ABSTRACT This research aimed at describing the effect sizes (ESs) of instructional strategies on student achievement in physics viewed from kinds of instructional strategies, levels of education, topics of physics, and kinds of student achievements. The phases in this research included: study acquisition for studies which have been carried out in Indonesia and been published from 2005 to 2014; study selection based on the statistics necessary for meta-analysis; and conducting the analyses. This research included 23 primary studies that have generated 29 sub-studies as the research subjects. The analysis revealed: (a) the mean of ESs on the topics of electricity, mechanics, and measurement were in big category, while on the topics of wave, modern physics, earth and space, and thermodynamics the mean of ESs were in moderate category, and (b) the mean of ESs on cognitive achievements were in big category, while on affective achievements were in small category. The major implication of this research is that we have generated empirical evidence supporting the effectiveness of non-conventional instructional strategies in physics with a big impact on grand mean of the ES that was found to be 1.020. Therefore, this research study suggests that the seven physics topics listed in this meta-analysis will be best delivered by the type of instructional strategies that characterized as student-led activities. Teachers’ pedagogical skill must enable them to apply this characteristic of instructional strategies. Furthermore, it is highly recommended to researchers to do more research study on affective and psychomotor achievement in physics so that these types of achievements can be included in future meta-analysis more effectively. Keywords: physics, meta-analysis, student achievement, effect size