Thanks for the reply @ricardoV94.
The shifted wald distribution used for response times in decision-making studies, substracts a non-decision time, expressed as a random variable, from the observed response time. The remaining observed response time is considered the true decision time.
But, the shifted wald distribution is for correct response times only. The censored shifted wald distribution adds a survival function to the logp to account for incorrect responses.
If you may be interested, this paper might be a good read On the Relation Between the (Censored) Shifted Wald and the Wiener Distribution as Measurement Models for Choice Response Times - PubMed (nih.gov).
Hence, I couldn’t figure out how to sample from the shifted wald or censored shifted wald distribution. E.g., While sampling from pm.Wald distribution to mimic shifted wald, should I add back the non-decision time to the mean or to the drawn sample?
And it becomes a bit more complicated for me for the censored shifted wald distribution.