Ngene design with values for status quo

[niyomip]

Hi,

I would appreciate it if anyone could answer my question.

I have 4 attributes in DCE,
Water clarity (WC)= 5 (worst),10,20,30 m visibility (5 is status quo value)
Seabed impact (SB)=35 m of negative impact from a fish pen (worst), 25m negative impact from a fish pen, 15m negative impact from a fish pen, 5m negative impact from a fish pen (35m is the status quo (worst one)
Jobs (EO)= 200 more jobs (worst), 300 more jobs, 500 more jobs (200 is status quo value)
Cost=0,20,40,60,80,100,120,140,160 (status quo value for cost is 0)
We consider WC=5m,SB=35m, EO=200, Cost=0 as a current situation.

I wrote the code to generate bayesian efficient design as follows:
Design
;alts = alt1*,alt2*,alt3
;rows = 48
;block = 8
;eff = (mnl,d)
;alg=mfederov(candidates =5000)

;require:
alt1.cost<>alt2.cost, alt3.water = 5, alt3.seabed = 35 , alt3.jobs = 200
;model:
U(alt1) = b1[[(n,0.0361,0.0109)]*water[5,10,20,30] + b2[(n,-0.00569,0.00657)]*seabed[35,25,15,5] + b3[(n,0.00151,0.000563)]*jobs[200,300,400,500] + b4[(n,-0.0122,0.00266)]*cost[20,40, 60, 80, 100, 120, 140, 160]/
U(alt2) = b1*water+b2*seabed+ b3*jobs+b4*cost/
U(alt3) = b1*water+b2*seabed+ b3*jobs$

I doubt whether there is a problem in this code as WC 5, SB35, EO200 appear in alternative 1 and alternative 2 (non-status quo alternatives) with a cost. I would appreciate it if you could have any idea of improving this code.

Cheers!
Niyomi

[Michiel Bliemer]

You had an error in your syntax, namely b1[[, I also made some other minor changes. Note that the mfederov algorithm does not guarantee attribute level balance, so please check the final design to see if all levels appear in the final design. You will have to run the syntax for a long time (at least several hours).

Code: Select all

Design
;alts = alt1*,alt2*,sq*
;rows = 48
;block = 8
;eff = (mnl,d,mean)
;bdraws = gauss(3)
;alg=mfederov(candidates = 5000)

;require:
alt1.cost > 0, alt2.cost > 0, alt1.cost <> alt2.cost,
sq.water = 5, sq.seabed = 35, sq.jobs = 200, sq.cost = 0

;model:
U(alt1) = b1[(n,0.0361,0.0109)]    * water[5,10,20,30]
        + b2[(n,-0.00569,0.00657)] * seabed[35,25,15,5]
        + b3[(n,0.00151,0.000563)] * jobs[200,300,400,500]
        + b4[(n,-0.0122,0.00266)]  * cost[0,20,40,60,80,100,120,140,160]
        /
U(alt2) = b1*water+b2*seabed+b3*jobs+b4*cost
        /
U(sq)   = b1*water+b2*seabed+b3*jobs+b4*cost
$

Michiel

[niyomip]

Thanks Michiel! appreciated.

Niyomi