choice-metrics.com

Hello everyone,

I use design with scenarios like:
....
;alg=mfederov
;require:
Alt1.B>Alt2.B
U(Alt1) = A[-0.6] * age[20,30,40,50]+B * income[1,2,3,4] /
U(Alt2) = A * age[age]+ B* income$

Now I like to program, that income in Alt1 is higher in each choice card than in Alt2. The require comand needs alg=mfederov, however, running the design results in an error that says mfederov is not suitable with scenarios.
Is there an other way to fix my problem?

Thank you!
Michael

There is an easy fix.

For scenarios, you need to use:

;require:
alt1.age = alt2.age

You cannot use shortcut age[age] as this only works in the default swapping algorithm.

If you want alt1 income to be higher than in alt2, then you need:

;reject:
alt1.income < alt2.income

Note that the modified Federov algorithm ignores attribute level balance, so you may want to impose some attribute level constraints, such as B*income[1,2,3,4](4-6,4-6,4-6,4-6) if you want all levels to appear in each column a minimum of 4 times and a maximum of 6 times.

Michiel