Tavemier and Turvey (2006). The Chi-square test statistic tests the null hypothesis that
the coefficients of all the independent variables equal zero. The null hypothesis is
rejected at the significant level of 0.02 and indicates that the model has significant
explanatory power. The Mc Fadden R2 value, 0.20, also indicates an extremely good fit
for the estimated model (see Bell et al., 1994). This statistic is generally low in binary
dependent variable models estimated with cross-section data (Pindyck and Rubinfield,
1991). The percent of correct predictions is also reasonably high. This measure predicts
whether or not an event will occur given a set of explanatory variables (Judge, et al.,
1982). The model correctly predicts 71% of the responses.
The estimated coefficients from variables that measure the age of farm operators
and the percent of family income earned from farming or ranching are positive, while
coefficients from variables measuring annual gross sales, including government program
benefits, and education are negative. The results support the hypothesis, that in general,
farm operators whose sales are likely to be adversely affected by the use of social
engineering as a policy tool are unlikely to express a preference for such a policy.
Specifically the results suggest that except for farm operators with sales under $10,000
and farm operators with sales between $250,000 and $499,999, farm operators across the
sales category are unlikely to favor a policy allowing countries to restrict trade to pursue
domestic economic and social policy goals if the policies affect international trade. For
example farm operators with annual gross sales including government payments between
$500,000 and $999,999 are 80 percent less likely to indicate that countries should be
allowed to restrict trade to pursue domestic economic and social policy goals if the
policies affect international trade.
15