5.2. Estimation results
We start by explaining the estimation results of the two-alternative
model, given in Table 6a. The total number of observations is 16000 considering
that 2000 respondents answer eight questions. McFadden's
pseudo R2 is 0.537. The estimation results indicate bothmeans and standard
deviations (S.D.) for two random parameters. Assuming a normal
distribution for nuclear power generation, we obtain a mean with the
expected sign (−0.117), which is significant at 1%. Further, assuming
a non-positive β distribution for electricity rate, we obtain a mean
with the expected sign (−0.200), which is again significant at 1%.
Note that the beta distribution assumes equality of mean and S.D.
values. The corresponding marginal effects, which measure the
probability change caused by a specific coefficient change, are −0.105
for nuclear power generation and −0.309 for electricity rate. The S.D.
estimates (0.117 and 0.200) are also significant at 1% for both random
parameters, indicating that respondents have diverse preferences. We
calculate the WTP values showing the electricity rate (per kWh) they
are willing to pay given that the dependency on nuclear power
generation decreases by 1%. The WTP value, which is defined as the
mean nuclear power generation coefficient divided by the mean
electricity rate coefficient with negative sign, is −β1n/β2n = −
(−0.117/−0.200)= −0.583 cents at the mean values