choice-metrics.com

Dear members,
I am designing choice sets using the code below. I had requested for 12 rows but the software generated 36. My concern is that in three of the choice sets, attribute levels in alternative 1 and alternative 2 are similar. What could have caused that??

Design
;alts = alt1, alt2, alt3
;rows = 12
;orth = sim
;block=6
;model:
U(alt1) = b1 + b2 * yield[0,1,2] + b3 * mtrty[0,1,2] + b4 * drghtol[0,1,2] + b5* tast[0,1] + b6 * grnsz[0,1] + b7 * pric[0,1,2] /
U(alt2) = b2 * yield + b3* mtrty + b4* drghtol + b5* tast + b6* grnsz + b7 * pric /
U(alt3) = 0

An orthogonal design for your specification does not exist with 12 rows. The smallest orthogonal design that exists is 36, which Ngene returns.
You selected ;orth = sim, but since alt1 and alt2 seem to be of the same type, you should use either ;orth = seq or orth = ood, where the latter ensures minimum attribute level overlap across alt1 and alt2.

Please post questions with Ngene scripts into the Ngene forums; this forum is for general questions about choice experiments only.

Michiel