A prerequisite for remediating student misconceptions is their accurate diagnosis by the teacher. However, studies on judgment accuracy show that teachers differ substantially regarding their judgments even though little is known about the reasons so far. The present study investigates whether teachers' diagnosis of misconceptions with decimal fractions is subject to judgment biases. For this purpose, we propose a cognitive model for diagnosing misconceptions based on the process of hypothesis testing. The study results show that the formulation of alternative hypothesis and the processing of relevant information are predictors of high judgment accuracy when diagnosing misconceptions. Furthermore, normative and confirmatory biased judgment processes could be distinguished. Implications for teacher education are discussed.