As a key outcome in environmental psychology, pro-environmental behavior is studied in various different ways. Researchers observe naturally occurring pro-environmental behaviors, ask people to complete self-report scales, or devise behavioral tasks with ecological consequences. These approaches have been claimed to yield measures of pro-environmental behavior. The present paper argues that this terminology is misleading. Some pro-environmental behavior researchers measure behavioral properties of pro-environmental behaviors (e.g., frequency, duration), while others apparently aim to measure more or less general person properties (e.g., pro-environmental propensity or preference). Behavioral properties and person properties are logically distinct, but they become conflated when researchers indiscriminately refer to observations, items, scales, and tasks as measures of pro-environmental behavior. As a result, researchers may end up evaluating their purported measures against irrelevant quality criteria, expecting convergence where it cannot be expected, using methods inconsistent with their research goal, and making spurious conclusions. They erroneously consider multi-item scales and behavioral tasks as measures of the same construct, examine the construct validity of procedures that cannot reasonably be assumed to capture a psychological construct, or mistake the correlates of person properties for the determinants of behavior. To promote a cumulative science of pro-environmental behavior and prevent misguided research efforts, researchers should carefully distinguish between measurement targets and avoid claiming that they “measured pro-environmental behavior”.