Threshold determination within the framework of the Block Up-and-Down, Two-Interval, Forced-choice (BUDTIF) method has been investigated. A computerized Monte Carlo technique was used to permit varying certain procedural parameters while maintaining an invariant “listener.” The basic approach involved a comparison of the threshold means and variances obtained for a representative set of parameter values. That set of parameters yielding minimum between-mean threshold variance and minimal bias was sought. Parameters considered and general results included (1) number of trials per run—inversely related to between-mean threshold variance, (2) number of trials per block—directly related, with a critical minimum size depending on target performance level, (3) number of blocks used in each level-change decision—directly related, (4) initial stimulus level—no effect, if reasonably close to true threshold, and (5) method of calculating thresholds—no clear effect. Parameters that yielded minimum variance within individual runs tended to yield maximum variance between successive threshold estimates.