0

I am using the support.CEs package in R for a choice experiment. I am attempting to calculate Marginal Willingness To Pay (MWTP).

The package seems to be relatively new, and I have reviewed the CRAN documents as well as the article found here:

Choice Experiments in R

As the example in the Choice Experiments in R article (Aizaki, 2012, page 18-19) demonstrates, the variables with positive coefficients in clogit results also have positive MWTP.

I have just completed a Choice Experiment survey and in reviewing the results it is very clear that respondents are willing to pay more for Organic products. The coefficient for the Organic variable is significantly positive, yet the MWTP results show a negative MWTP for for this attribute.

I suspect that this is either due to a negative coefficient for the ASC OR the fact that Price is showing as significantly positive (an unusual result, but given the audience it is not completely surprising).

I know that the Organic variable should have a positive MWTP as it has a positive coefficient. Is anyone able to help make sense of this?

Thanks so much.

dan1st
  • 12,568
  • 8
  • 34
  • 67
user2297202
  • 1

2 Answers2

2

The WTP calculations are simply the ratio of the attribute coefficient to the (negative) price coefficient. For example the Organic MWTP = -4.4918/5.7862 = -0.7762953.

Theoretically, price needs to have a negative sign and the MWTP then reflects the tradeoff between price and the positively valued attribute. A positive price coefficient will give you negative MWTPs.

BTW: This is not an R problem but rather a problem with (possibly) your data and the interpretation of the results.

acryd
  • 46
  • 2
0

the growth is exponential. if we assume the growth to be equal to n;it means that that the ability to respond to the needs to pay is positively increasing. if we take sum of b3X3i+y as the r^2-adjusted we get a positive value. also the price is positively increasing. the negative coefficient comes from the two n's; that is number of events-n=232-696=-464. my question is what happens when p=1? as i can see from the table, it shows that p=1 has been clogout. the command mwtp

user243620
  • 1
  • 3
  • has cloggedout 1 from the table. The rhetorical question has been answered. let B represent coeff, X represent exp(coeff) and Y represent se(coeff). i hope this answers your question. any more suggestions would help. – user243620 Feb 13 '21 at 17:26
  • d=c-1, d=c-0, d=c, g=n-1, g=n, muf=mpc/p=n-mpc/1, d>bp, d>bwp, d>bpw. from a marginal utility graph, equilibrum Do, y=d, x=p line graph. multiply by growth, g; gd>gbpw/p; g/g>bw/dp; GDP>GBW; 1>bw/dp; p>bw/d. when p=1; bw/d=bw/dp. p=dbw/dbw=1. – user243620 Feb 15 '21 at 01:39
  • it is due to a negative mwtp for Asc which is not tabulated because there cannot be a negative coeff and a negative mwtp for Asc. to prove this right, mwtp-p/3 for the organic is equal to the negative mwtp for Asc, which is wrong. so, therefore this is due to the negative mwtp for the Asc which is not tabulated from your results on the table. – user243620 Feb 15 '21 at 01:55
  • i hope these will help to answer your question. more suggestion would be appreciated. – user243620 Feb 15 '21 at 01:56
  • the organic has a negative mwtp, ''which is true''. the value of the mwtp is used to calculate the mwtp for the Asc which is negative and of the same value as the mwtp for the organic. – user243620 Feb 15 '21 at 02:09
  • s u m a i = { } – user243620 Mar 27 '21 at 13:17