choice-metrics.com

Hello Good Afternoon,
I have these few doubts:

1) When we compare two alternatives (unlabelled in this case, and one alternative we know is an existing tech, say Heating boiler and another one is a new future upcoming tech whose WTP we need to find before releasing in market) ....can we not add up their each individual attribute's WTP values (because we do know which level of which attribute corresponds to which alternatve amongst the two, even if unlabelled) in such a scenario to tell about something like ''people are willing to pay say 2+3+1 = 6£ (where 2,3, 1 are the WTP values estimated for each attribute level) for Alt. Future Tech over Alt.Current Boiler.''
For example say people from doing the DCE said they are willing to pay 2£ for the higher efficiency level from the base (which we analysts know is provided in Alt.Future Tech) , 3£ for the 'sleek Look and appearance' Attribute and 1£ for the '24x7 customer service' attribute.
Are we then allowed to say people are willing to pay 6£ + Base price for this new Future tech??
2)Can we compare across two levels of an attribute for an alternative after we get the estimate results (without comparing with the base level which we usually do)
2) Can we separately find the estimate values of sociodemo variables on each alternative, say effect of age and gender on Alternative 1 vs Alt. 2( New tech) alternative?

Thank u

1) The marginal WTP is defined as the ratio of the marginal utilities of a single attribute, where marginal utilities refer to a unit increase in the attribute level. The total WTP (TWTP) is the sum of marginal WTP for all units consumed. WTP values are therefore dependent on the unit, for example $5 for 1 hour less travel time in the train and $10 for wifi on the train. Then the WTP for a combination of 1 hour less time on the train AS WELL AS wifi on the train (versus no wifi previously) would be $15. For a train alternative that has 2 hours less travel time instead of 1 hour, it would become 2*$5 + $10 = $20.

2) Yes, you can compare between any two levels, which does not need to include the base level. It does not matter which level is the base level as only differences in utility matter.

3) Yes. You would include age/gender in either Alternative 1 or Alternative 2. For example:
U(alt1) = ... + b1 * age + b2 * male
U(alt2) = ...
Then if b1>0 then alt1 is more preferred by older people, and if b2>0 then alt1 is more preferred by males.
You can also include age and gender as interaction effects with the attributes to determine age and gender specific preferences for attributes.

Michiel