The estimation results for the WTP model are reported in Table 4. The magnitude of
the coefficients in the model cannot be interpreted per se, but the sign and the marginal
effects odds ratios are useful in interpreting the results. Food safety attitudes and a reduction
in beef consumption following the BSE outbreak have a statistically significant
positive effect on theWTP for BSE-tested beef at the 0.05 level of significance, and being
female has a statistically significant positive effect at the 0.057 level of significance.
As expected, the coefficient of the bid variable is negative with a P-value of 0.001,
which means that as the bid offered for BSE-tested beef increased, the respondent was
less likely to choose the product.
Because the magnitude of the coefficients of a binary logistic model are not directly
interpretable, we also report the marginal effects in Table 4. For the significant parameters
we find that, ceteris paribus, for a marginal increase in concerns with food safety
the probability of accepting the premium is 0.22, the marginal effect of consuming less
beef after the BSE is a 0.15 probability of accepting the premium, and women had
an additional 0.13 probability of accepting the premium bid compared to men. It is
also interesting that demographic variable such as age and income were not significant
and that the variables that were significant did not have a very strong effect. This may
indicate that the BSE scare affected all age and income groups, which may also explain
the major drop in beef sales.
The suitability of the model is evaluated using the Pearson Chi-squared goodnessof-
fit test with a null hypothesis that the distribution is logistic. This test is applied to
detect major departures from a logistic response function, but it is not sensitive to small