We model endogenous perception of private information in single-agent screening problems, with potential evaluation errors. The agent's evaluation of their type depends on their cognitive state: either attentive (i.e., they correctly perceive their type) or inattentive (i.e., they might misperceive their type). The mechanism's incentives structure determines the agent's cognitive state via costly investment in cognition. We derive a general representation of attention incentives, show how they vary with the mechanism's allocation rule, and define a notion of accuracy of perception. In applications we showcase how perception both shapes and is shaped by the design of mechanisms.