Closed bboudaoud-nv closed 3 years ago
Nickerson and Burnham in 1969 published "Response Times with Nonaging Foreperiods" (pdf) which makes the case that a uniform random distribution is less good for this purpose than a truncated exponential distribution.
I'm referencing this source in case we manage to update this PR with a truncated exponential distribution.
This branch adds support for random-in-range
pretrialDuration
. This does not break previous configs by allowing experiment designers to specify either a scalar value or a 2-array of values for randomization in range. This array is expected to be formatted[min value, max value]
.Exceptions are raised when the specified array is undersize (of length 0 or 1), and warnings are printed to the log when oversized (of length >2).
Merging this PR closes #309.