A cognitive model of how teachers plan instruction was validated in laboratory settings but remained to be tested empirically in authentic situations. The objective of this work is to describe and compare pedagogical reasoning in laboratory and authentic contexts and across expertise levels. The "state-driven hypothesis" and the "knowledge-driven hypothesis" were used in two studies to show how pedagogical reasoning was performed by novices and experts in laboratory (n = 18) and in authentic context (n = 14). Globally, the results show (1) similarities and differences in how pedagogical reasoning unfolds in laboratory and authentic contexts and (2) how domain knowledge influences only some aspects of this process. The work presented lays the foundations for the fine-grained study of how domain knowledge determines problem-solving in pedagogical-reasoning.