It was investigated whether test anxiety (TA), mathematics anxiety (MA), and reading anxiety (RA) can be traced back to some type of general academic anxiety or whether these are separable. A total of 776 fifth graders (M_age = 10.9 years) completed questionnaires on TA, MA, and RA, as well as a mathematics test. Also, mathematics and reading performance results from the National Tracking System were requested. The sample was randomly split into two halves. Confirmatory factor analyses showed that a three-factor model (factors: TA, MA, RA) had superior model fit compared with a one-factor model (factor: “Academic anxiety”), in both halves. The resulting anxiety factors were related to math performance measures using structural equation models. A scarcity of data on reading performance prevented the analysis of links between anxiety and reading performance. Anxiety–math performance relations were stronger for MA than for TA and MA. We concluded that TA, MA, and RA are separable constructs.

Leibovich et al. proposed that the processing of numerosities is based primarily on a “sense of magnitude.” The consequences of this proposal for how numerical symbols acquire their meaning are, however, neglected. We argue that symbols cannot be learned by associating them with a system that is not yet able to derive discrete numbers accurately because of immature cognitive control.