Reforming Curriculum in A Centralized System: An Examination of the Relationships between Teacher Implementation of Student-Centered Pedagogy and High Stakes Teacher Evaluation Policies in China

Mei Luo

Reforming Curriculum in A Centralized System: An Examination of the Relationships between Teacher Implementation of Student-Centered Pedagogy and High Stakes Teacher Evaluation Policies in China
Luo, Mei
Thesis Advisor(s):
Chatterji, Madhabi
Education Leadership
Permanent URL:
Ph.D., Columbia University.
Past research has shown poor implementation levels of a classroom-level curriculum strategy that forms the centerpiece of China's new national education reform policies, namely, Student-Centered Pedagogy (SCP). This dissertation set out to investigate the influences of selected school and teacher background characteristics, classroom-level SCP implementation variables, and a high stakes teacher evaluation policy variable, on SCP implementation levels reported by high school teachers in a selected school district in China. The overall aim of the research was to study relationships among a number of factors hypothesized to affect teachers' SCP implementation levels, guided by a theoretically-grounded, conceptual framework. The study particularly examined the potential adverse influence of an output-driven teacher evaluation policy on SCP implementation levels. The teacher evaluation policy is tied to secondary school students' performance on the high stakes, national college entrance examination in China, the Gaokao. Eight contextual and reform-related factors derived from a review of literature were tied together in the conceptual framework suggesting direct, mediating and moderating influences on SCP implementation. Based on the framework, paths by which the variables could affect SCP implementation levels directly or indirectly, were tested in stages. Data were collected and analyzed using survey research methodology. The first part of the analyses involved the design and validation of a bilingual teacher survey (English and Chinese), tapping the key variables. The second part of the analyses involved a series of hierarchical regression models to test hypothesized pathways and relationships among the measured variables. The theoretical premise of the study was that the large size and highly centralized structure of the Chinese educational system led it to adopt an output-control mechanism in the form of the high-stakes teacher evaluation policy tied to student performance on Gaokao. The adoption of such an output-control mechanism resulted in a mismatch between the philosophy underlying the newer SCP reforms and the pre-existing teacher evaluation policy, which in turn led to poor implementation levels of SCP in classrooms by teachers. Previously, researchers in China have overlooked the importance of policy incompatibility issues in examining effects of reforms at the classroom level. The study found that, consistent with the literature, teacher beliefs in SCP and teacher self-efficacy in practicing SCP had consistently positive, statistically significant influences on SCP implementation (for Beliefs in SCP, t(224)=3.745, p=.000, standardized β=. 22; for Self Efficacy, t(224)=3.387, p=.001, standardized β=.23). Also consistent with expectations, the influence of the survey measure tapping perceived control by the output-driven teacher evaluation policy on SCP implementation, was negative and statistically significant ( t(224)= -1.982, p=.049; standardized β=-.12). Perceived support for SCP implementation, including resources, professional development programs, support from principals and colleagues was a statistically significant predictor in initial models, but the factor was found to lose statistical significance when combined with the variable tapping perceived control by the output-driven teacher evaluation policy. With all the specified independent and mediating variables in the regression model, the cumulative variance explained on SCP Implementation levels was 20% ( = .199). The overall model was statistically significant (F[7,224)=7.935, p=.000). Together, these results confirmed the main hypotheses of the study.
Education policy
Educational leadership
Item views:

In Partnership with the Center for Digital Research and Scholarship at Columbia University Libraries/Information Services.