Abstract
The vast majority of psychoacoustic testing involves presenting a subject with a series of short, simple questions (trials). A common desire is to track the level of statistical confidence in the results after each trial so that one can stop the test once a threshold of certainty has been surpassed. Typically, this certainty statistic will take the form of a confidence interval (CI). Past methods to compute CIs involve assumptions that may impact the performance of the resultant test-stopping criterion. This talk proposes a method to compute CIs that is free of assumptions regarding: the testing procedure used (e.g., up-down staircase, PEST, non-adaptive), the testing modality used (e.g., n-alternative forced-choice, yes/no), and the model psychometric function used (e.g., logit, Gompertz). The method involves a fixed-quadrature integration of the likelihood function of the model parameters given the data retrieved after each trial. The execution time is demonstrated to be low enough so that the method ma...
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have