Conjoint Analysis Question: Estimating Willingness to Pay by Gender

[NCSUStudent]

H Conjoint Analysis Question: Estimating Willingness to Pay by Gender
Earlier last year, I fielded a choice-based conjoint survey to 770 respondents. The aim of this study was to how much people were to willing to pay to improve water quality. The survey utilized a fractional factorial design and included 4 choice tasks, each of which asked respondents to choose between 2 policies and a no-choice alternative.

I am currently analyzing the results and I would like to see how willingness to pay (WTP) for changes in water quality differ by gender. However, I am having some trouble.

Specifically, lets say the hypothetical policy we asked respondents to evalauate was described using only 2 attributes--price and water quality (wq):low, medium, and high (the actual policies had more attributes but i want to keep it simple for my question). If I code price as a continuous variable and wq as effects coded, then I can estimate the marginal utility of income and the part-worth of utility for each wq level using a mixed logit regression.

Below you will find these results estimated when interacting the water quality attribute levels with a binary variable for gender (0=male, 1=female)

MIXED LOGIT REGRESSION COEFFICIENTS
price:-0.00015
wq-low: 0.13
wq-medium: 0.20
wq-low x female:0.27
wq-medium x female: -0.14

Because water quality is effects coded, I can estimate the part-worth utility for the high level for males as such: -[(0.13)+(0.20)] = 0.33. Therefore, to estimate how much men are willing to pay to move from a low level of water quality to a high level (everything else held constant), we calculate: [-((0.33)-0.13))/(-0.00015)]=$1,333

But how do I estimate this same WTP for females?

[NCSUStudent]

Just a note for clarification. Here is how I thought about estimating how I am thinking about this. The part-worth utilities for females are the interaction terms PLUS the main effects coefficients are the part-worth utilities. So, I think this is how i should do it:
wp-low (female) = [(0.13)+(0.27)]=0.40
wp-medium (female) = [0.20+(-0.14)]=0.06
wp-high (female) = -[0.40+0.06] = -0.46

If that is right, then this implies that women responding to my survey actually recieved negative utility from water quality (assuming these results are statistically significant, which i am not quite sure how to test in this context, i guess by bootstrapping)?

If so, it seems to me from this point I can just estimate willingness to pay for women the same way as before, just using the part-worth utilities above. Right?Any help would be GREATLY appreciated. :confused: