Significance in 3d graphs

#1
hi all,
I have a problem:

I have to compare two 3D bar graphs. I have a total of 36 bars in a grid of 6x6.
I need to know if each bar is significantly different from the analogue bar of the other graph.
At the beginning I was doing a t test for each pair of bars but I think it is wrong as the values among a graph are part of set of data.

For those who work in drug synergy I am testing all the possible combination of six serially diluted concentration of two different drugs.
From the single dose responses I am creating a set of predicted responses to the different combinations then I'll compare the predicted values with my experimental data.
data that will fall above the predicted will be synergistic. But how can I define if an experimental data is significantly different from the predicted?

Thanks in advance!
Egle
 

bugman

Super Moderator
#2
One possibilty would be to comapre the two matrices using a Mantel test. This will test the strength and its significance between the two data sets. This would give you an idea of the overall picture. On the other hand, a MANOVA (two groups with 6 dilutions per drug - if I have this right) followed by a series of planned comparisons should answer your question.