Extant research has demonstrated that anxiety is negatively associated with self-efficacy, especially in science. However, social cognitive theory also posits that anxiety and self-efficacy are likely to dynamically interact (i.e., moderate), such that a student high in anxiety may not garner the benefits of high self-efficacy. It has been suggested that classrooms may also be characterized in terms of this problematic dynamic, such that class-average anxiety may impede potentially positive effects of class-average self-efficacy. Despite this, very little work has focused on the extent to which anxiety thwarts the positive effects of self-efficacy at either the student- or classroom-level, and none has done so among secondary school students or in science specifically. Thus, the present study examines the main effects of science anxiety and self-efficacy on science achievement, as well as the moderating effects of science anxiety on the relationship between science self-efficacy and achievement, at both the student- and classroom-level. With a sample of N = 1,075 high school students clustered in N = 99 science classrooms, doubly latent multilevel structural equation modelling demonstrated that science self-efficacy positively predicted science achievement at both student- and classroom-levels. Also, at the student-level (but not the classroom-level), science anxiety negatively moderated the effects of science self-efficacy on science achievement, such that students high in self-efficacy with higher anxiety scored lower in science achievement than those with lower anxiety. Interestingly, however, student-level (but not classroom-level) findings also suggested a potentially arousing role of anxiety for students low in self-efficacy. These findings have theoretical implications and suggest that a dual intervention approach (i.e., concurrently promoting science self-efficacy and reducing science anxiety) at student-level and interventions targeting self-efficacy at the classroom-level may be warranted to optimize science achievement.

现有的研究表明，焦虑与自我效能感呈负相关，特别是在科学领域。但是，社会认知理论也认为，焦虑和自我效能感可能会动态地相互作用（即适度），从而使焦虑程度高的学生可能无法获得高度自我效能感的好处。已经提出，教室也可以根据这种有问题的动态来表征，使得班级平均焦虑可能会阻碍班级平均自我效能的潜在积极影响。尽管如此，很少有研究集中于焦虑在多大程度上阻碍了学生或教室一级自我效能的积极影响，而在中学生或科学领域，则没有这样做。从而，本研究从学生和课堂两个层面考察了科学焦虑和自我效能感对科学成就的主要影响，以及科学焦虑对科学自我效能感与成就之间关系的调节作用。带有样品N  = 1,075名高中生聚集在N中 = 99个科学教室，双重潜在的多级结构方程建模表明，科学自我效能感在学生和教室两级均积极预测科学成就。同样，在学生级别（而不是课堂级别），科学焦虑消极地减轻了科学自我效能感对科学成就的影响，因此，自我效能感高且焦虑较高的学生在科学成就方面的得分低于那些降低焦虑。有趣的是，学生水平（而非教室水平）的发现也表明，自我效能感低下的学生可能会引起焦虑。这些发现具有理论意义，并建议采用双重干预方法（即，同时促进科学自我效能和 减少学生层面的科学焦虑感）和针对课堂自我效能的干预措施可能需要优化科学成绩。