校長教學領導是發展優質教育和使學校成功的關鍵因素,校長若能營造可使教師成功合作的工作環境並藉以促進教師協作,則此支持性的學校環境可提高教師自我效能並能提升教學成效。本研究旨在探討學校層次之校長教學領導與學校教師協作,以及教師層次之教師協作,對於教師自我效能的影響,除瞭解校長教學領導與教師協作對教師自我效能的直接效果之外,更進一步分析教師協作的跨層次中介效果。由於不同教育階段(國小、國中和高中)的學校文化與學生特性並不相同,因此本研究將檢視教育階段在校長教學領導與教師協作之 間的調節效果。本研究使用2018年教學與學習國際調查(TALIS)之臺灣地區553所國小、國中與高中校長及10,129位教師資料進行多層次中介效果模式分析,其中,學校教師協作同時作為學校層次的脈絡變數與教師層次解釋變數,利用分效模式來檢視不同層次變數對教師自我 效能的中介影響。研究結果顯示,教師層次的教師協作對教師自我效能具有顯著的正向影響,學校層次的學校教師協作對教師自我效能有顯著的脈絡效果。其次,校長教學領導會透過教師協作的完全中介對教師自我效能產生間接影響。第三,國小和國中教育階段在校長教學領導與學校教師協作之間具有正向調節效果。最後,本研究依據研究結果進行理論與實務的討論,並提出未來研究之建議。 Principal instructional leadership and teacher collaboration are key characteristics of effective schools (Scheerens, 1990; Teddlie & Reynolds, 2000) and the primary factors that influence teacher self-efficacy. On the basis of the definition provided by the Organisation for Economic Co-operation and Development (OECD, 2018), the present study defined principal instructional leadership as the process of supporting and developing high-quality instructional practices, developing and implementing policies that support student achievements, developing learning communities, providing feedback on instructions, modeling effective instructions, and supporting the use of assessment data. Teacher collaboration refers to the process through which teachers exchange information with their colleagues and learn from each other (Geijsel et al., 2009). The concept of self-efficacy is based on social cognitive theory proposed by Bandura (1977). In the present study, teacher self-efficacy was defined as the competence of teachers to effectively address problems in their classrooms and schools and improve the quality of their teaching, which could eventually improve the learning outcomes of their students (Ainley & Carstens, 2018; Battersby & Cave, 2014; Geijsel et al., 2009; Tschannen-Moran & Hoy, 2001). Purpose Principal instructional leadership is crucial for high-quality education and schools’ success. Principals are responsible for shaping the work environments of teachers; thus, they can strengthen and promote teacher collaboration. A supportive school environment can improve teachers’ self-efficacy. The aim of this was to elucidate the direct and indirect — through teacher collaboration — effects of principal instructional leadership on teacher self-efficacy. Because teaching is profoundly influenced by school culture and environmental factors, work pressure at different education levels may create different work environments and exert different effects on principal instructional leadership and teacher collaboration. Therefore, we performed additional analyses regarding the contextual moderating effects of education levels (International Standard Classification of Education, ISCED). Design/Methodology/Approach This study performed a secondary data analysis by using the OECD Teaching and Learning International Survey (2018) dataset. The sample included principals and teachers from elementary (200 and 3,494, respectively), junior high (202 and 3,835, respectively), and high (151 and 2,800, respectively) schools in Taiwan. Multilevel linear modeling was performed to investigate relationships between principal instructional leadership, teacher collaboration, and teacher self-efficacy in these schools. Teacher collaboration at the individual level and principal instructional leadership at the overall level were used as independent variables to evaluate the effects of both levels on teacher self-efficacy. Findings/Results 1. Teacher Collaboration Exerts a Direct and Contextual Effect on Teacher Self-efficacy The analysis results revealed a positive relationship between individual- and school-level teacher collaboration and individual-level teacher self-efficacy, indicating that a higher level of teacher collaboration is related to higher perceptions of self-efficacy among teachers; these results supported the study assumption. These two variables were related at both the individual and school levels, which suggests that the use of both individual- and school-level variables is a precise approach for evaluating collaboration. 2. Principal Instructional Leadership Exerts as a Significant Effect (p < .05) on Teacher Collaboration Principal instructional leadership had a positive and significant direct effect on teacher collaboration. This demonstrates that principal instructional leadership encourages teachers to be productive in their classes, share their ideas with others, and create a cooperative culture. 3. Teacher Collaboration Significantly Mediates the Correlation Between Principal Instructional Leadership and Teacher Self-efficacy Principal instructional leadership exerted a significant indirect effect on teacher self-efficacy through teacher collaboration. Thus, principal instructional leadership may play a key role in teacher collaboration and indirectly enhance teachers’ self-efficacy through teacher collaboration. 4. Moderating Effect of Education Level on the Correlation Between Principal Instructional Leadership and Teacher Collaboration Education level (ISCED) positively moderated the relationship between principal instructional leadership and teacher collaboration. The relationship between principal instructional leadership and teacher collaboration was found to be significant in elementary and junior high schools but nonsignificant in high schools. Originality/Significance Studies have reported principal instructional leadership as an imperative predictor of teacher self-efficacy. However, few studies have focused on the indirect effects of mediating variables, such as teacher collaboration, on the relationship between principal instructional leadership and teacher self-efficacy. Accordingly, this study comprehensively analyzed various school and teacher factors by using a single model and evaluated the effects of teacher collaboration on the relationship between principal instructional leadership and teacher self-efficacy from a multilevel perspective. Principal instructional leadership was found to be associated with teacher self-efficacy, both directly and indirectly, through teacher collaboration. Suggestions/Implications 1. Implications for Educational Practice On the basis of the findings, this study proposes some recommendations for strengthening teacher collaboration, promoting teachers’ professional development, and creating a supportive work environment to enhance teacher self-efficacy. Principal instructional leadership is crucial;nevertheless, because principals in Taiwan are administrators and not teachers, they rarely have time to invest in the professional development of teachers. Furthermore, to engage more strongly in instructional and staff development, principals must possess relevant skills. Policymakers should develop training and professional development programs for principals that are centered on the aforementioned topics. 2. Methodological Implications Preacher et al. (2010) indicated that in a mediation model for 2-1-1 data, when the b effect (the effect of a mediator on a dependent variable) conflates the within and between effects, the indirect effect that necessarily operates between two groups is confounded by the within-group portion of the conflated b effect. This study compared the analysis results of the 2-1-1 unconflated and conflated models to determine the difference between the individual and overall mediating effects of teacher collaboration. The study findings support those of Preacher et al. (2010) in that the between and within effects of a mediator must be separated when a multilevel mediation model is used.