choice-metrics.com

Hi everyone,

I recently purchased Ngene and is learing to make some simple SP designs.

One simple question:

I can run the following codes without error:

Design
? This will generate a full factorial design
;alts = alt1, alt2
;rows = all
;fact
;model:
U(alt1) = b1 + b2 * A[0,1,2] /
U(alt2) = b3 * B[2,4,6,8] $
However, I always got an error if I reduce the attribute levels of B from 4 to 3:

Design
? This will generate a full factorial design
;alts = alt1, alt2
;rows = all
;fact
;model:
U(alt1) = b1 + b2 * A[0,1,2] /
U(alt2) = b3 * B[4,6,8] $

Can anybody tell me why? Thank you in advance.

Best Regards,

Zuduo

This seems a bug, as when I put ;rows = 9 it does work. Somehow "all" does not get computed correctly into a number, we will investigate this. Thanks for pointing this out to us, not many people use full factorials in Ngene, so that is probably why we did not hear about this before.

Yes, indeed this is a bug, that affects cases where the full factorial has less than 12 rows. We will fix this in the next point release. In the meantime, just set the number of rows explicitly.

Andrew

Thanks a lot, Michiel and Andrew.
I was new to Ngene. So, I wanted to use some simple example to test output from Ngene with the mannual calculation. That's how I caught that bug.

Anyway, I agree that in practice few people would be interested in doing a full factorial design.