Support for genetic modification in agriculture mainly stems from approval of food and agricultural goals. It is facilitated by trust in the judgment of scientific authorities and undermined by anxiety about the risks involved. But there are symptoms of danger: Any public opinion data that show significant correlations between perceptions of fact (risk, trust etc.) and background characteristics (age, sex, religion, politics) or goals (environmental, medical, economic) typically reflect emotional feedback effects as well as rational scientific ones. Estimates from regression are then biased and more complex models required. Our structural equation analyses of five large, representative national surveys of Australia (N = 8730) provide precise estimates of the magnitude of these effects, including reciprocal effects reflecting emotional influences. The author also finds that: (1) acceptance of the scientific worldview modestly increases support both directly and also indirectly through its influence on trust; (2) family socio-economic background increases knowledge of genetic engineering but is otherwise inconsequential; and (3) religious belief greatly hinders acceptance of the scientific worldview and slightly increases anxiety about risks.