Question: is non-decision time estimated in relation to the fastest RT in the data set or the fastest RT for each participant?

[Camilla Eva Andersen]

Hi all,

I have been trying to understand how the non-decision time parameter is estimated. If I want to estimate the mean non-decision time for each of my two conditions and and for each participant for the two conditions, is it then bounded by each participants fastest RT or only the overall fastest RT (or something else)?

If some of my participants have made a response by error which is super fast, then that would be the non-decision time that will be the minimum non-decision time I assume? So in that case, I would need to be careful about removing unlikely fast RTs?

Best regards,
Camilla

[Michael J Frank]

Hi Camilla,

 all parameters are estimated based on the likelihood of the full set of choices and RTs for each subject (and also the likelihood of those parameters given the group distribution, which is the hierarchical part). The non-decision time shifts the whole distribution by a constant, and indeed any RT for subject i < non decision time (subject) i is very unlikely -- EXCEPT that we also usually include p_outlier =.05 which means that 5% of the data are assumed to come from a generative process other than the DDM and that includes a uniform distribution that can allow for very fast (or other RTs). Nevertheless it is common to remove very fast RTs that are completely implausible, assuming that there are only a small number of them.

--
You received this message because you are subscribed to the Google Groups "hddm-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hddm-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hddm-users/0b3d01ca-9651-4acc-a68c-a0f346aa5824n%40googlegroups.com.

[Camilla Eva Andersen]

Hi Michael,

Thank you for your response, that makes sense to me.

The data I am looking at is based on a task where participants had to respond if stimuli were warm or cold. We can feel cold much faster than warm, because of a difference in fiber latency, so I expect a difference in non-decision time between warm and cold. I have defined t as such: 't ~ 0 + C(stim, Treatment('warm'))'. In this case, would HDDM bound t by the minimum RT for each participant for cold and warm separately? So that there is not just a min RT for each participant but there are two min RTs for each participant; one for cold and one for warm?

Best regards,
Camilla

[Michael J Frank]

with that regression you are estimating a separate nondecision time parameter for each condition so its estimate would be informed by the minimum (and other) RTs only for that condition.  

To view this discussion on the web visit https://groups.google.com/d/msgid/hddm-users/a36bf058-a815-48d3-91fc-06f18140848bn%40googlegroups.com.