SPSS -- ANALYZE -- MIXED MODELS - LINEAR
Manolis Giahnakis
Subject, Hearing, Gender, ConditionA, ConditionB, Result
Subject: 1, 2, 3, ..., 20 (nominal)
Hearing (for each subject): normal/impaired (nominal)
Gender (for each subject): male/female (nominal)
ConditionA: 0, 1, 2 (3 different values, with a certain meaning
each, nominal)
ConditionB: 0, 1, 2, 3 (4 different values, with a certain meaning
each, nominal)
For each Subject i applied combinations of the two conditions
ConditionA kai ConditionB receiving - for each combination - the value
of the variable Responce (the dependent variable; scale; e.g. 31.8),
repeating that 15 times for the same combination of conditions. (The
repetition has nothing to do with time).
Trying to apply SPSS -- ANALYZE -- MIXED MODELS - LINEAR ...
a) Should i use a new variable (Rep) describing the repetitions? its
values?
b) What should i specify for:
Subjects, Repeated, Repeated covariance type, Factors(s), Covariate
(s), and (in Random:) Subject Groupings Combinations ?
Thanks in advance,
DrM.G.
Bruce Weaver
It sounds like this is an experimental study with complete data
for every subject, is that right?
Why do you have 15 repetitions per combination? Do you care about
the raw measures across those repetitions? Or are they simply
there to allow computation of a mean for each combination (because
the mean allows a more precise estimate of the effect of the
various combinations on the DV)? (E.g., in research that uses
response time as the DV, ANOVA is often performed on the median,
or trimmed mean RT for each condition.)
If you are happy to analyze the means for each combination, it
reduces to a 2x2x3x4 mixed design ANOVA. If you have complete
data for each subject, I think you would be better off using GLM -
Repeated Measures.
--
Bruce Weaver
bwe...@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/
"When all else fails, RTFM."