Re: ANOVA and Interaction

Bernhard Laback wrote:
Hi,

Thanks for your offer. Here are the data:

A B Dependent
200 1 25.28
200 1 30.49
400 1 31.58
400 1 18.96
500 1 30.91
500 1 29.97
540 1 38.34
540 1 47.56
588 1 298.2
588 1 131
200 2 37.89
200 2 27.04
400 2 30.98
400 2 16.02
400 2 21.22
500 2 36.55
500 2 57.04
540 2 44.21
540 2 58.63
540 2 50.47
588 2 216.4
588 2 164.2
200 3 28.56
200 3 29.86
400 3 48
400 3 43.26
500 3 118.9
500 3 80.38
540 3 298.6
540 3 177.2
588 3 600

The particular question I would like to answer is at what levels of A the diffences between the levels of factor B are significant.

Thanks in advance
Bernhard



If I understood OMU's analysis, he B as if it was a continuous (scaled) variables--e.g., there is only one t-test for B in the table of coefficients. But when I look at these data, I see B as a nominal variable, with the values 1-3 representing 3 independent groups of observations. If this is correct, B needs to be represented by two indicator variables (or some other type of contrast coefficients) in the regression model, and will have 2 degrees of freedom associated with it.

--
Bruce Weaver
bweaver@xxxxxxxxxxxx
www.angelfire.com/wv/bwhomedir
.