Hello !
I have fitted a LOG LINEAR MODEL
for analysing a 3 dimensional Contingency Tables with no empty cells. After
completing the necessary tests the final model was found to be
LOG ( Fijk) = m + ai + bj + ck + (ac)ik +
(bc)jk
where the factor A has 4 levels
B has 2 levels
C
has 2 levels
I fitted the above LOG LINEAR MODEL on the analogy
of Simple Analysis of Variance.
I am now interested to compare different levels of
A or different combinations of two or three factors like A
significance of effects of A1 and A3
significance of A1B2 and A3C1 etc.
I request the members of the list to kindly suggest
suitable statistical procedures, Statistical tests, Standard Errors etc. (
Preferably in the form of computational formulae ) for doing above.
Thnaks in advance
D. V. Ratnalikar
Assoc. Prof. Stat.
Deptt. Econ & Stat
Dr. PDKV, Akola ( Maharshtra )
INDIA
mial me
|