Two experiments were conducted to examine (1) whether a trait hypothesis about a target person sets up expectancies for the type of evidence that will be received and (2) how subjects use a trait hypothesis and actual evidence about the target to draw conclusions about the validity of the hypothesis and, thus, to form an impression of the target person. In Experiment 1, regardless of the type of question selected (i.e., hypothesis true or alternative true), subjects expected to receive hypothesis consistent evidence. In Experiment 2, subjects′ impressions of the target were a joint function of the initial trait being tested and the type of evidence received. Hypothesis consistent evidence led to dispositional attributions that the target possessed the hypothesized trait. Consistent with Reeder and Brewer′s (1979) analysis of trait schemata, hypothesis inconsistent extraverted evidence led to dispositional attributions of extraversion, whereas hypothesis inconsistent introverted evidence led to hedging and situational attributions. Discussion focuses on the limitations of the question selection methodology that predominates the trait hypothesis testing literature and on the importance of considering the properties of traits to more fully understand the process of testing trait hypotheses.