Factors Impacting Mathematics Achievement In Shanghai And Peru:A Multilevel Analysis Based On PISA Dataset

The multilevel analysis started with the null model where none of the level-1 or level-2 explanatory variable was included.Results from this model suggest that considerable differences in total average school mathematics performance existed across the two countries.Similarly,the intra-class correlation(ICC)was found homogeneous for the two countries.The student background model was developed to examine the extent to which preschool education attendance for one year or less,preschool education attendance for more than one year,gender difference and student's family economic,social and cultural status in Peru and Shanghai correlated with math attainment.The results suggested that this model worked differently across countries.Specifically,in Peru,all of the student background variables were statistically significant predictors of math performance.Meanwhile,the relationship between math performance and preschool education attendance for one year or less seemed to vary significantly across schools in Shanghai.The school background model was developed to determine whether mathematics teacher-student ratio and quality of school educational resources are related to students'math performance in Peru and Shanghai.In this study,math teacher-student ratio was negatively related to math achievement in Shanghai;it implies that a one unit increase in math teacher-student ratio will reduce student performance across schools.It should be recalled that math teacher-student ratio appeared to vary significantly across schools in Peru.The different relationship pattern between each undertaken student-and-school level variables and mathematics achievement were identified.The cultural differences in the two countries might have serve as possible reason in explaining the different interpretations of the undertaken variables at both the student and school level.It should be cited that Shanghai came 1st,while Peru came 65th among the 65 countries that participated in PISA 2012.In addition to this,findings of this study provide strong evidence to support the view that countries vary and a learning model that is functional in a country might not be applicable to another country.Furthermore,with detailed descriptions of research design,method and data analysis steps,this research work serves as an example for other researchers to replicate this study either with other countries that participated in PISA 2012 and 2015 assessments or with other international databases such as TIMSS,SACMEQ and PIRLS.