[ACT-R-users] model validation

[Fehmida Hussain]

Sorry I accidently hit the send key and incomplete reply was sent, here is
the complete email:


Hi,

Briefly describing what I am tryging to do:

I have re-implemented an ACT-R 6 model of a computer-based reaction time
test called Attentional Network Task(ANT ref1) which is a combinatin of
cueing and flanker tasks, having various cues and flanker condtion presented
to the subject/model. The objective of my model is not only to simulate the
experiment producing reaction times but also to see whether the same
interaction/disassociation exits between the networks of attention(alerting,
orienting, control -Posner's work on attentional networks ref2).

So the model code has 2 parts: lisp code which is the experimental setup and
the productions consisting of the ACT-R part for rule firing, parameter
setting etc.


So when you talk about replication and as Rick suggested "do the same
experiment with the model", does that mean let the human subject interact
with the screen setup by my model code?

I am a little confused with the use of word replication. In the context of
'replicating the experiment' does it mean repeating the experiment and
observing results and in the context of replicating the model does it mean
come up with another model for the same experiment and then compare model
results? Is my understanding correct?



ref1: Fan J, McCandliss BD, Sommer T, Raz M, Posner MI. (2002). Testing the
efficiency and independence of attentional networks. J. Cogn. Neurosci.
3(14):340–47

ref2: Posner, M.I.  & Petersen, S.E. (1990). The Attention system of the
Human Brain. Annual Rev. of Neuroscience, 13: 25-42

thanks
Fehmida
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/act-r-users/attachments/20080915/90dad746/attachment.html>