Purpose: This study aimed to examine the effects of Cambodia's New Generation Schools (NGS), as compared to their traditional counterparts, in enhancing the seven constructs--(1) science and math achievement, (2) science and math self-efficacy, (3) science and math outcome expectations, (4) attitudes toward science, (5) interactive science and math lessons, (6) support from science and math teachers and (7) encouragement and support in science from family. Design/methodology/approach: The two observations data was collected using self-rated questionnaire from 301 11th graders from five upper secondary schools located in three provinces of Cambodia. Independent sample "t"-test, Analysis of Covariance (ANCOVA) and dependent sample "t"-test were used to analyze the repeated measures data. Findings: The study revealed that students from the NGS exhibited statistically higher scores for most constructs, especially for attitudes toward science. However, when controlling for the differences in the first observation data, the significant effects of the three constructs have been neutralized. Moreover, though there was an increase in science activities outside school, there was a negative trajectory in the other two sub-constructs of attitudes toward science and support from science and math teachers, both in NGS and traditional schools. Practical implications: These findings point some practical implications for enhancing the effectiveness of the two school types and further research. Originality/value: There is heavy investment in new form of schools across the world to enhance students' learning and academic achievement in science and math in K-12 and to promote their interest in STEM in higher education. However, there is little document on the effectiveness of this new form of school, particularly in the Cambodian context.