Human behavior is probabilistic. Even in the simplest of tasks, some randomness or unpredictability creeps into human responses. This makes it difficult to study the details of how people respond to subtly different stimuli. To gain a good understanding of such responses, it is necessary to collect large amounts of data, which can be expensive and time-consuming. This research seeks to develop new procedures for efficient data collection, and new methods for evaluating such procedures. It applies information theory and the theory of stochastic processes to develop adaptive algorithms for extracting as much information as possible from experiments. A particular test case for these methods will be experiments in visual psychophysics, which is notorious for requiring extremely long and numerous data-collection sessions. However, the methods will be developed abstractly, in order to maximize their applicability to a wide range of problems, such as experiments in economics and the evaluation of reliability of electronic components.
