choice-metrics.com

Hi,
I want to use the “;eval=” option for evaluating more than one design (for evaluating a heterogeneous design). I’ve prepared the . csv file as it is written in the manual, page 22. The first column states the design numbers, the second column states the choice sets, etc.
The syntax I have is:
Design
;alts = alt1, alt2, alt3
;rows = 15
;eval = …\design.csv
;eff = (rppanel,d)
;rep = 1000
;model:
U(alt1) = …/
U(alt2) = … /
U(alt3) = … $
Ngene gives an error as “More designs found in data that in model specification.” Actually, this is what I have expected, but I’m confused about writing the proper syntax.
What is the proper and the easiest way to tell Ngene that there exist 100 different designs to be evaluated in model specification?
Thanks in advance.
Regards,
Deniz

I am not sure I understand it correctly, but you seem to have multiple designs while you only specify a single model. For each design you need to specify a different model (using ;model(m1): etc). So you shpuld either copy your model 100 times (have not tested whether this works) or you should copy you entire syntax 100 times with designs in different file names.