| Subjects were asked to estimate their preferences for intact objects (i.e., neckties) without experiencing the intact objects, using information about decomposed components of the objects (i.e., pattern, background color, and foreground color). It was examined whether subjects' estimations became better with immediate feedback over multiple trials. Because the task was about one's own preferences, 'experiential feedback' was introduced, which was to have direct experience of the intact objects immediately after estimation. The results from two experiments showed that people improved their estimation with experiential feedback by having better memory or simulation of the intact objects, figuring out which components were relevant and how components combined to determine their preferences, and getting more consistent in their estimations. With more experience of the intact objects, subjects' intact judgments became more consistent as well. Experiential feedback was compared to outcome feedback, and the results suggested that these different kinds of feedback induced different kinds of learning. Experiential feedback was more effective in inducing rule learning (e.g., figuring out the relationships between the components and the criterion, and among the components) than outcome feedback, which induced more rote learning (e.g., remembering the ratings given as outcome feedback). Learning from experiential feedback seemed not to be easy because there were interactions among the components, which often makes learning difficult. In addition, what subjects had learned was not provided explicitly by experiential feedback, but instead subjects had to figure it out for themselves by engaging in a hypothesis-testing-like process. |