Dear All,
a) I did some programming on the weekend, and we have now the possibility
to present colored bars, or circles for an distractor task (e.g. how many
red ones? or Any X ones? with X a different colour each time). Even a
simple task like that can be pretty demanding if one puts time pressure on
the subject. Moreover, we're now able to collect the responses by letting
the subjects type what they've hear into the computer, which should
eventually help us with counting the errors/hits.
b) I did some planning of the statistics: If we concentrate in the
factorial test on the interaction of the three basic effects
(time,f0,resonance) than I see two viable options:
1) a Latin square this would require 8 subject groups
and 8 sentence group (i.e. the number of sentences
should be a multiples of 8, contrary to opinions I had
previously..., sorry)
2) a completly confounding design, which wouldn't require
the numbers of sentences to be a multiples of 8, but which
would be less powerfull because we couldn't control the
influence of the various subject groups. We thus would
have to rely on the fact that the groups are more or less
equivalent, which would mean that we would need more subjects
in each group.
There might be another options putting less constraints on the number of
subjects or sentences but requiring more intricate designs. I'll tell
you if I find out anything.
c) concentrating on the three main factors, would mean that we have to
go ahead with the pretest rather quickly, because we cannot gain
any knowledge from this experiment about different possibilities of
e.g. f0-alignment, but only the general fact that it helps (or not). In
order to have homogenous groups of sentences it would be very helpful, to
know as much as possible about the effects of the various treatments in
advance.
Sebastian
*********************************************************************
Sebastian Heid Email: sh276@cam.ac.uk
Phonetics Laboratory Phone: +44 1223 33 50 50
Dept. of Linguistics Fax: +44 1223 33 50 53
University of Cambridge
Sidgwick Avenue
Cambridge CB3 9DA
United Kingdom
*********************************************************************
This archive was generated by hypermail 2b29 : Mon Jan 17 2000 - 16:37:46 GMT