Attribute level remains fixed despite assigning 3 levels

[izakro]

Dear all,
The design below has several restraints which are warranted given to nature of the problem. For some reason, the attribute NOLED.intensity receives the same level in all 18 scenarios (i.e. 20% reduction) despite the fact that it can be assigned with an additional level (i.e. no change). I was trying hard to figure out the problem with no success.
BTW, for those of you who wonder, we handled an issue of dependency between LED.esaving and LED.quality in a awkward way, knowing what we will be only able to estimate preference for LED.quality (the dependency is that If LED.quality=1 --> esavings =60% and if LED.quality=2 --> esavings=40%).
design
;alts = led*, noled*, bau*
;rows = 18
;block = 3
;eff = (mnl,d)
;alg = mfederov(candidates=3900)
;require:
bau.intensity=2
;reject:
led.quality=1 and led.cost_l>5,
led.quality=2 and led.cost_l<40,
noled.esaving_nl=20 and noled.cost_nl=150,
noled.intensity=3
;model:
U(led) = led[0]
+intensity.dummy[0.001|0.002]*intensity[2,3,1]? 1=20%reduction, 2=nochange, 3=20% increase
+quality.dummy[0.001]*quality[2,1]? 1=high glare and discomfort, 2=moderate glare & discomfort
+cost[-0.00001]*cost_l[5,40,80,120]
/
U(noled) = noled[0]
+ intensity * intensity (4-6, 4-6,4-6)
+ esaving[0.00001] * esaving_nl[20,30,40]
+ cost * cost_nl[150,200,250]
/
U(bau) = intensity * intensity
+ esaving * esaving_b[0]
+ cost * cost_b[290]

;formatTitle = 'Scenario <scenarionumber>'
;formatTableDimensions = 4, 6
;formatChoices = choice 1
;formatTable:
1,1 = 'attributes'/
1,2 = 'Light intensity'/
1,3 = 'Light glare and discomfort' /
1,4 = 'Energy savings'/
1,5 = 'Cost'/
1,6 = 'your choice'/
2,1 = 'LED'/
2,2 = '<led.intensity>' /
2,3 = '<led.quality>' /
2,4 = 'If glare is high savings =60% if glare is moderate savings=40%'/
2,5 = '<led.cost_l>' /
2,6 = '<choice 1>' /
3,1 = 'Savings without replacing streetlights' /
3,2 = '<noled.intensity>'/
3,3 = 'no change'/
3,4 = '<noled.esaving_nl>' /
3,5 = '<noled.cost_nl>' /
3,6 = '<choice 1>' /
4,1 = 'status quo'/
4,2 = 'no change'/
4,3 = 'no change'/
4,4 = '<bau.esaving_b>'/
4,5 = '<bau.cost_b>'/
4,6 = '<choice 1>'
;formatStyleSheet = Blue buttons.css
;formatAttributes:
led.intensity(1=20% reduction, 2=no change, 3= 20% increase)/
led.quality(1=high glare and discomfort, 2= moderate glare and discomfort)/
led.esaving_l(40= #%, 60= #%)/
led.cost_l(5=NIS #, 40=NIS #, 80=NIS #, 120=NIS #)/
noled.intensity(1=20% reduction, 2=no change, 3=20% increase)/
noled.esaving_nl(20= #%,30= #%, 40= #%)/
noled.cost_nl(150=NIS #, 200=NIS #, 250=NIS #)/
bau.cost_b(290=NIS #)
;formatTableStyle:
1,1 = 'default' /
1,2 = 'headingattribute' /
1,3 = 'headingattribute' /
1,4 = 'headingattribute' /
1,5 = 'headingattribute' /
1,6 = 'headingattribute' /
2,1 = 'heading1' /
2,2= 'body1' /
2,3= 'body1' /
2,4= 'body1' /
2,5= 'body1' /
2,6= 'choice1' /
3,1 = 'heading2' /
3,2= 'body2' /
3,3= 'body2' /
3,4= 'body2' /
3,5= 'body2' /
3,6= 'choice2' /
4,1 = 'heading3' /
4,2 = 'body3' /
4,3 = 'body3' /
4,4 = 'body3' /
4,5 = 'body3' /
4,6 = 'choice3'
$

[Michiel Bliemer]

This is because you have heavily constrained your attribute levels in this design, there only exist 96 feasible choice tasks. You can see them using the syntax below. Note that your dominance checks have a large impact since you only have three attributes, each with a few levels, so for Ngene to avoid dominance across the choice tasks there are not many feasible choice tasks left. Given that you seem to have a labelled experiment, I think your dominance constraints may be too strict.

Code: Select all

design
;alts = led*, noled*, bau*
;rows = all
;fact
;require:
bau.intensity=2
;reject:
led.quality=1 and led.cost_l>5,
led.quality=2 and led.cost_l<40,
noled.esaving_nl=20 and noled.cost_nl=150,
noled.intensity=3
;model:
U(led)   = led[0]
         + intensity.dummy[0.001|0.002] * intensity[2,3,1]   ? 1=20%reduction, 2=nochange, 3=20% increase
         + quality.dummy[0.001]         * quality[2,1]       ? 1=high glare and discomfort, 2=moderate glare & discomfort
         + cost[-0.00001]               * cost_l[5,40,80,120]
         /
U(noled) = noled[0]
         + intensity                    * intensity
         + esaving[0.00001]             * esaving_nl[20,30,40]
         + cost                         * cost_nl[150,200,250]
         /
U(bau)   = intensity                    * intensity
         + esaving                      * esaving_b[0]
         + cost                         * cost_b[290]
$

Michiel

[izakro]

Thank you Michiel, we will release some of the constraints.