Hi Statalist,

I will appreciate help with this problem. First, here is my data:

Code:
* Example generated by -dataex-. To install: ssc install dataex
clear
input long(group treatment) byte time double ratio byte(_est_model1 _est_model2)
1 1 1  .9983001875221594 1 0
1 1 2 1.0053739531421417 1 0
1 1 3 1.0030314592038265 1 0
1 1 4  .9932944001318715 1 0
1 2 1 1.0898057417568519 1 0
1 2 2 1.0721209699002034 1 0
1 2 3 1.0826934924466416 1 0
1 2 4 1.0756296048445315 1 0
1 3 1 1.0655604477977156 1 0
1 3 2  .9813086269478382 1 0
1 3 3  .9783118217772869 1 0
1 3 4   .983333805554108 1 0
1 4 1   .984920213199492 1 0
1 4 2  .9911889207870019 1 0
1 4 3  .9723713002587203 1 0
1 4 4  .9641145313352677 1 0
2 1 1  .9869416069631516 0 1
2 1 2  .9879977506775215 0 1
2 1 3  .9944675490019041 0 1
2 1 4  .9895596930121271 0 1
2 2 1 1.0027612460924615 0 1
2 2 2   .989941671061605 0 1
2 2 3  .9899114163046291 0 1
2 2 4  .9927223347255462 0 1
2 3 1  .9901865879888093 0 1
2 3 2 1.0029885911645473 0 1
2 3 3  .9939653025024692 0 1
2 3 4  .9978908163525125 0 1
2 4 1  .9998596837924634 0 1
2 4 2 1.0047053051308992 0 1
2 4 3   .997293411572137 0 1
2 4 4   .998141599504501 0 1
end
label values group group2
label def group2 1 "Control Virus", modify
label def group2 2 "Control+HERG Virus", modify
label values treatment treatment2
label def treatment2 1 "A", modify
label def treatment2 2 "B", modify
label def treatment2 3 "C", modify
label def treatment2 4 "D", modify
I run the following commands:
Code:
regress ratio i.treatment if group==1
estimates store model1

regress ratio i.treatment if group==2
estimates store model2

estimates table model1 model2, varlabel varwidth(25)  b(%6.3f) star(0.05 0.01 0.001) stats(N r2_a)
Next, I would like to graph the mean difference between treatment A (reference group) versus treatments B, C, D. I want to do this for the difference between models 1 and 2. The research question is whether the addition of the HERG virus to the control virus (model 2) would change the mean differences between the four treatments that were observed in model 1 (control virus only).

Your help will be appreciated.

Al Bothwell