Various inquiry approaches have been promoted in science classrooms as a way for students to engage in, and have a deeper understanding of scientific discourse. However, there is a paucity of empirical evidence to suggest how children's actions and engagement in these approaches, or "practical epistemologies" (Sandoval, Science Education 89(4): 634-656, 2005), may contribute to the development of their "personal epistemologies", or their views about the nature of knowledge and knowing and the nature of learning. This paper puts forth the conceptualization and development of the Practical Epistemology in Science Survey (PESS) instrument, a 26-item Likert-scale self-assessment which measures how students view their individual and social participation in the classroom scientific community. Data were collected from 4th-6th-grade students (n = 1019) in the USA and a psychometric evaluation of the reliability, validity, and dimensionality of the instrument was conducted. The Cronbach's alpha value indices for all subsets of items of the PESS suggest a strong reliability of the instrument (a = 0.80). The development of the PESS may be useful in science education research to (a) detect changes to students' beliefs about knowledge and knowledge development; (b) identify dispositions and beliefs which may or may not be in line with the aims and values of various pedagogical approaches; (c) monitor the process of change, e.g., time it takes for students to change their approaches and beliefs with respect to teacher practice; and, (d) overall, to provide an understanding of how students' formal epistemologies are developed and informed by the affordances in science classrooms.