Measuring Students' Representation Level on Arithmetic Word Problems: Based on the Relational-Representational Complexity Model

XIN Zi-qiang,ZHANG Li
2009-01-01
Abstract:Students often have difficulty in solving arithmetic word problems.According to the relational-representational complexity model developed by Xin Ziqiang(2003-2008),the difficulty of an arithmetic word problem is determined partly by the relational complexity of variables in the problem.Based on the model,in the present study,a test consisting of 24 arithmetic word problems with different levels of relational complexity was designed for measuring children.The test was administrated to 165 students from grade 4,5,6 and 7.Results showed that:(1) prior analyzed problems' relational complexity(involving analysis of relational complexity of variables in every item) could predict most variance of task difficulty(73.7% of error percentage,78.7% of a difficulty index of Rasch model),and the criterion validity with the SOLO taxonomy test was 0.65;(2) students' representation levels on the test improved with grades.Such results indicted that the test is an effective and reliable instrument for measuring children's representation levels changing with grades.
What problem does this paper attempt to address?