"auda" <[EMAIL PROTECTED]> wrote:
>Thanks for your advice. I should describe my experimental
design more
>detailed.
>
>I have two within subject factors, say A (2 level) and B (4
level). Subjects
>do 24 trials for each combination of A and B. The dependent
variable is the
>judgement of target position. What I concern is: whether
certain levels of A
>or B lead to more variation in subjects' judgements of target
position. Of
>course, I can compare the means in every condition. However, I
am more
>interested in whether every subject performs more variablely in
the
>precision of position judgement in some condition than others.
For example,
>for A1xB2 condition, subject 1 gives out 24 responses, and thus
has a
>variance caculated form these 24 responses. Subjects can have
variances for
>their trials in every condition, just as they will have means
for their
>trials in every condition. Finally we can obtain the following
variance
>table:
>
>
>subj A1B1 A2B1 ........ A1B4 A2B4
>1 var111 var121 var114 var124
>2 var211 var221 var214 var224
>3
>4
>..
>..
>..
>n varn11 varn21 varn14 varn24
>
>
>The hypothesis I would like to test is:
>H0: variances in A1B1 is not different from A2B4.
>
>If the null hypothesis in rejected, I can infer that subjects'
judgements
>are MORE VARIABLE in condition A1B1 than in A2B4.
>
>Erik
>
>auda <[EMAIL PROTECTED]> wrote in message news:8m4q92
$r7e$[EMAIL PROTECTED]
>> Hi,
>> What kind of procedure should I use to compare the varience
between
>> different experimental groups?
>>
>> Thanks
>> Erik
As a simple preliminary analysis, you can try Levene's test
(with sample median, http://www.recursive-partitioning.com/hov)
to compare the groups. For a more thorough analysis, a
random/mixed effect modeling is a possibility. I'd suggest that
you consult with your local statistician.
--
T.S. Lim
[EMAIL PROTECTED]
www.Recursive-Partitioning.com
-----------------------------------------------------------
Got questions? Get answers over the phone at Keen.com.
Up to 100 minutes free!
http://www.keen.com
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
