Page 1 of 1

dummy variables with interaction

PostPosted: Wed Aug 07, 2019 2:57 pm
by mmina
Dear Ngene team,

I have done a D-efficient design using Ngene for my study using dummy var with zero priors and interaction as below:
;alts = alt1, alt2
;rows = 240
;eff = (mnl,d)
;block = 20
;model:
U(alt1)=
b1[0] * DUR[1,3,5,7,10](38-58,38-58,38-58,38-58,38-58) +
b2.dummy[0|0|0|0] * Prt[1,2,3,4,5] +
b3.dummy[0|0|0|0] * Accpt[1,2,3,4,5] +
b4.dummy[0|0|0|0] * Pain[1,2,3,4,5] +
b5.dummy[0|0|0|0] * Emthn[1,2,3,4,5] +
b6.dummy[0|0|0|0] * comm[1,2,3,4,5] +
b7.dummy[0|0|0|0] * slp[1,2,3,4,5] +

i1[0] * DUR * Prt.dummy[1] +
i2[0] * DUR * Prt.dummy[2] +
i3[0] * DUR * Prt.dummy[3] +
i4[0] * DUR * Prt.dummy[4] +
i5[0] * DUR * Accpt.dummy[1] +
i6[0] * DUR * Accpt.dummy[2] +
.
.
.
/
and I have done a pilot to measure the priors for the main design, I have some uncertainty about whether to put the the priors just the interaction part (in front of i's) or I need to put it with the dummy line as well (dummy[0|0|0|0] )?

Many thanks in advance,

Regards,
Mina

Re: dummy variables with interaction

PostPosted: Wed Aug 07, 2019 3:36 pm
by Michiel Bliemer
If you have priors available, you can put them both in the main effects (b1 to b7) as well as the interaction effects (i1 to i6).

Michiel

Re: dummy variables with interaction

PostPosted: Wed Aug 07, 2019 6:34 pm
by mmina
Thank you Michiel for your fast respond.