The types of knowledge captured using three knowledge acquisition (KA) methods, conceptual graphs, backward thinking, and sorting tasks, were examined. One hundred and forty-four participants were assigned to one of four conditions: (a) procedural-spatial, (b) procedural-verbal, (c) declarative-spatial, and (d) declarative-verbal. Participants in procedural conditions learned a virtual environment's layout before performing one of three KA tasks. Participants in declarative conditions received a tutorial about McCune-Albright Syndrome before completing a KA task. The procedural and declarative information was presented either spatially or verbally, depending on the condition. The predictive validity of each KA method for each information type (procedural vs. declarative), task type (i.e., spatial vs. verbal), and combinations of information and task types was evaluated by comparing the accuracy of the knowledge depicted by KA methods to various performance measures. Findings indicated that both task type and information type affect the validity of KA methods for capturing knowledge structures.