[Is there a feature-positive effect in classical conditioning in humans?]

Jenkins and Sainsbury (1969, 1979) first described an asymmetry in pigeon's operant discrimination learning which they called a feature-positive effect. Subsequent studies demonstrated that other animals as well as humans also show a similar asymmetry. The effect is often explained as being caused by a bias in information processing which results from the neglect of negative information as a discriminative cue. The effect has also been observed in classical conditioning experiments, but not yet with human subjects. Sometimes, however, a reversal of the asymmetry, i.e., a feature-negative effect, has been observed. The experiment reported here primarily intended to demonstrate a feature-positive effect in a classical conditioning procedure with human subjects. Secondly, we also wanted to test an alternative explanation of the feature-positive effect. We found that a substantial feature-positive effect does indeed occur during classical conditioning in human subjects. Our results, however, did not fully support the alternative explanation based on rule learning.