The direction of the next stimulus change depends on the participant’s previous response. A correct response leads to a lower intensity, while an incorrect response leads to a higher one. This alternating adjustment keeps testing focused near the participant’s changing performance boundary, allowing trials to move toward the sensory or behavioral level being estimated rather than sampling intensities indiscriminately.
Reversal points mark changes in the direction of the staircase, such as a decrease in intensity followed by an increase, or the reverse. Because these points occur as responses alternate between correct and incorrect, they identify the region where performance is near the targeted threshold. Researchers use the collected reversal points to calculate a threshold estimate.
Fixed-level testing distributes trials across predetermined intensities, including levels that may be far from a participant’s threshold. The adaptive staircase instead uses each response to select the next intensity, concentrating measurements near the relevant performance boundary. This response-guided sampling can reduce the number of trials needed to characterize detection, discrimination, or another behavioral limit.
A researcher presents a stimulus at an initial intensity and records whether the participant responds correctly. The next trial uses a lower intensity after a correct response or a higher intensity after an incorrect response. Across successive trials, the researcher tracks the changing intensity and identifies reversal points, which are then used to estimate the threshold or performance level of interest.
The procedure can assess several types of behavioral limits, including sensory detection, discrimination, and pain sensitivity. It is also applicable when researchers want to estimate another boundary in performance rather than compare responses across many fixed stimulus levels. The selected task determines what the resulting threshold represents, while the staircase provides an efficient way to locate it.
Threshold estimates summarize where each participant’s responses approach the performance level selected for the task. Researchers can compare these estimates across individuals or across perceptual and behavioral conditions to characterize differences in sensitivity or performance. Reversal points provide the trial-level basis for those estimates, linking the final comparison to the response pattern observed during testing.