The document discusses a study that tested how knowledge of previous convictions influences judgments of guilt. It describes two predictions: 1) those with a logical framework will conclude "not guilty" regardless of convictions, while those with a probabilistic framework will conclude guilt is more likely with convictions; and 2) underlying guilt ratings will not/will be increased by convictions depending on the framework. The methodology involved a scenario of a child's death with/without mention of previous convictions. Results showed convictions biased verdicts and guilt ratings, indicating probabilistic rather than logical reasoning. The implications for law and theory are discussed.