本研究从工作记忆容量的角度入手，探讨数学焦虑对工作记忆子系统及数学心算成绩的影响，并寻求三者之间关系的理论解释。本研究为3×2×2混合实验设计，即三组被试（高、中、低数学焦虑组）、工作记忆两个子系统（言语工作记忆和视空间工作记忆）和两种数学心算任务难度，并记录数学心算的反应时和正确率作为其成绩指标。结果表明在要求较高工作记忆容量的数学心算任务中，数学焦虑通过言语工作记忆和视空间工作记忆都会对数学心算成绩产生影响，其中视空间工作记忆是数学焦虑与复杂数学心算反应时之间的中介变量。

Mathematics anxiety not only affects the development of individual emotion, but also affects the development of individual mathematics cognition. Ashcraft (1994,1998) et al. attempted to develop a program of research that examines the possible cognitive consequences and correlates of mathematics anxiety and got a lot of valuable results. But there are some questions which were still unclear, for example, mathematics anxiety impacts on which working memory subsystems more greatly, how mathematics anxiety impacts on mental arithmetic achievement and so on. So in this article, we will discuss the effects of mathematics anxiety on working memory subsystems and mental arithmetic from work memory capacity. And then we will seek the theory explanation that can explain the relationship. It’s 3×2×2 mixed design: Mathematics anxiety (high, medium, low), working memory (the phonological loop, the visuospatial sketchpad) and mental arithmetic tasks. Mathematics anxiety was measured in classroom by group test. The working memory subsystem and mental arithmetic task were measured in a quiet laboratory using the computer by individual test. And the reaction time and accuracy are regarded as the mental arithmetic performance index. The experimental results indicate that mathematics anxiety impacts on visual-spatial working memory more greatly. In mental arithmetic tasks which request high work memory capacity, mathematics anxiety is correlated negatively with mental arithmetic achievement, and mathematics anxiety can influence mental arithmetic achievement by the phonological loop and the visuospatial sketchpad and that the visuospatial sketchpad is the medium variable between mathematics anxiety and mental arithmetic (carry) reaction time. According to these findings, we can integrate Processing Efficiency Theory proposed by Eysenck and Calvo (1992) and Inhibition Theory proposed by Hasher and Zack (1991) to explain what we discovered in this experiment. That is, for those who are in high mathematics anxiety, their mandatory thoughts or fear and other negative emotions can struggle for limited processing resources in working memory with ongoing cognitive tasks ,and then reaction time will increase or accuracy will cut, that is to say the cognitive efficiency reduces. From the perspective of inhibition, we also can say those who are in high mathematics anxiety are confronted with some difficulties including failure of inhibition of these ideas. But the theory of Arousal-Performance Function can’t be used to explain these findings. It may indicate that the mathematics anxiety is a special type of anxiety. It may also be associated with the mathematics achievement evaluation index used in the experiment (the reaction time and accuracy of mental arithmetic task, rather than the students' exam results).