Background: Epistemological beliefs are subjective theories of the structure and acquisition of knowledge. The instruments used to measure epistemological beliefs in educational psychology (see Duell & Schommer-Aikins, 2001) typically consist of questionnaires tapping general, decontextualized beliefs about knowledge or knowledge acquisition in a specific field or in general. Aims: Using specific theories as stimuli, we determine the degree of topic-specificity of certainty beliefs as well as the association between certainty beliefs and the learning environment. Sample: Participants were 662 upper secondary school students (Study 1) and 211 college students (Study 2). Method: A global instrument and a topic specific instrument were used to collect responses to up to 10 stimulus theories. Factor analysis, multiple regression analysis and multi-level modelling were carried out. Results: Students' topic-specific certainty beliefs varied markedly across the stimulus theories. Furthermore, students in different academic environments differed more strongly on global certainty beliefs than on topic-specific certainty beliefs, and global certainty beliefs were only loosely connected to topic-specific certainty beliefs. Conclusion: Researchers should critically assess the validity of decontextualized global questionnaires for assessing certainty beliefs. If possible, global measures should be complemented by topic-specific measures.