each level of education fall below the desired level. Likewise, students’ academic engagement levels also appeared to be poor. Similarly, there were frequent complaints by teachers, supervisors, principals, and families of secondary school students about less student engagement in their learning (less time and effort devoted to doing classwork, homework, and project work, as well as low learning motivation). However, scarce empirical evidence is available regarding the level of student behavioral, emotional, and cognitive engagement and their determinant factors in Ethiopia.

Teachers’ belief in their competence influences students’ outcomes positively (Zee & Koomen, 2016). Teachers’ motivational beliefs influence their professional decision-making, teaching practices, efforts, and ways of accomplishing tasks, and approaches to teaching (Lauermann & Berger, 2021; Shahzad & Naureen, 2017). Teacher self-efficacy, or beliefs about one’s abilities, is related to better student engagement (Altun, 2017). Teacher self-efficacy is a teacher’s belief about their teaching capabilities, such as using different teaching strategies and managing and engaging students in learning activities (Tschannen-Moran & McMaster, 2009). It is a multidimensional construct that includes efficacy in instruction (capacity to use various instructional methods), classroom management (belief in maintaining classroom order), and student engagement (belief in motivating and engaging students in learning) (Liu et al., 2020).

Zee and Koomens’ (2016) meta-analysis of forty years of studies showed that teacher self-efficacy was one of the determinant variables consistently associated with student engagement. It might be because more efficacious teachers will be more likely to plan appropriate activities, persist with learners experiencing difficulties, and use appropriate teaching materials (Erawan, 2010). The level of self-efficacy is more likely to impact teachers’ work persistence, dedication to careers, and instructional decisions. Likewise, teachers with high efficacy tend to demonstrate high levels of planning and organization and spend more time monitoring and checking teaching tasks (Shaukat & Iqbal, 2012). Moreover, teachers with high self-efficacy have confidence in their knowledge and expect high performance from their pupils (Stronge, 2018), which helps them believe in their ability to make a difference in students’ learning. If teachers believe they can affect their students’ learning, they stipulate higher learning outcomes and exert more effort to achieve them to influence students’ engagement in learning (Thien & Chan, 2020).

Teacher self-efficacy affects teachers’ decisions and practices related to instruction, which can impact students’ engagement in learning activities. Classroom instruction is under the control of teachers; hence, it holds promise as a way to engage students better (Turner et al., 2014). In their review of previous research, Zee and Koomen (2016) revealed a positive link between teacher self-efficacy and teacher classroom practices, which affect the quality of the classroom environment and student performance. However, more research focused on the relationship between teacher efficacy and other teacher-related factors such as job satisfaction, instructional practices, and job stress (Klassen et al., 2011; Zee & Koomen, 2016). Some Western-based studies have examined the relationship between teachers’ efficacy and students’ engagement in learning activities. Although self-efficacy and student engagement vary due to variations in institutional characteristics and contexts, empirical evidence in non-Western collectivist samples about the constructs appears limited (Klassen et al., 2011). Likewise, student engagement is vital for positive pupil outcomes; however, little is known about it and its linkages with teacher self-efficacy (Lu & Mustafa, 2021) and in the current study context.

Therefore, the study’s primary purpose was to investigate the effect of teacher self-efficacy on the learning engagement of secondary school students in Amhara National Regional State, Ethiopia. We hypothesized that high levels of efficacy beliefs (efficacy for student engagement, efficacy for classroom management, and efficacy for instructional strategies) impact teacher instructional decisions and practices, in turn influencing students’ behavioral, cognitive, and emotional engagements. Hence, teachers’ self-efficacy beliefs are more likely to be associated with student engagement.

Hence, the study raised the following questions: (a) What is the level of student learning engagement (behavioral engagement, emotional engagement, and cognitive engagement) in secondary schools of the Amhara region? (b) How confident are teachers in engaging students in learning, managing classrooms, and using different instructional strategies? (c) Is there any significant relationship between student engagement and teachers’ self-efficacy? And (d) Does teacher self-efficacy significantly influences students’ engagement in learning activities?

2. Theoretical background

2.1. Student engagement

The theoretical foundation of student engagement is associated with Astin’s (1999) theory of involvement and Tinto’s (1975) interactionalist theory of student departure. The theory of involvement focuses on the amount of physical and psychological energy students dedicate to their academic work (Astin, 1999). On the other hand, the interactionalist theory of student departure posits the social and academic integration of students with the school communities and learning activities, which increases student retention and commitment (Harper & Quaye, 2009).

As a theoretical model, engagement first emerged to understand school dropouts, and later, as a construct, it received scholarly attention due to its pivotal role in positive school and student outcomes (Fredricks et al. 2004). Since then, it has been widely studied and linked to positive outcomes for student success and development, including satisfaction, persistence, academic achievement, social engagement, and low dropout rates (Ali, & Hassan, 2018; Archambault et al., 2009; Kuh, 2009; Walker & Greene, 2009). Fredricks et al. (2016) mentioned four significant reasons for the increased popularity of engagement in research. First, student engagement is pivotal to learning and academic success, such as higher grades, achievement test scores, and school completion rates. Second, it has appeal because it is a “meta-construct” that includes observable behaviors, internal cognitions, and emotions. Third, student disengagement is the biggest challenge many teachers face in their classrooms. Finally, evidence shows that student engagement is adaptable and responsive to teacher and school activities changes (Fredricks et al., 2016).

Although there has been much disagreement regarding how engagement has been defined and studied, there is some consensus in the literature that it is a multidimensional construct that involves behavioral, emotional, and cognitive engagement (Ben-Eliyahu et al., 2018; Fredricks et al., 2004; Fredricks et al., 2016). On the other hand, Appleton et al. (2006) viewed engagement as a construct comprised of four subtypes: (1) academic (time on task, credits earned toward graduation, and homework completion); (2) behavioral (attendance, suspensions, voluntary classroom, and extra-curricular participation); (3) cognitive (self-regulation, relevance of school work to future endeavors, value of learning, and personal goals and autonomy); and (4) psychological (belonging, and relationships with teachers and peers Alternatively, Finn (1989) conceptualized the word engagement as a construct comprised of behavioral (participation in class and school) and affective or emotional components (school identification, belonging, and valuing learning). In conclusion, despite the various classifications, it is reasonable to infer that engagement is a multidimensional and elusive concept that needs more studies to understand the construct and its antecedents better. According to Trowler (2010), engagement involves various variables; using one variable to study it is flawed. However, Kuh (2009) conceptualized student engagement as the time on task, and the quality of effort students devote to productive learning activities associated with desired student outcomes.

Empirical evidence indicates that student engagement decreases as students’ progress from elementary to middle school and reaches its lowest levels in high school (Martin & Torres, 2016). Marks (2000), for example, estimated that 40 to 60 percent of youth are disengaged in secondary schools. Moreover, as students’ progress through high school, their levels of engagement in educationally purposeful activities decrease (Klem & Connell, 2004).

Studies showed that teachers’ support and encouragement were salient for students to actively engage in learning tasks (Sahil, 2010; Trowler, 2010). Besides, Thien and Chan (2020) found that teacher self-efficacy influences students’ engagement in learning. Moreover, Ali and Hassan (2018) corroborated that collaboration among school leaders, teachers, and parents is necessary for planning and accomplishing relevant activities to promote and improve the level of engagement among students. However, knowledge and empirical evidence about student engagement in the context of developing countries seem scant.

2.2. Teacher self-efficacy

The theoretical underpinning of self-efficacy is grounded in social cognitive theory, which highlights the capacity of human beings to exercise control over the nature and quality of their lives (Bandura, 1997; Bandura, 2006). According to social cognitive theory, individuals possess self-beliefs that allow them to control their thoughts, feelings, and actions (Bandura, 2006). Moreover, according to this theory, an individual’s behavior and actions are influenced by observation of others and interaction with their environment (Dimopoulou, 2014). Furthermore, the theory suggests four goal realization processes: self-observation, self-evaluation, self-reaction, and self-efficacy (Bandura, 1995).

Self-efficacy deals with the belief one has in their ability to perform or complete a specific task, or a course of action successfully (Bandura, 1997). Likewise, teacher self-efficacy reflects the extent to which a teacher’s belief about their capability of instructional practices, classroom management, and student engagement (Zee & Koomen, 2016). Efficacy can be seen in terms of collective efficacy or self-efficacy. Collective efficacy refers to the shared beliefs of the group members in their group’s capabilities to achieve desired outcomes (Katz-Navon & Erez, 2005). On the other hand, self-efficacy is an individual teacher’s perception of their capability to make a difference in student learning (Tschannen-Moran & Woolfolk Hoy, 2001). According to Bandura (1997), collective efficacy is an extension of self-efficacy, and its sources and processes partly overlap with those that influence self-efficacy. However, he added, collective efficacy is an emergent group-level property that reflects the group as a whole. On the other hand, the self-efficacy construct covers internal resources that people ascribe to themselves and plays a pivotal role in determining the quality of one’s performance. Katz-Navon and Erez found that self-efficacy emerged as a significant construct influencing individual performance in low-task interdependence situations.

Self-efficacy influences what a teacher chooses to do, how much effort they put into achieving objectives, and their persistence when their efforts fail to produce desired results (Bandura, 1997). Based on Bandura’s insight, one’s self-efficacy can be shaped by information collected from four sources: (1) performance or mastery, based on the teacher’s success and failure experiences; (2) vicarious experiences, by modeling others’ success or failure; (3) verbal or social persuasion, due to negative or positive feedback gain; and (4) physiological and emotional states, based on the interpretation of one’s physical and emotional reactions. Teachers typically overestimate or underestimate their capabilities due to such exposure, and these estimations may impact their efforts and the courses of action they choose to pursue (Oh, 2011).

Tschannen-Moran and Woolfolk Hoy (2001) categorized teacher self-efficacy into three dimensions: self-efficacy for instructional strategies (confidence in utilization of various instruction methods), self-efficacy for student engagement (beliefs in actively involving students in learning activities), and self-efficacy for classroom management (confidence in managing classroom order). Hence, self-efficacious teachers have confidence in their knowledge and skills of effective teaching behaviors concerning instructional strategies, classroom management, and student engagement (Liu et al., 2020).

Teachers’ belief in their capacity influences student learning and how teachers handle classroom things (Liu et al., 2020). Moreover, teachers’ efficacy beliefs impact how they feel, behave, think, and motivate themselves (Pierce, 2014). Likewise, social cognitive theory suggests that teachers’ perceptions of themselves affect their actions (Bandura, 1997). Moreover, teachers with confident self-efficacy are more likely to positively influence students’ behavior and learning outcomes (Halim & Ahmad, 2015). A study conducted by Holzberger et al. (2013) in German secondary schools found that teachers with higher self-efficacy beliefs had higher instructional quality of classroom management, cognitive activation, and individual learning support. Likewise, the instructional strategies used by efficacious teachers comprise open-ended inquiry methods, student-centered teaching strategies, critical thinking, analysis, and synthesis, and expect high performance from their pupils (Butucha, 2014; Stronge, 2018).

2.3. The effect of teacher self-efficacy on student engagement

Empirical evidence has shown that teachers’ self-efficacy beliefs significantly influence student achievement and behavior (Klassen et al., 2011). Teacher self-efficacy is reportedly linked with better student engagement, learning opportunities, and positive student outcomes (Gibbs & Powell, 2012). Efficacious teachers tend to demonstrate high levels of planning, organization, and passion for teaching and spend more time on teaching (Shaukat & Iqbal, 2012). Teachers with a higher sense of efficacy devote more time to students’ learning, supporting and reinforcing students’ intrinsic motivation (Bandura, 1997). Teacher efficacy is crucial for students to succeed academically, and student achievement is linked to student engagement (Shoulders & Krei, 2015).

More research was conducted focusing on the relationship of teacher efficacy with other teacher-related factors (such as job satisfaction and job stress) compared with the association between teacher efficacy and student outcomes (Klassen et al., 2011). However, teachers’ beliefs about their efficacy have been found to influence student learning directly. In this regard, Shoulders and Krei (2015) asserted that efficacious teachers develop multifaceted learning activities that foster students’ learning engagement. Moreover, instruction is under the control of teachers; hence, it holds promise as a way to engage students better (Turner et al., 2014). Likewise, while teachers have confidence in their teaching, they use innovative instructional practices. As to Shoulders and Krei, efficacious teachers are more likely to manage their classrooms effectively. Moreover, Havik and Westergård (2020) demonstrate that teachers who provide high-quality classroom interactions, such as caring and supporting behavior and a well-structured classroom, foster high student engagement.

After reviewing teacher efficacy studies conducted from 1998 to 2009, Klassen et al. (2011) revealed that most studies focused on Western (North America and Europe) samples. Specifically, research in self-efficacy theory is scarce in the African context (Kinde & Asfawossen, 2016). Exploring teacher self-efficacy and its linkage with student engagement in diverse settings is noteworthy because teacher preparation, practices, and teaching policies show considerable variation across countries, and variations in teaching environments may influence teachers’ beliefs about their duties and responsibilities.

3. Materials and methods

3.1. Research design

This study was conducted based on post-positivist philosophy, which is associated with a quantitative approach and makes claims of knowledge based on objectivity, standardization, deductive reasoning, and guides within the research process (Creswell & Plano Clark, 2017; Mertens, 2015). Hence, this philosophical assumption helped researchers formulate hypotheses, examine the relationship between study variables, and deduce conclusions. Accordingly, a descriptive and correlational survey design was used in the study. As a descriptive, it describes the students’ engagement status in their learning and teacher self-efficacy. As a correlation design, it examined the link between student engagement and teacher self-efficacy. Correlational design assisted in determining the direction and strength of relationships between variables (student engagement and teacher self-efficacy) under consideration (Creswell, 2015).

3.2. Population and sample

Regarding the study context, a secondary school in Ethiopia ranges from grades 9 to 12 and it is a foundation and preparation place for students to join higher education institutions or the world of work. On average, students aged 15 to 18 join these grade levels. First- and second-degree-graduated teachers teach students at the secondary school level. The conceptual framework of the study was presentented in Figure 1 below.

Figure 1. Hypothesized Conceptual Framework of the Study.

The population of this study was 39,247 teachers who were teaching in 652 secondary schools in Amhara National Regional State. In the region, there were 15 zones. Multi-stage sampling was used to choose respondents. In the first stage, out of 15 zones, five zones (East Gojjam, Awi, Bahir Dar City, South Gondar, and South Wollo) were selected randomly. In the second phase, 11 districts (Machakel, Enemay, Debre Markos Ketema, Guagsa Shikudad, Dangila Ketema, Bahir Dar City, Dera, Farta, Dessie Zuria, Kutaber, and Tehulederie) were choosen using proportional random sampling since the numbers of districts were not equal in each zone. Finally, sample respondents were selected from sample districts proportional to the available number of teachers (Gay et al., 2012). In the sample districts, there were 3753 teachers based on 2020/21 regional statistics. The Cochran (1977) sample size determination formula was employed to determine the sample size. The Cochran formula was employed in the study because of its appropriateness for large populations; it allowed researchers to calculate an ideal sample size given a desired level of precision and confidence interval; it reduced sample errors; and considered effect size. Accordingly, out of 3753 teachers, with 0.05 precision, and considering the number of stages the study went through and the ten percent non-returnable rate, the sample size for this study was 771 (Cohen et al., 2007). From these, 714 (92.6%) teachers with valid and completed data were used for further analysis.

3.3. Instrumentation

3.3.1. Student engagement scale

The student engagement scale was a self-developed tool based on the available literature. Twenty-seven items were developed: 13 for behavioral engagement, 4 for emotional engagement, and 10 for cognitive engagement. The items were subjected to exploratory and confirmatory analysis. Hence, five items were rejected due to inappropriately loading on components other than their respective. Twenty-two items (8 for behavioral, 4 for emotional, and 10 for cognitive engagement) were retained and used in this study. The scale was a 5-point Likert scale, where 1 represents “strongly disagree,” and 5 represents “strongly agree.” A pilot study used 92 respondents and found a .83, .87, and .91 alpha coefficient for behavioral, emotional, and cognitive engagement, respectively. It indicated that the items were reasonably reliable. The content validity of the instrument was checked to see if it could measure what it was supposed to measure (Collier, 2020). Twenty subject experts (psychology instructors) rated items as “essential,” “useful, but “not necessary” (Lawshe, 1975) and found .83, .87and .90 content validity ratios (CVR) for behavioral, emotional and cognitive engagement, respectively, indicating that items had good content validity to measure their respective construct. As per Lawshe’s suggestion, 50% and above agreement of experts offers assurance of content validity.

3.3.2. Teacher self-efficacy scale

The Tschannen-Moran and Woolfolk Hoy (2001) Teacher Sense of Efficacy Scale (TSES) was used. The scale is also known as the Ohio State Teacher Efficacy Scale. It has close congruence with Albert Bandura’s self-efficacy theory (Klassen et al., 2011) has empirical support for satisfactory construct validity and reliability across grades and several countries (Klassen et al., 2009) and is considered a better measure due to its factor structure stability (Woolfolk Hoy & Burke-Spero, 2005). Permission was obtained through email to use the scale. The instrument consisted of 12 items, 4 for each of the three sub-factors: self-efficacy for student engagement, classroom management, and instructional strategies. It is a 5-point Likert scale where 1 stands for “none at all,” 2 for “very little,” 3 for “some degree,” 4 for “quite a bit,” and 5 for “a great deal.” The alpha coefficient of internal consistency of the TSES was .90 (Tschannen-Moran & McMaster, 2009). Before using the scale in this study, it was piloted using 92 teachers and found .82, .81, and .84 alpha values of self-efficacy for student engagement, classroom management, and instructional strategies, respectively. Hence, the Cronbach alpha values of the pilot study exceeded the acceptable level value (.70), indicating reasonable reliability of the tool for the main study (Pallant, 2020).

3.4. Data analysis

The data collected were coded and subject to data cleaning and preliminary analysis (unengaged responses, missed values, multivariate outliers, and testing multivariate assumptions). Descriptive statistics were done for participant demographic characteristics. An exploratory factor analysis (EFA) was employed to examine whether indicators were measuring more than one construct (Collier, 2020). Confirmatory factor analysis (CFA) was used to verify the result of the EFA and test if the data supported the conceptual factor structures (Tabachnick & Fidell, 2019). Moreover, EFA and CFA were undertaken to ensure the convergent and discriminant validity of the instruments. In accepting the facture structures of AMOS (Analysis of Moment Structures) outputs, the following model fit indices and cutoff criteria were used: chi-squared ((χ2/df ≤ 5), the Comparative Fit Index (CFI ≥.90), the Tucker-Lewis Index (TLI ≥.90), the Root Mean Square Error of Approximation (RMSEA ≤.08), the Root-Mean-Square Residual (RMR ≤.08), and Standardized Root Mean Square Residual (SRMR ≤.08) (Collier, 2020; Hair et al., 2019; Tabachnick & Fidell, 2019). One sample t-test was computed to examine the status of student engagement and teacher self-efficacy. Pearson product-moment correlation coefficient was used to scrutinize the association between teacher self-efficacy and student engagement. Multiple regressions were employed to identify teacher self-efficacy’s contribution to student engagement.

3.4.1. Demographic information of respondents

The survey was distributed to 771 respondents. Of those, 736 (95.5%) questionnaires were returned. However, prior to data analysis, the completeness, outliers, and unengaged responses were checked. Accordingly, in the data screening process, 22 questionnaires were removed due to unengaged and incomplete responses (Collier, 2020). Hence, 714 participants’ responses were used for further analysis.

As presented in Table 1, the majority of teachers 527 (73.8%) were male, and 187 (26.2%) were female. Concerning educational qualification, 497(69.6%) teachers were first-degree holders, 207(29%) were second-degree graduates, and the rest, 10 (1.4%), were diploma graduates. Concerning professional experience, 321 (44.96%) teachers had 17 or more years of service, and 276 (38.6%) had 11–16 years of experience. 108 (15.13%) had 5 to 10 years of service, and the remaining 9 (1.3%) had served less than 5 years. The average length of service of respondents in the current job was 17.4 years.

Table 1. Respondents’ personal information.

Demographic variables		Frequency	Percent
Sex	Male	527	73.8
	Female	187	26.2
Qualification	Diploma	10	1.4
	First degree	497	69.6
	Second degree	207	29.0
Teaching Experience	Below 5 years	9	1.26
	5–10 years	108	15.13
	11–16 years	276	38.60
	17 and above years	321	44.96
	Total	714	100

3.5. Ethical considerations

The study was conducted after the authors’ institution granted ethical clearance. The respondents were told the purpose of the study and engaged in the study after their consent was obtained. Participants were also told that their participation was voluntary and that the information they provided was kept confidential. Personal identifiers were excluded from the questionnaire to ensure participants’ confidentiality.

4. Results

4.1. Validation of measurement tools

To ensure the constructs’ validity and assess indicators’ loading on a construct exploratory and confirmatory factor analysis (CFA) using AMOS 23 were employed for both student engagement and teacher self-efficacy scales.

4.1.1. Exploratory factor analyses for measurement tools

A total of 27 and 12 items were subjected to exploratory factor analysis for student engagement and teacher self-efficacy scales, respectively. Multiple criteria were used to determine the number of components to be retained: previous theory, eigenvalues greater than 1, the cumulative percentage of variance extracted, and the scree plot (Henson & Roberts, 2006; Tabachnick & Fidell, 2019). Accordingly, both scales satisfied those criteria and the results supported the factorability of the data well. Moreover, the results of the rotated dimension matrix revealed a three-factor solution for student engagement scale, which explained a total of 62.41% of the variance. Component 1 (cognitive engagement) explained 49.87% of the variance with an eigenvalue of 13.47, component 2 (behavioral engagement) contributed 6.76% with an eigenvalue of 1.83, and 1component 3 (emotional engagement) raised 5.78% of the variance with an eigenvalue of 1.56. Of the 27 items subjected to rotation, five items (EN9, EN10, EN11, EN12, and EN13) were rejected due to inappropriate loading on components other than their respective. Therefore, 22 items were loaded high on three dimensions (see items in Appendix I). Hence, student engagement in the Ethiopian secondary school context can be conceptualized in line with student behavioral engagement, emotional engagement, and cognitive engagement.

On the other hand, exploratory factor analysis of teachers ‘self-efficacy revealed a three-component solution with eigenvalues greater than one. The three components jointly accounted for about 72.05% of the total variance. Hence, component 1 (self-efficacy for student engagement) accounted for 49.98% of the variance with eigenvalue 5.99, component 2 (self-efficacy for classroom management) accounted for 11.99% of the variance with eigenvalue 1.44, component 3 (self-efficacy for instruction) accounted for 10.08% of the variance with eigenvalue 1.21. In general, except one item (item 8, which omitted in the CFA), the pattern and the standard estimates of factor loading provided satisfactory support for the previous structure of the Tschannen-Moran and Woolfolk Hoy (2001) teacher self-efficacy scale.

4.1.2. Confirmatory factor analysis for engagement scale

The exploratory factor analysis for the student engagement scale yielded a three-factor structure for the student engagement scale: student behavioral engagement, emotional engagement, and cognitive engagement. To verify the EFA results, a confirmatory analysis was conducted. The CFA outputs of the student engagement scale in Table 2 proved the EFA results that the scale retained the three-factor structure of the student engagement scale, where the student behavioral engagement sub-scale can be measured in eight items, the emotional engagement dimension can be assessed in four items, and the cognitive engagement sub-scale can be examined in ten items. Moreover, standardized co-efficient values ranged from 0.6 to 0.9, which suggests the items in the constructs can adequately measure what they are supposed to measure. In this regard, Collier (2020) asserted that the standard co-efficient of the item that exceeds 0.5 explains more than half of its respective construct.

Table 2. Results of confirmatory factor analysis for student engagement.

	Student behavioral engagement	R^2	β	CR
1	Students come to class completing their individual and group assignments	.52	.72
2	Students enthusiastically come class finishing home works	.66	.81	20.91
3	Students regularly attend classes in this school	.56	.75	19.25
4	I feel that students demand extra lessons outside the normal school timetable	.46	.68	17.39
5	Students invest most of their time in learning	.52	.72	18.48
6	I feel that students do more than required	.48	.69	17.82
7	Students actively participate in class room instruction	.55	.74	19.09
8	Students are actively involved in various co-curricular activities	.37	.60	15.50
	Student emotional engagement
9	In my class students follow their lesson enthusiastically	.81	.90
10	In my class, students feel happy	.68	.83	29.47
11	I feel that students enjoy when they learning new things	.69	.83	29.57
12	I feel that students enjoy when they do their class work	.69	.83	29.53
	Student cognitive engagement
13	I feel that students seem hopeful for their further education	.57	.76
14	I believe that students plan achieving better goals	.62	.79	22.01
15	I feel that students believe that education will create a better opportunity for them	.61	.78	21.70
16	I feel that students in this school consider education as an important tool achieving future goal	.68	.83	23.28
17	Students feel that learning in school is most important	.60	.78	21.61
18	Students feel that the school academic assessment results are based on their effort	.54	.73	20.28
19	When I do class and home works students check to examine whether they understand what I'm doing	.53	.72	20.00
20	I feel that students ask teachers if the lesson they are learning is unclear	.49	.70	19.14
21	I believe that students feel that they have achieved a good result by working hard	.48	.70	19.08
22	I feel that students strive for appropriate effort for the task given	.61	.78	21.80

Note: R² = squared standardized regression weight, β = standardized regression weight, C.R. = critical ratio, the model fit indices: χ2 = 746.022, df = 206, χ²/df = 3.621, CFI = 0.95, TLI = 0.94, RMSEA = 0.061 and RMR=. 0332.

4.1.2.1. Structural model and hypotheses testing of student engagement scale

The CFA model in Figure 2 showed the constructs’ standard estimates ranging from 0.6 to 0.81 for student behavioral engagement, 0.83 to 0.91 for student emotional engagement, and 0.69 to 0.83 for cognitive engagement dimensions. Results indicated that most standard estimates were approaching or exceeding 0.7, suggesting that the variance in the indicators explained half or more of the respective unobserved construct, and the model fitted the data (Collier, 2020).

Figure 2. CFA Model of Student Engagement with Standardized Estimates.

Confirmatory analysis results of student engagement scale yielded a good fit indices, with χ2 = 746.022, df = 206, χ²/df = 3.621, CFI = 0.95, TLI = 0.94, RMSEA = 0.061 and RMR=. 0332. Accordingly, the measurement model indices met the cutoff criteria and fit the data well. Therefore, it is inferred that with the elimination of some items, the hypothesized model of three factor structure of student engagement scale was found to be accepted. Moreover, the sample data reflected the model for the population and the structural model was well-fitted model (Collier, 2020; Hair et al., 2019; Tabachnick & Fidell, 2019).

4.1.3. Confirmatory factor analysis for teacher self-efficacy scale

To evaluate the validity and the measurement model of the teacher self-efficacy scale, CFA was performed. The CFA outputs in Table 3 verified the EFA results that the measurement scale maintained the three factor structure of teacher self-efficacy scale: efficacy for student engagement, efficacy for classroom management and efficacy for instructional strategies. The standardized co-efficient values ranged 0.586 to 0.9, which indicates items in the scale can adequately measure their respective construct (Collier, 2020).

Table 3. Confirmatory Factor Analysis results for Teacher Self-efficacy.

	Efficacy for Student Engagement	R^2	β	C.R
1	Craft good questions for your students	.71	.84
2	Motivate students who show low interest in school work	.75	.86	27.67
3	Get students to believe they can do well in school work	.74	.86	28.70
4	Help your students’ value learning	.69	.83	26.00
	Efficacy for Classroom Management
5	Control disruptive behavior in the classroom	.66	.81
6	Get children to follow classroom rules	.73	.85	24.11
7	Calm a student who is disruptive or noisy	.65	.81	22.96
	Efficacy for Instructional Strategies
9	Use a variety of assessment strategies	.63	.79
10	Provide an alternative explanation or example when students are confused?	.57	.76	18.11
11	Assist families in helping their children do well in school?	.30	.55	12.59
12	Implement alternative teaching strategies in your classroom	.44	.66	16.19

Note: R² = squared standardized regression weight, β = standardized regression weight, C.R. = critical ratio, the model fit indices: χ2 = 182.826, df = 38, χ²/df = 4.81, CFI = 0.97, TLI = 0.97, RMSEA = 0.073 and RMR=. 0335.

4.1.3.1. Structural model and hypotheses testing of teacher self-efficacy scale

The measurement model of teacher self-efficacy in Figure 3 indicated the standard estimates of each construct indicator, ranging from 0.83 to 0.86 for teacher self-efficacy for student engagement, 0.81 to 0.85 for teacher self-efficacy for classroom management, and 0.55 to 0.79 for teacher self-efficacy for instructional strategies. Results specified that most indicators described over half of the variance in their construct.

Figure 3. The measurement model for teacher self-efficacy with standardized estimates.

The model fit indices of teacher self-efficacy scale revealed reasonably good model fit, with χ2 = 182.829, df = 38, χ2/df = 4.8111, CFI = 0.968, TLI = 0.967, RMSEA = 0.073 and RMR= .0355 (Tabachnick & Fidell, 2019). Hu and Bentler (1999) suggested that the RMSEA values closer to 0.08 or below CFI and TLI values near .95 or above specify a good model fit. Accordingly, the results suggested that the hypothesized model of three factor structure of teacher self-efficacy scale was accepted. Moreover, the structural model was well-fitted model (Hair et al., 2019; Tabachnick & Fidell, 2019).

4.1.4. Construct reliability and validity analyses

As indicated in Table 4, the Cronbach’s alpha coefficients and the composite reliability values ranged .89 to .93, suggesting items in the constructs had good internal consistency and sufficiently measured they supposed to measure. Likewise, estimating convergent and discriminant validity in measuring the measurement scale provided credible evidence. Convergent validity assesses the interrelatedness of indicators in theoretically similar constructs, whereas discriminant validity estimates the uniqueness of indicators in different constructs (Collier, 2020). The CFA results of convergent and discriminant validity of the student engagement scale in Table 4 revealed that the average variance extracted (AVE) value of each dimension (EME=.716, BHE=.565, and CGE=.572) yielded above 0.5, indicating that indicators of theoretically similar constructs were interrelated or indicators load better in a single theoretically related construct and confirmed adequacy of convergent validity of the scale (Collier, 2020, Brown, 2015). Discriminant validity was employed to determine whether a construct was distinct from other constructs in the measurement scale. Hence, the result in Table 4 showed that the square root of the AVE for each scale dimension (EME=.716, BHE=.545, and CGE=.572) exceeded the maximum shared value ((EME=.531, BHE=.513, and CGE=.558); suggested that constructs were unrelated, contained no double-loading indicators, and showed good discriminant validity (Collier, 2020).

Table 4. Reliability, convergent and discriminant validity of student engagement dimensions.

Constructs	α	CR	AVE	MSV	MaxR(H)	EME	BHE	CGE
EME	.909	0.901	0.716	0.531	0.915	0.846
BHE	.890	0.894	0.565	0.513	0.889	0.654	0.751
CGE	.929	0.930	0.572	0.558	0.933	0.729	0.747	0.756

Notes: α = Cronbach’s alpha coefficient, CR = composite reliability, AVE = average variance extracted, MSV = maximum shared variance, MaxR(H)=maximum reliability, EME: Emotional Engagement; BHE: Behavioral Engagement; CGE: Cognitive Engagement.

As displayed in Table 5, the Cronbach’s alpha coefficients and the composite reliability values ranged .79 to .91, implying items had good internal consistency and sufficiently measured they were intended to measure. The AVE of teacher self-efficacy scale constructs were found to be 0.68, 0.72, and 0.51 for teacher self-efficacy for classroom management (ECM), teacher self-efficacy for student engagement (ESE), and teacher self-efficacy for instructional strategies (EIS). All the values of AVE exceeded the recommended value (0.5), confirming the adequacy of convergent validity (Collier, 2020).

Table 5. Reliability, convergent and discriminant validity of Teacher Self-efficacy Scale.

Constructs	Α	CR	AVE	MSV	MaxR(H)	1	2	3
ECM	0.864	0.864	0.679	0.397	0.866	0.824
ESE	0.909	0.912	0.722	0.397	0.913	0.913	0.850
EIS	0.804	0.787	0.510	0.392	0.810	0.810	0.626	0.714

α = Cronbach alpha coefficient, CR = composite reliability, AVE = average variance extracted, MSV = maximum shared variance, MaxR(H) = maximum reliability, ECM: efficacy for classroom management, ESE: efficacy for student engagement, EIS: efficacy for instructional strategies.

The results in Table 5 noted that all the AVE values of teacher self-efficacy constructs (ECM=.679, ESE=.722, and EIS=.510) exceed MSV coefficients ((ECM=.397, ESE=.397, and EIS=.392), suggested that each construct is unique from others and had no discriminant validity concern (Brown, 2015).

4.2. The status of student engagement

Table 6 provides student engagement subscales mean and standard deviation values. The highest mean value of all subscales was emotional engagement (M = 2.941, SD=.862). Student cognitive engagement mean value (M = 2.768, SD=.816) was the next. The lowest level was the student behavioral engagement mean (M = 2.594, SD=.794). As depicted in Table 6, the grand mean value (M = 2.768, SD = 0.722) indicated that student engagement was below the average (3:00).

Table 6. Mean and standard deviations of student engagement subscales.

Student engagement sub-scales	Mean	SD
Behavioral engagement	2.594	0.797
Emotional engagement	2.941	0.862
Cognitive engagement	2.768	0.816
Total	2.768	0.722

N = 714, SD = Standard deviations.

In Table 7, it was revealed that students’ behavioral engagement was significantly less than expected (t = −13.600, df = 713, P = .000). As Cohen’s d (Cohen 1998) suggestion indicates (d = .80, .50, and .20 for large, medium, and small effects, respectively), the result of the effect size (d =.5) is labeled moderate, implying that the obtained mean of behavioral engagement was moderately lower than expected. Similarly, students’ cognitive engagement was significantly below average (t = −7.601, df = 713, P = .000). Its effect size (d =.28) is a medium in Cohen’s d, indicating that student cognitive engagement was below expected. However, the study disclosed that students’ emotional engagement was average (t = −1.823, df = 713, P = .069). The effect size (d =.10) was considered small, supporting that students’ emotional engagement did not significantly differ from the average.

Table 7. One-sample test statistics of students engagement subscales.

	Test value = 3
Dimensions	t	df	Sig. (2-tailed)	Mean Difference	Effect size
Behavioral engagement	−13.600	713	.000	−.40578	.50
Emotional engagement	−1.823	713	.069	−.05882	.10
Cognitive engagement	−7.601	713	.000	−.23198	.28

4.3. The status of teacher self-efficacy

Teacher self-efficacy was measured based on an individual teacher’s self-reported survey. As demonstrated in Table 8, the mean scores of the teacher self-efficacy subscales showed that overall, teachers had more than average self-efficacy in student engagement, classroom management, and instructional strategies. To compare, the highest mean score was teacher efficacy for classroom management (M = 4.071, SD = 0.807); this indicates that teachers were confident enough in their classroom management competency. The lowest mean score was teacher self-efficacy for instructional strategies (M = 3.345, SD = 0.824), indicating that teachers were less confident in their ability to use different teaching strategies than their ability to manage students’ behavior.

Table 8. Mean and standard deviations of subscales of student engagement.

	Mean	SD
Efficacy for student engagement	3.565	0.857
Efficacy for classroom management	4.071	0.807
Efficacy for instructional strategies	3.345	0.824
Total	3.666	0.680

N = 714, SD = Standard deviations.

Table 9 provides one-sample t-test results for subscales of teacher self-efficacy. Results indicated that teachers had significantly above average self-efficacy in their efficacy for student engagement (t = 17.622, df = 713, P =.000), classroom management (t = 35.475, df = 713, P =.000), and instructional strategies (t = 11.191, df = 713, P =.000). The effect sizes of the variables (d =.66 for efficacy for student engagement, d = 1.32 for efficacy for classroom management, and d =.42 for efficacy for instructional strategies) showed teachers rated their sense of self-efficacy significantly higher than average, as Cohen’s effect size benchmark suggests.

Table 9. One-sample test statistics of teacher self-efficacy subscales.

	Test value = 3
Dimensions	T	df	Sig. (2-tailed)	Mean difference	Effect size
Efficacy for student engagement	17.622	713	.000	0.57	.66
Efficacy for classroom management	35.475	713	.000	1.07	1.32
Efficacy for instructional strategies	11.191	713	.000	0.35	.42

4.4. Relationship between student engagement and teacher self-efficacy

Pearson correlation was used to examine the relationship between student engagement and teacher self-efficacy and the results are illustrated in Table 10. Students’ behavioral engagement had significant weak positive relationships with teacher efficacy for student engagement (r =.294), efficacy for classroom management (r =.228), and efficacy for instructional strategies (r =.354). Compared with others, student behavioral engagement had the highest association with teachers’ confidence in utilizing various instructional techniques.

Table 10. Pearson’s correlation coefficients among variables (N = 714).

	ESE	ECM	EIS	BHE	EME	EIS
ESE	1
ECM	.565**	1
EIS	.517**	.466**	1
BHE	.294**	.228**	.354**	1
EME	.304**	.231**	.302**	.594**	1
CGE	.360**	.251**	.397**	.683**	.672**	1

** Correlation is significant at the 0.01 level (2-tailed).

Similarly, student emotional engagement had slight positive associations with teacher efficacy for student engagement (r =.304), efficacy for classroom management (r =.231), and efficacy for instructional strategies (r =.302). Likewise, it was found that there was a significant slight to moderate positive association between student cognitive engagement and teacher efficacy for student engagement (r =.360), efficacy for classroom management (r =.25′), and efficacy for instructional strategies (r =.397). These results indicated that while teachers have confidence in their competency in classroom management and instructional techniques, students’ engagement is more likely to be enhanced.

Results in Table 11 revealed that teacher self-efficacy accounted for 19.4% of the variance in student engagement (R² = 0.194). The Cohen’s f result is f²=.23, indicating a medium effect (In Cohen’s f 2 multiple regression benchmarks, 0.02, 0.15, and 0.35 are small, medium, and large effects, respectively). It implied that efficacious teachers significantly influence student engagement in their learning activities.

Table 11. Regression results model summary.

Model	R	R^2	Adjusted R^2	Std. error of the estimate	Cohen’s f (f^2)
1	.441^a	.194	.191	.649	0.23

The ANOVA summary of Table 12, the regression model overall results (F (3, 710) =57.113, p =.000) confirmed that teacher self-efficacy significantly predicts student engagement.

Table 12. ANOVA results.

Model		Sum of squares	df	Mean square	F	Sig
1	Regression	72.300	3	24.100	57.113	.000^b
	Residual	299.599	710	.422
	Total	371.899	713

Table 13 shows that teacher self-efficacy for instructional strategies was the main significant contributor (ß=.400, t = 11.65, P = .000) to student engagement. It raised the variance of student engagement by 16% (R²=.16). Teacher self-efficacy for student engagement accounted for 3.4% (R²=.034) of the variation in students’ engagement (ß=.215, t = 5.48, P = .000). Teacher self-efficacy for classroom management was found to have a non-significant effect on students’ engagement (ß=.022, t=.514, P = .607). Hence, as teacher self-efficacy for instructional strategies increases, student engagement more likely increases; as teacher self-efficacy for student engagement increases, so does student engagement.

Table 13. Coefficients of variables.

Model		Unstandardized coefficients		Standardized coefficients	t	Sig	Collinearity statistics
		B	Std. error	Beta			Tolerance	VIF
1	(Constant)	1.594	.104		15.371	.000
	EIS	.351	.030	.400	11.650	.000	.598	1.672
	ESE	.182	.033	.215	5.478	.000	.639	1.564
	ECM	.019	.038	.022	.514	.607	.688	1.453

Note: Dependent Variable: Student Engagement.

Predictors: ESE = Teacher Efficacy for Student Engagement, ECM = Efficacy for Classroom Management,.

EIS = Efficacy for Instructional Strategies.

5. Discussion

This study aimed to investigate the status of student engagement and teacher self-efficacy and determine the effect of teacher self-efficacy on student engagement. Prior to data analysis, instrument validation and preliminary analysis were employed. Thus, the reliability, convergent, and discriminant validity of the instruments were ensured by conducting EFA and CFA. The normality of the collected data and statistical assumptions were maintained through preliminary analysis. Finally, descriptive and inferential statistics were computed to answer the research questions.

Hence, the study revealed significantly low level of student behavioral, emotional, and cognitive engagement, suggesting that secondary school students in Ethiopia were reluctant to participate in classroom activities, involve themselves in co-curricular activities, establish good relationships with others, feel connected to their schools, and give value to learning. These findings are similar to the Ethiopian Ministry of Education’s desk review report that secondary school students had less learning motivation and interest in engaging in academic activities (MoE, 2018). Consistently, previous evidence showed that as students’ progress through high school, their levels of engagement in learning activities tend to decrease (Klem & Connell, 2004; Marks, 2000; Martin & Torres, 2016).

The mean student engagement scores in the current study vary between 2.59 and 2.94 on a five-point Likert scale. To compare, student behavioral engagement had the lowest mean score (M = 2.59), and emotional engagement was the highest (M = 2.94). This result should be used cautiously because behavioral engagement exhibits relatively observable acts like attention, participation, and effort in academic activities, which teachers could easily rate in a survey. On the other hand, students’ emotional engagement comprises affective deeds such as feelings towards school, teachers, peers, and learning, which look difficult for teachers to rate. In this regard, Reschly and Christenson (2012) sort out behavioral engagement as observable, which teachers can easily track in the classroom, and affective engagement as internal or affection deeds, which teachers face difficulty in reporting it. Likewise, in reviewing empirical evidence published between 2003 and 2015, Lei et al. (2018) found substantial variations in self-reported or other-reported results of student engagement (mainly for behavioral and emotional engagement).

The study also revealed that teachers had above-average self-efficacy for student engagement, classroom management, and instructional strategies. The mean teacher self-efficacy score was between 3.35 and 4.07 on a five-point Likert scale. These values showed that teachers perceived confidence in using various instructional methods, maintaining classroom order, and motivating and engaging students in educational tasks (Liu et al., 2020). The results of the study are, to some extent, in line with a study done in primary schools by Hiruy et al. (2021). Their study revealed that teachers’ self-efficacy scores, except community involvement efficacy, ranged from 3.38 to 4.09. However, the study setting was secondary schools since teacher efficacy beliefs may vary due to differences in school contexts and teacher experiences (Martin & Sass, 2010). In this regard, Hiruy et al.’s study revealed that contextual factors such as educational qualification levels affected teachers’ self-efficacy beliefs, where diploma-qualified teachers (who comprised 66% of their study) had a high sense of efficacy compared with first-degree holders. However, in our study, only 1.4% (n = 10) of the participants were diploma holders; the majority, 98.6% (704), had a first degree or higher qualification.

Correlation results revealed that higher teacher self-efficacy beliefs were significantly and positively related to student engagement. Teacher self-efficacy (efficacy for student engagement, classroom management, and instructional strategies) had a slight (r =.228) to moderate (r =.397) significant positive association with student engagement dimensions (behavioral, emotional, and cognitive engagement). Relatively, the highest significant positive relationship was found between student cognitive engagement and teacher efficacy for instructional strategies (r =.397 or R²=.16). According to Gay et al.’s (2012) correlation coefficient categorization (between 0.35 and 0.65), the strength of the relationship between variables is classified as a moderately positive relationship. These findings suggest that the more teachers are confident in their ability to engage students, employ various instructional strategies, and manage classrooms, the more students are likely to be engaged behaviorally, emotionally, and cognitively in their learning. Likewise, this might be the reason that efficacious teachers provide high-quality classroom interactions, such as by developing multifaceted learning activities, managing their classrooms effectively, and utilizing various assessment techniques that foster students’ learning engagement (Havik & Westergård, 2020; Shoulders & Krei, 2015). This indicates that when teachers are highly efficacious, they can improve their instructional skills and practices, and in turn, their students are found to have a high level of engagement in their learning activities.

The study results suggested that 19.4% of the variation in student engagement was explained by teacher self-efficacy. Moreover, teacher self-efficacy for instructional strategies and teacher self-efficacy for student engagement accounted for 16% and 3.4% of the variation in student engagement, respectively. When teachers have a high sense of belief in their abilities with various instructional strategies and the importance of students’ involvement, they are more likely to devote more time and effort to student learning and make resources available, which increase the likelihood that students engage in practical educational activities.

However, in this study, teacher self-efficacy for classroom management was found to have a non-significant effect on students’ engagement. Similarly, Martin and Sass (2010) found that the relationship between teachers’ perceived ability to manage classroom instruction and student engagement effectively was relatively small. They claimed that teachers with high efficacy favor a less directive or humanistic approach to managing student behavior and their instruction.

The finding of this study is consistent with a previous meta-analysis of forty years of studies by Zee and Koomen (2016), who identified that teacher self-efficacy, had a significant positive link with student engagement. The more teachers become efficacious in their competency of student engagement and instructional techniques; the more likely students will be engaged in their learning activities. Erawan (2010) contended that efficacious teachers plan appropriate activities and suitable teaching materials. Similarly, highly efficacious teachers develop good instructional plans, spend more time, and exert more effort checking their teaching tasks, which helps them make a difference in students’ engagement (Shaukat & Iqbal, 2012; Thien & Chan, 2020). It supports the notion that teacher self-efficacy beliefs influence their classroom practices, which in turn influence student engagement because teachers with high efficacy engage in professional learning and try out and implement different new instructional methods (Zee & Koomen, 2016).

6. Conclusion

To conclude, from the study findings, it was inferred that students exhibited below-average behavioral, cognitive, and emotional engagement in their learning activities. However, teachers had above-average self-efficacy beliefs. It was also concluded that teacher self-efficacy beliefs had a significant positive association with student engagement. Moreover, the study deduced that overall teacher self-efficacy had a significant positive influence on student engagement. Specifically, high levels of teacher self-efficacy beliefs for student engagement and instructional strategies had a significant positive effect on student engagement. This indicated that the more teachers are efficacious in their abilities in terms of student involvement in instruction and utilization of various instructional strategies, the more students are engaged behaviorally, emotionally, and cognitively in their learning. On the other hand, it was inferred that teacher efficacy in classroom management had a non-significant effect on students’ engagement. This finding may support the notion that efficacious teachers have a less directing approach to classroom management, which makes students reluctant to their learning.

The study has implication that teachers’ high sense of beliefs in their capabilities in instructional strategies and the merits of students’ involvement significantly influence students’ in-class and out-of-class engagement toward educationally effective activities. Specifically, teachers who have high confidence in their ability to encourage student participation in instruction can enhance student engagement in educational activities. Similarly, confident teachers in their ability to utilize various strategies in their instruction are more likely to improve their students’ engagement in learning activities. Hence, given its importance to student engagement, school leaders and experts should focus on improving teachers’ sense of efficacy.

The study has limitations, as do other research works. First, teacher self-efficacy data were collected through a self-reported survey, which may have a social desirability bias or over-reporting of their efficacy level. On the other hand, the data about student engagement were rated by teachers, who might face difficulty in scoring the affective deeds of students’ emotional engagement. The current study was investigated on the effect of one variable (self-efficacy) on student engagement using correlation and regression analysis. Other determinant variables that might have impact on student engagement could exist. The scope of current study was secondary schools, which limits its generalizability. Taking into account the limitations of the current study, the following directions can be made for future research work: The first suggestion is that future research should involve diversified samples, such as students and school leaders, to ensure the comparison of results among different populations. Likewise, to better understand the determinant factors of student engagement, future studies should be conducted incorporating more variables such as family support, students’ demography, and teacher classroom practices and employing sophisticated statistical analysis such as structural equation modeling. In addition, to ensure the consistency of the current study findings, similar studies should be replicated in other educational contexts (such as kindergarten, primary, and tertiary levels) in future research endeavors.

Author contributions

EE: the corresponding author ensured the quality of the data and worked on the collection, analysis, and interpretation of the data.

MG: ensured the data quality, supervised the study, proofread and edited the final paper for publication.

Acknowledgments

The authors would like to express their truthful thanks to the study participants for their consent and dedication of their time in providing data.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The raw data supporting the conclusion of this article will be made available by the authors without undue reservation.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Notes on contributors

Eshetu Kibret Emiru

Eshetu Kibret Emiru, MA, is a PhD student in Education Policy and Leadership at Bahir Dar University. He has been involved in higher education teaching, research, and community service. He has served as a teacher and educator for the last nine years. His research interests involve policy, leadership, teaching and learning, etc.

Mateb Tafere Gedifew

Mateb Tafere Gedifew is an associate professor of education policy and leadership at Bahir Dar University. He has published several articles in the areas of educational leadership, policy, and higher education. He has been teaching courses at the Masters and PhD levels. He also served as dean of the college, executive director and vice president of Bahir Dar University, head of the Amhara Regional State Leadership Academy, and education bureau. His major areas of research comprise higher education, instructional leadership, policy, organizational development, etc.

Appendix I.

Questionnaires

Sample questionnaire for teachers

Part I: Background Information

1. Sex: - Male______ Female____

2. Level of Education:-

   A. MA/MSc _____B. BA/BSc_____C. Diploma/level_____D. other please specify_______

3. Year of experience in teaching

   A. 5 years & below B. 6-10 years C. 11-15 years D. 16-20 years E. 21 years & above

Part II Student Engagement Scale

The following statements are items intended to measure student engagement in learning activities. Please indicate your agreement by putting a (√) mark in one of the five alternatives.

Table

No	Items	Level of agreement
		Strongly agree	Agree	Agree to some extent	Disagree	Strongly disagree
	Student Behavioral Engagement
EN1	Students come to class completing their individual and group assignments
EN2	Students enthusiastically come class finishing home works
EN3	Students regularly attend classes in this school
EN4	I feel that students demand extra lessons outside the normal school timetable
EN5	Students invest most of their time in learning
EN6	I feel that students do more than required
EN7	Students actively participate in class room instruction
EN8	Students are actively involved in various co-curricular activities
EN9	I feel that students respect teachers
EN10	Students show good manners in class
EN11	I believe students have a good relationship with teachers
EN12	I consider students adhere to the rules of the school
EN13	In my class students follow their lesson enthusiastically
	Student Emotional Engagement
EN14	In my class students follow their lesson enthusiastically
EN15	In my class, students feel happy
EN16	I feel that students enjoy when they learning new things
EN17	I feel that students enjoy when they do their class work
	Student Cognitive Engagement
EN18	I feel that students seem hopeful for their further education
EN19	I believe that students plan achieving better goals
EN20	I feel that students believe that education will create a better opportunity for them
EN21	I feel that students in this school consider education as an important tool achieving future goal
EN22	Students feel that learning in school is most important
EN23	Students feel that the school academic assessment results are based on their effort
EN24	When I do class and home works students check to examine whether they understand what I'm doing
EN25	I feel that students ask teachers if the lesson they are learning is unclear
EN26	I believe that students feel that they have achieved a good result by working hard
EN27	I feel that students strive for appropriate effort for the task given

Part III Teacher Self-Efficacy Scale

The following statements are items intended to measure teacher self-efficacy beliefs. Please indicate your agreement by putting a (√) mark in one of the five alternatives.

Table

No	Items	Level of agreement
		A great deal	Quite bit	Some degree	Very little	Note at all
	Efficacy for Student Engagement
EF1	Craft good questions for your students
EF2	Motivate students who show low interest in school work
EF3	Get students to believe they can do well in school work
EF4	Help your students’ value learning
	Efficacy for Classroom Management
EF5	Control disruptive behavior in the classroom
EF6	Get children to follow classroom rules
EF7	Calm a student who is disruptive or noisy
EF8	Establish a classroom management system with each group ofstudents
	Efficacy for Instructional Strategies
EF9	Use a variety of assessment strategies
EF10	Provide an alternative explanation or example when students are confused?
EF11	Assist families in helping their children do well in school?
EF12	Implement alternative teaching strategies in your classroom

Source of teacher self-efficacy scale: Tschannen-Moran and Woolfolk Hoy (2001). Teacher efficacy: Capturing an elusive construct.