Experiment 1 Results
Trials to learn t(16) = 1.91 p <.05
- Logical A ? (B???C) ? FireInFall 27.6
- Feature C ? (B ??A) ? FireInFall 50.4
Total Errors t(16) = 2.135 p <.05
Summary- Prior knowledge is being used to constrain the search for hypotheses (in addition to features), but too many errors to be explained by Explanation-Based Learning Methods