Thu, 08/21/2014 - 12:21

Hello,

I'm a novice in structural equation modeling. I am curious if Onyx works for the following experimental design:

There are 2 treatment variables (T1, with 2 levels; and T2, with 3 levels; each with 3 replicates) and 3 response variables (R1, R2, R3). I'm interested in how each of the T variables affects the R variables, as well as if this relationship changes depending on the level within each T variable. I'm also interested to know how R3 is related to R1 and/or R2.

I would like to repeat this simple design with more replicates and more R variables later. I tried putting the data in Onyx with help from the guide but wasn't sure how to specify some of the variables/paths and other options to see the relationships quantified...

Thanks for any input.

Yev

Here is some sample data with this design:

T1 T2 R1 R2 R3

0 0 7.7 2.8 7

0 0 7.5 1.8 7

0 0 7.3 4.0 8

0 1 7.2 4.9 9

0 1 7.0 22.0 8

0 1 7.0 1.9 7

0 2 7.4 1.3 9

0 2 7.0 1.8 8

0 2 7.4 1.7 9

1 0 5.9 302.6 7

1 0 6.6 141.4 9

1 0 6.7 113.7 7

1 1 7.5 56.2 9

1 1 5.5 89.5 9

1 1 6.0 99.2 8

1 2 7.6 23.9 11

1 2 6.5 74.7 10

1 2 6.6 33.1 12