Millisecond Forums

#1: Your questions primarily seem to be concerned with the properties of the (A)SA procedure. For that see e.g. http://www.researchgate.net/profile/Francesco_Pavani/publication/6278477_Small-sample_characterization_of_stochastic_approximation_staircases_in_forced-choice_adaptive_threshold_estimation/links/0deec528bdbf5526e9000000.pdf

#2: Since "just noticeable difference" (JND) is usually taken to indicate the amount something must be changed in order for a difference to be detectable at least half the time, the closest analogue within the (A)SA framework used by this procedure would be 0.50, I believe.

Hi Dave,
Thank you for your reply. May I ask another question？how can I set p to calculate the differential threshold known as just notice difference？If I set p=0.25, it will measure the value up to which the comparison increase when making the correct response（comparison is perceived to be longer than the standard) 25% the time. And when i set standard=1000ms, the 25% threshold is about 1036ms. Does this 25% threshold measure lower threshold through which the differential threshold is usually calculated as (upper threshold-lower threshold)/2?
Best,
Nakayama

The script has a <values> entry that allows you to set the desired threshold

***************************************************************************************
*********************************************
Editable Values
*********************************************
***************************************************************************************
...
/probability_p:                                    target threshold probability (default: 0.75)
...
<values>
...
/probability_p = 0.75
...
</values>

Set it to 0.25 to change the target threshold to 25%.