Fixed step staircase objects

An object of type fixedStepStaircase is an “n-down/m-up” (i.e. n correct responses → step harder, m incorrect responses → step easier) staircase with fixed step sizes. Subject threshold is calculated as the average staircase value across reversals. In an experiment script you can make a fixedStepStaircase object outside trials using function fixedStepStaircaseObject and add it using addToExperiment.

val1

No default

Initial staircase value (a number).

stepRule

Default: step on 2 correct / 1 incorrect responses in a row

A 1×2 vector setting how many [correct incorrect] response trials in a row are needed to step the staircase value in the respective direction. e.g. <cd>[2 1]<cd> gives a 2-harder/1-easier staircase.​

stepSizes

No default

A 1×2 vector setting staircase value step sizes for [correct incorrect]. One number must be positive (increase staircase value), the other negative (decrease). Which one you make positive or negative depends on whether it’s an increasing or decreasing psychometric function.

numFastReversals
fastStepSizes

Default: numFastReversals = no fast reversals
Default: fastStepSizes = stepSizes above

You can set a fixed step staircase so that for the first few reversals it steps on every response and maybe takes larger steps in order to get close to the subject’s threshold fast. These “fast reversals” count toward numReversals and maxNumReversals but not threshold.

numFastReversals is number of reversals to do fast. 0 = none.

fastStepSizes is like stepSizes above. <cd>[]<cd> = same as stepSizes.

Input properties all objects have

report
info

Input properties all staircase objects have

min
max

maxNumReversals

PsychBench uses record properties to record information during experiments. You can't set record properties but you can see them in experiment results by listing them input property report.

Record properties all staircase objects have

val

responseScore

reversed

threshold

numTrialsRan
numReversals