Correcting response measures for guessing and partial information

Demonstrates that proportions, response time moments, distributions and histograms, and response confidence when only a single stimulus category (such as words) is used can be corrected just as guessing on a response choice and response time test can be corrected. An example is presented in which response time for Korsakoff patients was corrected. Corrections under equal and unequal bias are addressed, and a stochastic guessing model is presented. A complete set of parameter estimates is obtained by analyzing only responses to word stimuli; therefore, false alarms (such as responses to nonwords) are not a necessary component of a correction for guessing. The correction for guessing illustrates the theory that beneath the surface of data lie relations between psychological processes that can be revealed in a mathematical format. (13 ref)