Research On The Application Of Multilevel Rasch Model In Preschool Children’s REMA-SF Measurement

Mathematical development is an important part of children’s cognitive development.The development of children’s mathematics level in the preschool period not only directly affects their future primary school mathematics learning,but also has important significance for their future academic development and life.Preschool children’s understanding and application of mathematics-related knowledge have shown certain individual differences.Scientific measurement of children’s early mathematics ability is an important prerequisite for providing effective mathematics education for them later.REMA-SF(Part A of the Research-based Early Mathematics Assessment-Short Form)scale is one of the classic tools to test children’s mathematics development level based on children’s mathematics learning and development trajectory.In this study,360 preschool children with an age range of 28 to 84 months(57.58±10.75 months)from a place in eastern China were selected as the research objects.Applying the multilevel Rasch model to the evaluation results of the research subjects on the REMA-SF scale can not only initially explore the functional differences of the items in the REMA-SF scale,but also deeply explore the relevant factors at each level and their effect sizes that affect the evaluation results.The first study was based on the multilevel Rasch model,and the Lord’s chi-square test was used to test the functional differences of items at the two classification levels of gender(boy/girl)and region(city/township).The results showed that: Item 27,Item33,and Item 35 have slight differences in item function between boys and girls,indicating that the above three items were more inclined to boys;Item 1,Item 3,Item5,etc.a total of nine items have small functional differences between city children and township children,indicating that the above nine items are more inclined to city groups;however,the difference levels in the above two results are relatively small in the item function difference test,so they can be ignored in the practical application of the scale;in addition,the difficulty values of item 1 to item 8 and item 10 are all less than 0 in the above group,so the above items can be considered to be easy to a certain extent.The second study was based on the response results of 360 subjects in the REMASF measurement and the class course evaluation results of the "Towards High Quality-China’s Kindergarten Education Quality Evaluation Standard"(referred to as "High Quality Standard")scale.The multilevel Rasch model was used to analyze the factors affecting the mathematical development of preschool children from the individual level and class level.The results showed that: "gender" doesn’t affect children’s answering results on the REMA-SF scale,and its effect size was also small;"age" and "studentteacher ratio" have positive and negative effects,respectively,on children’s responses on the REMA-SF with high effect sizes;"mother’s educational level" and "teacher teaching age" have medium and high effect sizes,respectively,on children’s responses on the REMA-SF;the effect size of "class curriculum quality" is medium,but it doesn’t affect children’s answering results on the REMA-SF scale;"gender*student teacher"ratio","age*teacher’s teaching age","mother’s educational level*student-teacher ratio",and "mother’s educational level*curriculum quality" all have significant positive predictive effects on children’s mathematics development,among them,the effect size of "gender*student-teacher ratio" is low,the effect size of "age*teacher’s teaching age"reaches a medium level,and the effect sizes of the other three all reach a high level.In conclusion,the following conclusions are drawn:(1)There are no non-negligible item functional differences between genders and regions in the REMA-SF scale.(2)From the perspective of individual-level predictors: "gender" does not affect the level of children’s mathematics development;"age" and "mother’s educational level" have significant positive predictive effects on preschool children’s mathematical development.Among them,"age" has the greatest positive predictive effect,followed by "mother’s educational level".(3)From the perspective of class level predictors: the "student-teacher ratio" has a negative impact on children’s answering results on the REMA-SF scale with a high level of effect size;"teacher teaching age" has a positive impact on the children’s answering results on the REMA-SF scale with a high level of effect size;"class curriculum quality" also has a positive predictive effect,but it is not significant.(4)From the perspective of cross-layer interaction effects: "class quality *mother’s educational level","student-teacher ratio * mother’s educational level" and "age * teacher’s teaching age" can all significantly affect children’s mathematics development;although "student-teacher ratio*gender" has a significant effect,its effect size is extremely small and has no practical significance.